| Age | Commit message (Collapse) | Author |
|---|
|
Changelog
=========
Crosscompiling: forgot to 'git add' one file in last commit (90997f4)
Crosscompiling: improve SBCL download, ectraction and installation. Update crosscompiling documentation (README.txt and TODO.txt) (b77f502)
Crosscompiliing: Improve VTK download, extraction and installation. (cc63d8a)
Docs: add newlines so paragraphs and examples don't run into each other (c12dae7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6dee329)
Tell gnuplot not to round down everything < 1e-8 to zero. (e2a6f50)
trap floating point overflow in ecl (588ff7e)
use signal for floating point overflow (dae685a)
Fixed bug #3246: "Integrating u'(x) * f(u(x) + c) fails for f any inverse trigonometric/hyperbolic function" by adding look-up table entries for the antiderivatives of all inverse trigonometric/hyperbolic function. Added test cases and modified an existing test case because integrate(atan2(y,x),y) now returns an equivalent, but slightly different form than before. (67fd622)
test for overflow in exp(1000),numer (06b469f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (47c7d24)
Remove redundant definition of copy from share/affine, and move documentation from linearalgebra.texi to Expressions.texi. Resolves bug reported to mailing list 2016-11-12: "copy question" (74800c2)
Crosscompiling: Update SBCL (15c5d17)
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Crosscompiling: improve SBCL download, ectraction and installation. Update crosscompiling documentation (README.txt and TODO.txt) (b77f502)
Crosscompiliing: Improve VTK download, extraction and installation. (cc63d8a)
Docs: add newlines so paragraphs and examples don't run into each other (c12dae7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6dee329)
Tell gnuplot not to round down everything < 1e-8 to zero. (e2a6f50)
trap floating point overflow in ecl (588ff7e)
use signal for floating point overflow (dae685a)
Fixed bug #3246: "Integrating u'(x) * f(u(x) + c) fails for f any inverse trigonometric/hyperbolic function" by adding look-up table entries for the antiderivatives of all inverse trigonometric/hyperbolic function. Added test cases and modified an existing test case because integrate(atan2(y,x),y) now returns an equivalent, but slightly different form than before. (67fd622)
test for overflow in exp(1000),numer (06b469f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (47c7d24)
Remove redundant definition of copy from share/affine, and move documentation from linearalgebra.texi to Expressions.texi. Resolves bug reported to mailing list 2016-11-12: "copy question" (74800c2)
Crosscompiling: Update SBCL (15c5d17)
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Docs: add newlines so paragraphs and examples don't run into each other (c12dae7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6dee329)
Tell gnuplot not to round down everything < 1e-8 to zero. (e2a6f50)
trap floating point overflow in ecl (588ff7e)
use signal for floating point overflow (dae685a)
Fixed bug #3246: "Integrating u'(x) * f(u(x) + c) fails for f any inverse trigonometric/hyperbolic function" by adding look-up table entries for the antiderivatives of all inverse trigonometric/hyperbolic function. Added test cases and modified an existing test case because integrate(atan2(y,x),y) now returns an equivalent, but slightly different form than before. (67fd622)
test for overflow in exp(1000),numer (06b469f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (47c7d24)
Remove redundant definition of copy from share/affine, and move documentation from linearalgebra.texi to Expressions.texi. Resolves bug reported to mailing list 2016-11-12: "copy question" (74800c2)
Crosscompiling: Update SBCL (15c5d17)
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6dee329)
Tell gnuplot not to round down everything < 1e-8 to zero. (e2a6f50)
trap floating point overflow in ecl (588ff7e)
use signal for floating point overflow (dae685a)
Fixed bug #3246: "Integrating u'(x) * f(u(x) + c) fails for f any inverse trigonometric/hyperbolic function" by adding look-up table entries for the antiderivatives of all inverse trigonometric/hyperbolic function. Added test cases and modified an existing test case because integrate(atan2(y,x),y) now returns an equivalent, but slightly different form than before. (67fd622)
test for overflow in exp(1000),numer (06b469f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (47c7d24)
Remove redundant definition of copy from share/affine, and move documentation from linearalgebra.texi to Expressions.texi. Resolves bug reported to mailing list 2016-11-12: "copy question" (74800c2)
Crosscompiling: Update SBCL (15c5d17)
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
|
|
Changelog
=========
Fixed bug #3246: "Integrating u'(x) * f(u(x) + c) fails for f any inverse trigonometric/hyperbolic function" by adding look-up table entries for the antiderivatives of all inverse trigonometric/hyperbolic function. Added test cases and modified an existing test case because integrate(atan2(y,x),y) now returns an equivalent, but slightly different form than before. (67fd622)
test for overflow in exp(1000),numer (06b469f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (47c7d24)
Remove redundant definition of copy from share/affine, and move documentation from linearalgebra.texi to Expressions.texi. Resolves bug reported to mailing list 2016-11-12: "copy question" (74800c2)
Crosscompiling: Update SBCL (15c5d17)
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (47c7d24)
Remove redundant definition of copy from share/affine, and move documentation from linearalgebra.texi to Expressions.texi. Resolves bug reported to mailing list 2016-11-12: "copy question" (74800c2)
Crosscompiling: Update SBCL (15c5d17)
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Crosscompiling: Update SBCL (15c5d17)
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Another fix for summation of rational functions: make sure there are no poles in the summation area. (8ec3a05)
Pass all required arguments to changevar in recursive calls. (985ee62)
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Merge branch 'scratch-pad' removed atomson in matcom.lisp. It was just a simple version of "(remove-if-not #'atom list)". (e354f84)
removed atomson. this function was used only in two places in matcom.lisp and was a recursive version of "(remove-if-not #'atom list)" no errors in tests. run_testsuite has tests that use this branch of the code. (9c63bed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (e042ecb)
In parse_timedate, handle timezone offsets, ignore leading and trailing whitespace, and allow any number of digits in fractional seconds. (27df459)
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Comment out wrong code. (989086a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (c7f1018)
In package facexp, remove "&&" because load(facexp) complains about that. Fixes bug reported to mailing list 2016-11-01: "facexp, pls fix" (394ca66)
In RGRP, rephrase test for ECL to avoid comparing float infinity with rational, which tickles a bug in ECL (reported as https://gitlab.com/embeddable-common-lisp/ecl/issues/299). Fixes SF bug #3235: "ECL lisp arithmetic error in definite integration with large limits" (e916dd0)
Use return value from complex numerical eval (537cbc0)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5f3c61a)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Fix uses of complex-float-numerical-eval-p (dd6fbb1)
complex-float-numerical-eval-p should work like bigfloat (4088e70)
Use return value from complex-bigfloat-numerical-eval-p (c674047)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (4c967aa)
Reindent badly indented code. (b7aaad5)
Use return value from complex-bigfloat-numerical-eval-p (e371633)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (f60c713)
Reindent badly indented code. (6bd6b84)
Use return value from complex-bigfloat-numerical-eval-p (01c6d1e)
Merge branch 'master' into rtoy-complex-numerical-eval-returns-values (5656996)
Reindent badly indented code. (4fb8843)
Convert maxima rational to lisp rational (06306dc)
Use results from complex-float-numerical-eval-p (657f405)
Don't convert reals to complexes. (2a0ec86)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Use the new complex numerical eval function for elliptic_pi computation (3250089)
Complex numerical eval returns args (766f655)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (314ae8e)
Ensure that *OPR-TABLE* associates "{" with $SET instead of ${. Fixes bug reported to mailing list 2016-10-27: "Subtle change in apply (with simp:false)" (b7ee537)
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Reindent badly indented code. (b7aaad5)
Reindent badly indented code. (6bd6b84)
Reindent badly indented code. (4fb8843)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (d8f72ec)
In package Solver, remove solver.mac since it causes trouble on case-insensitive file systems. (5ebc1ba)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Ran simplifications.texi through update_examples: There were no errors in the results this time. But they partially differed with what maxima currently outputs as response. (8d80db1)
Documentation: A link from factor to collectterms. (5aae6c8)
Documentation: Tried to improve the googleability of collectterms and added an example to its manual entry. (71ddb4e)
Ran update_examples on Simplification.texi and corrected the only example that was broken. (9f43849)
Reindent badly indented code. (b7aaad5)
Reindent badly indented code. (6bd6b84)
Reindent badly indented code. (4fb8843)
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
|
|
Changelog
=========
Documentation: Scanmap was lacking an anchor leading to warnings during compilation. (30e12db)
Documentations: Cross references between append, addrow and addcol, since these functions all append elements to objects. (79f030d)
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Neatly indent simp-$elliptic_pi (f3a3c59)
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
* tests/rtest_elliptic.mac: Save and restore fpprec. (d7c6950)
Handle complex phi in elliptic_pi (f2788d6)
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Fix bug in running test files. Variable test-file-path not updated, same file run twice. Introduced in [22d564bc3f1b0d2412da7b3d5343e3c724923beb] (07363c2)
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Add tests from but #3221 (7060221)
Fix #3221: elliptic_pi(n,z,m) wrong for z > %pi/2 (2204fca)
Renamed solver.mac to Solver.mac (6620590)
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
In package Solver, create new solver.mac which prints a message and loads Solver.mac. (35ffdc3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (2a6493f)
In package Solver, rename solver.mac to Solver.mac to make package name coincide with function name. Also correct name of function in solver/test.mac. (6aebc96)
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
* tests/rtest_algsys.mac: Add two tests from maxima-dicuss list. (1b38672)
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Make run_testsuite(share_tests=true) work (22d564b)
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Crosscompiling: Update SBCL. (3ba4f23)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bc706f7)
Documentation: A "see also" section for map (). (4cf489c)
In package stringproc, return default encoding for SBCL. (98067d7)
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Avoid asksign question about sign of 0 in diff(elliptic_pi ...). (c262392)
* src/algsys.lisp:simplify-after-subst Limit expression growth from rectform (e49d612)
* tests/rtest_algsys.mac: More tests. (e1ed80f)
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
|
|
Changelog
=========
* tests/rtest_algsys.mac: Enable two tests. (4deba9a)
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
*src/algsys:simplify-after-subst Set ($rootsconmode t) ($radexpand t) (8650644)
[bug:#453] algsys fails in simple case (cb3dc1c)
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
* tests/rtest_algsys.mac: Add examples from Beyer (1984) (b782138)
In package charsets, flatten.lisp was renamed charsets_flatten.lisp. (4735682)
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
i* tests/rtest_algsys.mac: Add tests from SF bug 2059. (7efa7eb)
* tests/rtest_algsys.mac: Add tests from SF [bugs:2038] and [bugs:#2736] (4851e2c)
SF [bug:#1266][bug:#3208] Don't use $radcan when simplifying in algsys. (640ca75)
update timestamp example (b277e59)
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Add PLY support (a0cfb94)
Fix SF [bugs:#3208]. Use sqrtdenest in algsys to simplify partial solutions. (2986fc0)
* src/sqrtdenest.lisp: New file with lisp sqrtdenest function. (d0a0a0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (0ee6576)
* share/simplification/rtest_sqdnst.mac: Add new tests. (28d174e)
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Use REDUCE instead of APPLY in UCONTENT (3672c77)
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Restructure src/algsys.lisp:ebaksubst1 to separate call to radcan (86f2ae3)
src/algsys.lisp:ebaksubst New function ebaksubst1 (95d60c0)
[bugs:#3208] algsys.lisp:presultant: resimplify result (5fd2f55)
Tidy up recent commit to tests. (9eeff35)
tests/rtest8.mac: Delete two algsys tests moved to rtest_algsys.mac (e4df541)
New algsys tests collected while investigating bugs. (5d801a5)
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Commit f6cf5718 triggered commit 4734ec8a (which updated the list of share subdirectories) by making "make html" create additional subdirectories on systems running the current version of texi2any. (2da3703)
Manual: Updated the documentation of openr and openw. (d6bba2f)
share-subdirs.lisp is auto-generated *and* in version control - which is afaik unavoidable for the lisp-only build system. It also currently is out-of-sync with the repo => Updated the repo version. (4734ec8)
Removed two auto-generated documentation from git: Otherwise a "make clean" deletes them and a "make html" regenerates them either as .html files or as .html directories containing one file per chapter. Both makes doing a "git bisect" an nightmare. (a775e97)
Another error message that perhaps needs a googleable explanation. (3bacf3f)
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Manual: A example value for maxima_tempdir that matches what is to be expected of the next version of maxima. (cb7928f)
Manual: Added documentation for another frequently-asked-for error message. (bd385c6)
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Apply patch from Leo (a3263ea)
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
resultant() algorithm mod is commented out. Document this. (b9fdad7)
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
tests/rtest_elliptic.mac: Remove debugging prints from test_deriv() (13b1e70)
Document system variable %rnum in doc/info/Equations.texi (4af1e76)
tests/rtest_elliptic.mac: a working version of test_deriv(). (ce5caf9)
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Update comments for test_deriv() in tests/rtest_elliptic.mac (0926376)
Test derivatives of elliptic_{f,e,pi} using numerical differentiation. (8871ff4)
Define derivatives of elliptic_pi() wrt all arguments. (106c95d)
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Crosscompiling: removed (obsolete and now wrong) comment. (a7fde4b)
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Update docs for elliptic_pi(). Promise less. (172b8fa)
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Allow subscripted variables in ode_check (dcc8f9b)
Increase output when running share/contrib/diffequations/tests (f91d677)
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Crosscompiling: updated documentation. (8f6ad20)
Correct ODEs kamke 2.232 and 2.233 (e45aa77)
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Call kovacicODE() from contrib_ode for 2nd order linear ODEs * share/contrib/diffequations/contrib_ode: Add call to kovacicODE() * share/contrib/diffequations/tests/rtestode_kamke_2_*.mac: Update expected solutions to ODEs (2f89b26)
share/contrib/diffequations/tests/rtestode_odelin.mac testsuite update - Enable tests that were disabled due to algsys regression - Update ode soultions that have changed form (9ef3062)
Enable tests that were disabled due to algsys regression * share/contrib/diffequations/tests/rtestode_kamke_1_1.mac * share/contrib/diffequations/tests/rtestode_kamke_1_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_1.mac * share/contrib/diffequations/tests/rtestode_kamke_2_2.mac * share/contrib/diffequations/tests/rtestode_kamke_2_3.mac * share/contrib/diffequations/tests/rtestode_kamke_2_4.mac * share/contrib/diffequations/tests/rtestode_kamke_2_5.mac (a52dd0e)
SF bugs 3210 and 3212. Fix algsys regressions. (a5873ba)
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
expintegral_e_simp was renamed to expintegral_e_simplify. Fix call. (d6dc64b)
Tweak solution for Kamke ODE 1.518 (fbab61f)
Tweak solution for Kamke ODE 1.26 (eeb479d)
Bug #2796 can't load ode2 with n declared constant (9a4694f)
Bug report ID: 1621 Wrong solution to ode2 (3c6c473)
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Triage contrib_ode testsuite to enable it to run to completion. Disable 39 tests that "hang" in odelin. Fix a few minor differences in results. (4b0bb5d)
Actually rename expintegral_e_simp to expintegral_e_simplify. This was started 2015-09-06 but missed the actual rename. (61a0ec1)
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
Declare split_sum as evfun. (1ea2ca7)
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|
|
Changelog
=========
throw 'limit on failure (840643d)
limit(floor(sin(x)), x, 0, minus); (7823bb5)
Crosscompiling: Update SBCL to version 1.3.9 (fda193d)
remove assumption that lim-epsilon < 10 ^ -8 (c54f533)
Fix sign for li[3](-1.0) (e698e55)
Documentation: Mentioned the 'geomview' plot_format. (4acbf2a)
Update documentation on xlabel, ylabel and zlabel (65ca7d1)
Update documentation on xlabel, ylabel and zlabel (6b5187e)
* Add support for axes labels in 2D VTK scenes. * Revert default axes labels to empty strings. We want the same behavior for Gnuplot and VTK. (cf8ac7f)
Remove background rectangle in epslatex_standalone files. (af53a13)
Additional test cases for mod. (753dd2d)
In TeX processing for exponents, distinguish trig-like functions from others. Output foo^n x for trig-like foo and foo(x)^n otherwise. Fixes SF bug #2667: "tex1 of f(x)^n" Thanks to Leo Butler for the patch. (cf9cbb2)
Update list of built-in symbols by extracting all @defvr and @deffn items from doc/info/*.texi. (179efa3)
In rtest_log, additional tests for SF bug #3105. (bf4cf94)
In rtest_limit, li[4](1.0) now simplifies to a float. (15155a4)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8ed2077)
In rtest5, change the documentation index that is loaded to one that exists now. share/MYTOPIC.info is the output from makeinfo template-standalone.texi in share/. (ffa4ef1)
Adjust the numerical tolerances of 10 tests for Allegro CL (5be9543)
Crosscompiling: Updated wxwidgets to 3.1.0 (recommended by Gunter) (does currently work with the current wxMaxima Git version, but I assume, there will be a wxMaxima release before the next Maxima release) (d688caf)
VTK: Windows was not correctly identified, then the space between the vtkpython and the script was missing (the command to be executed was: vtkpythonC:) Hopefully fixed the issue. (7c85cff)
Documentation: Fix 'make html' - the make process failed in share/contrib/maxima-odesolve (723bb65)
Documentation: Changed abs_integrate.texi, so that the generated PDF is readable. (some lines were way too long) Some results changed in current Maxima. The referenced URL is redirected to a new one, changed the URL. Made the URL clickable. (6457769)
Hyperrefs from isolate to disolate and vice versa. (27e09fc)
Fix building logic.pdf from the texinfo source. Remove the now generated logic.pdf from the repository. (3af27a6)
Remove background rectangle in eps files. (d7a16ce)
Crosscompiling: Use @SEVENZIP_EXE@ variable in sbcl.sh "installation script" Add information about the included software (not only Maxima, but also Gnuplot, TCL/TK, VTK, ...) and their licenses to the start menu. (bce4e64)
Cut out unreachable code and the special *MACSYMA-EXTEND-TYPES-SAVED* (bfa29d3)
Fixed bug #3148 ("sign can't figure out sign(a - b) but it knows sign(b - a) where a and b are exponentials"). Added test cases. Note regarding the change in compar.lisp: Most of the change shown by Git is simply a change in indentation. See the discussion on #3148 for a detailed explanation. (9ef3fee)
Changed a semicolon to a dollar sign in a test case (279c5c3)
Fixed bug #3194 ("No simplification of tan(x+n*%pi) and cot(x+n*%pi) with n being a declared integer"). Added test cases. (8f523fc)
Update to mention the "website" repo instead of the old "site-xml" repo (e85b382)
Bug fix: csign of a product depended on the order of the factors (ea9e4b8)
Port VTK interface from TCL to Python (27091ad)
Removed Toolchain-mingw32.cmake but did not remove it from Makefile.am (dcdd027)
Crosscompiling: CMake-Changes for using a VTK python interface. (not finished yet! Marios must commit his changes for vtk.lisp that that will work.) (c9382f6)
The documentation change needed after adding .wxm to file_type_maxima in my last commit (953017e1f4) (8835cde)
Now maxima can directly load .wxm files from wxMaxima (953017e)
Remove a broken EQ comparison in SIMPLIM%GAMMA_INCOMPLETE (0902ac8)
Crosscompiling: Update SBCL, TCL, TK (d51d278)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (91d2ddc)
In f90, bind special variables to ensure that exponent markers are printed appropriately for Fortran. New lines of code are copied from src/fortra.lisp. (3c71534)
Fix some shell quoting issues (f8e1509)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (fc2d2c3)
Another test for triangularize, this time with tellrat. (53345c2)
In triangularize, bind algebraic according to logic copied from echelon. Fixes SF bug #3158: "triangularize gives incorrect result on a matrix containing %i" (c9375b3)
Draw/vtk: Minor corrections to a few comments (0ab4c89)
Draw: Explicitely check if all temporary files can be created. If draw doesn't manage to create temporary files that allow it to communicate with gnuplot the result is either a lisp error (which is fine) or a blank plot without a description why (which leads to long debugging sessions. Added a few lines of code that should catch the latter case. (acfc005)
In LIMIT-CONTEXT and MAKE-LIMIT-ASSUMPTIONS, assume integer or rational values instead of floats. Fixes SF bug #3185: "Error in sum with non-false modulus" (2f7df6f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5cd9c91)
Revert commit 99a8ba2, so that totalorderp returns indices in the sorted list, per discussion on mailing list. (b06e9b8)
Additional tests for numerical evaluation of elliptic functions, this time for complex bigfloats. Mark two tests as expected failures since elliptic_eu and jacobi_am cannot handle complex bigfloats. (d6f86ee)
Verify that f(u, k) = f(rectform(u), k) for elliptic functions, where u is a complex float. This tests commits c93feaa, 372582e, 133408b, and 80cab10, which fix the issue raised on the mailing list 2016-07-14: "Jacobi elliptic functions, maxima 5.38.1" (4c4acd7)
In inverse_jacobi_cd, turn on numer flag to cause sqrt to simplify to a numerical value when we punt to inverse_jacobi_sn. (80cab10)
Further improvements for numerical evaluation of elliptic functions. (133408b)
In some elliptic functions, handle float, complex float, bigfloat, and complex bigfloat arguments uniformly. In particular, call rectform on complex floats and complex bigfloats, since the functions which detect those arguments (COMPLEX-FLOAT-NUMERICAL-EVAL-P and COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P) are also working with the rectangular form, via TRISPLIT. (372582e)
Make COMPLEX-BIGFLOAT-NUMERICAL-EVAL-P exactly analogous to COMPLEX-FLOAT-NUMERICAL-EVAL-P. (c93feaa)
Make maxima.bat no more deviate the temp dir location to the user's home dir. (144eb0b)
Crosscompiling: It might be necessary to add the i386 architecture. (de943ed)
README.rpms: Changed CVS to GIT. (0df86ed)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (381a646)
Bug fix: now things like fourier_elim(2^x < 2^y, [x, y]) yield [x < y] (86fd249)
Reorder the results of some fourier_elim tests (7915cf7)
Crosscompiling: Update Gnuplot to 5.0.4 (95451f2)
Fix #3186 and other get_plot_option issues. (fdadebc)
Merge branch 'tmpdir' (a0c55de)
Try to find a real temp directory for temporary files (48f3100)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (ded646e)
In rtest1, modify totalorderp to return indices in original input instead of indices in sorted input, in hope of making results more informative. (99a8ba2)
In rtest1, paste in totalorderp from mailing list 2016-07-18 "Testing for total order (for orderlessp testing)" courtesy of Stavros Macrakis. Create 2 new test cases and mark them expected failures. (4c4c446)
Crosscompiling: Updated TODO.txt Some of the TODOs are already done. (8095139)
Fix bug #3022: Update the RPM spec file (3fd2b4c)
In package boolsimp, remove comments which are no longer relevant. (5ccb06d)
In package boolsimp, cut out explicit resimplification in evaluation. It is unneeded since simplification is handled after evaluation in general. (26c95ba)
In package boolsimp, remove comments about examples that are already handled in Maxima. No need to load boolsimp to handle those. (a981d32)
In package boolsimp, remove redundant comments about simplification of conditional expressions. (There is another comment block describing the same topic.) (e7705a2)
In package boolsimp, cut out redefinition of INTEXT of unknown purpose. rtest_boolsimp runs without errors, so it is not required for that, and I do not remember what problem it was supposed to solve. (2e356b0)
In package boolsimp, cut out commented-out definitions of $EVERY and $SOME. I forget what I was attempting to do there, and at this point I think it is probably a poor idea to redefine stuff like that. (0d763c5)
In TEST-BATCH, print line number of each problem as well as problem number. (d80b253)
Fix an egrep pattern which broke the finding of the share directories (d642fe7)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (bd58cfd)
Mentioned ABCL in README.lisps. (18cd7c0)
Remove some cruft from the default props information list (6fe1cc9)
Add --enable-quiet-build option (659c535)
Another .gitignore entry that handles another temp directory created by "make pdf". (9584036)
Documentation/Lists: Robert has informed me that there are lisp implementations that allow for fast random access to lists. Updated the documentation so it doesn't say that random access times need to be proportional to the number of list elements. (e0e600f)
Adjust the numerical tolerances of 6 tests for ABCL (874e521)
Work around a COERCE bug in ABCL (b1c9c0d)
Don't need defpackage for INTL if lisp already has INTL. (4564ad1)
Ignore *.t2p directories in the share directories. (012756f)
.gitignore: added several files autogenerated by "make pdf". (c5b60d3)
Resolved a build failure for "make pdf". I hate build failures as they tend to make "git bisect" a nighmare. (4aaa2b0)
.gitignore: Tell git to ignore files generated by a run of "make all". (3d14139)
Documentation/error messages: Added missing catbox macros (52da717)
Documentation: Alphabetically sorted the chapter about error messages. (e2be6c83)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (181e8d3)
Mark tests that require the docs as known failures in Lisp-only builds (94eecb5)
Define *AUTOCONF-LISP-ONLY-BUILD* which is T in Lisp-only builds (ad3cf4e)
Update documentation for continuous_freq and discrete_freq, which now take arrays as input as well as lists. (0ad5565)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (4ce4be1)
In continuous_freq and discrete_freq, accept arrays as well as lists. (33e9c87)
TexInfo docu: Documented another error message I get frequent questions about. The "Zero to a negative exponent" actually describes completely what has happened. But documenting that this is a "division by 0 error" improves this error's googleability. One of the next steps will be sorting the error messages chapter alphabetically as sorting the error messages by themes or similar doesn't make much sense as long as the user doesn't know what caused them and therefore won't know in which section to look, anyway. (87c2620)
Force evaluation of expressions in transform_sample. (e9514e4)
Use ccl64 for ccl64 tests. (6bfefdd)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (a4a0777)
In SIMPLIFYA, cut out unneeded special cases for MPLUS, MTIMES, and MEXPT. (55f8004)
Updated information about Clisp in our README.lisps (the same than on our website). (42c230a)
Crosscompiling: update SBCL (caf3faf)
Did forget to change configure.ac, when removing the crosscompile-windows/wxwidgets directory. Sorry. (9b0cb16)
No longer compare numbers with EQ [part 4] (99b101f)
No longer compare characters with EQ [part 2] (361740b)
In package z_transform, update links and add a few more. (cd985e4)
Crosscompiling: found the reason, why wxwidgets created libraries with names different from them reported by wx-config. One must use --host and --build when configuring. Very strange. (882cf6d)
Remove an outdated (incorrect in CL) comment in src/limit.lisp (5903981)
Add rule for generating kovacicODE.pdf. (763a6a3)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (97f6468)
In continuous_freq, handle empty input correctly. (c55e791)
In discrete_freq, count data by sorting and then searching for indices. This is faster than the previous method. Also include more tests for discrete_freq. (d1e854b)
In continuous_freq, do not bother to compile count_by_bins and friends, as it generates confusing warning messages, and does not appear to make continuous_freq any faster. (3e1fa68)
In continuous_freq, rename count_em_up to count_by_bins, and have it return the bin counts instead of assigning values to an argument. (7d7ed0d)
Add tests for `print_graph' and `edge_coloring' (468eb90)
In continuous_freq, exclude the left end of the bin and include the right end. This is consistent with the default behavior of hist and truehist in R (and also consistent with the previous behavior of continuous_freq). (0907d88)
In continuous_freq, count data falling into bins by sorting the data and then finding the indices of the bin boundaries in the sorted data. It is anticipated that this method is faster than the previous. (dc8ef8d)
Fix bug 3180: declare(foo, antisymmetric) has no effect on subscripted foo (aa0a365)
Draw: Corrected the indentation of the lisp code contained in commit 14b326f543 that provides the possiblility to choose transparent lines colors if gnuplot >=5.0 is used. (1adebc1)
Test for SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" This tests commits 2247c1f and b1a14c3. (9ad3f2f)
In COMPLEX-ERF, be careful about sqrt(z^2)/z for pure imaginary z. This (along with commits 9203cc3 and 2247c1f) fix SF bug #3090: "erfi switches sign at approximately -0.476 and 0.476" (b1a14c3)
In share directories, remove generated documentation files from Git. (f6cf571)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (472c3ea)
A cross-reference from abs to cabs. wxMaxima sometimes gets bug reports that can be resolved by replacing abs by cabs and this might spread awareness that both functions exist. One example I've sent to the mailing list: (7310a17)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8dc9f50)
Enable building xmaxima in an out-of-source-build. (f73f6b5)
No longer compare characters with EQ (d62d12f)
Cut out the broken and unused macro DEF$FUN in share/affine/polyd.lisp (b9f889a)
Fix/update a couple of comments (73e2f0f)
Fix error message in BF-INVERSE-ERFC (wrong function names) (df9a5ac)
Remove unused lexical variables in the contrib package gentran (f9051b6)
Declare unused variables to be ignored in some contrib packages (e2bf248)
Fix read_binary_list lisp error when reading into an existing list (84ca942)
No longer compare numbers with EQ [part 3] (050553b)
Add test for SF bug #3170: "Error in simtran" (already fixed by commit 3e4e107). (a57ff0c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (76bcb5a)
Limits of erfc (5b5e352)
Documentation/Draw: Corrected a typo. (10b195b)
Define the variable $OLD_STDERR (instead of $OLD_STDIN) (da74d5f)
Eliminate a superfluous PROG1 (e5fb199)
Fix limit((x^(1/x) - 1)*sqrt(x), x, inf) (0fc7634)
Quote Clisp maxima image in maxima.bat (more work, so that Maxima might be installed in a user defined directory) (7b877bf)
Crosscompiling: update SBCL (8f235a2)
texinfo documentation and example images for the new transparent lines in draw. (b0522c1)
draw: Add transparency support for lines for gnuplot >5.0 the vtk interface ignores transparency when it sees it. (14b326f)
Windows: Determine the installation directory from the working directory maxima.bat was started in. This should be another step in making maxima no longer depend on hardcoded installation directory names. After this change the following things should still worked on my computer even if the installation directory has been moved: - The current wxMaxima version still finds the maxima it was installed with - The current Xmaxima version still finds the maxima it was installed with - maxima still finds its help system: "?? test" still works - maxima still is able to load packages (tested this one with load("draw"); ) (07ec0e0)
In COMPLEX-ERF, compute z^2 via (* Z Z) instead of (EXPT Z 2.0) since for some Lisps, the latter yields a small spurious imaginary part when Z is pure imaginary. In GAMMA-INCOMPLETE, squash signed zeros by adding zero. In rtest_gamma, adjust a test tolerance to ensure test passes on all Lisps. (2247c1f)
Revert commit 8f10856 in preparation for another attempt to fix COMPLEX-ERF. (9203cc3)
Changed references to unencrypted Website to https. (one get redirected to https anyway, so one could use https directly) (ebca8d5)
Resolve merge conflicts in rtest15. (3bdbb8a)
Additional tests for eigenvectors. Also, cut out an existing test, which Maxima cannot solve at present, which had gone unnoticed due to a bug in eigenvectors which was recently fixed. (8d73716)
In bromberg, print an error message if integrand evaluates to a nonnumeric value. (caa5a33)
Reformat a few lines for readability and consistency. (855d059)
Add simple test case for changevar of a summation. (d4cbc63)
Simplify the code a litte more. (b63c316)
Update changevar to support reversing the index for summation. (612dd83)
Add docstring for meqhk. (d7822b1)
Clean up tfun usage. (efe8692)
adjust tolerances for GCL 2.6.12 (c4e24ca)
Remove invfun and nfun from declare-top. (a80e88e)
Remove trans, ovar, nvar, and tfun from declare-top. (5c10158)
In eigenvectors, iterate over all eigenvalues. (3e4e107)
xmaxima: Header.tcl was included twice. (caae80f)
Add more missing files to autoconf 'make dist-gzip'. (still not finished...) (030712b)
Crosscompiling: Update wxMaxima to 16.04.2 (1818d3b)
Documentation: Got rid of an unneeded comma. (eefb9c1)
Include all loop keywords in documentation. Fixes SF bug #3157: "next" (a3c0913)
Add documentation item for %union. (09ddc2a)
In to_poly_solve documentation, correct statement about empty set. (604bd4f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (6f3d687)
Attempt to clarify description of simp. (e463d52)
Crosscompiling: update SBCL (3956e39)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (5e6df90)
Started fixing the autoconf system, so that "make dist-gzip" includes all (necessary) files. Lots are still missing (e.g. info files in other languages). Some I am unsure, if they are required (directories archive (is that current code or should it be removed (we have a version control system, so old files can be recovered) and admin, Fortran files (some are included in the tarball, but are they really needed?), etc. (e3728a9)
Additional tests for integrate: Borwein integrals (a cute sequence). (f9dae91)
Avoid trouble with ratvars when solving cubics and quartics. (b433655)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (8b92ee5)
In defsystem, do not bind Lisp verbosity flags. Otherwise the user cannot affect the verbosity of load and compile messages. (e7e444e)
Update from Richard. (d2fd9c5)
Merge branch 'master' into rtoy-polylog-eval (a28c64f)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45b9c67)
Commit chebformax.lisp (Chebyshev approximations) by Richard Fateman. (f09b737)
When parsing Lisp options, honor preceding --very-quiet if present. (b087ee9)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (cfd4fbb)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. Added test cases. (6d1b6fe)
Partially fixed assumptions regarding products and powers. We can now do assume(2*x/y^3>0) and get sign(x*y) = pos. (639dd40)
Merge branch 'master' of git://git.code.sf.net/p/maxima/code (3b37cdb)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (b6e2ce9)
Avoid unbounded recursion in PTIMES%E. Fixes SF bug #3144: "stackoverflow in integral" (4568937)
Actually add *debug-quadpack* (64a78a2)
Add rudimentary debugging for quadpack routines. (8da6033)
Update thresholds for ccl32 and ccl64 (2392025)
Merge branch 'master' into rtoy-li-eval (d831d14)
Add numeric tests for li[s](z) (45a7488)
Cleanup implementation of li-s-simp. (7c7101b)
Comment out some prints. (7d06835)
Implement the log series for li[s]. (49ce4ec)
In src/cl-info.lisp, change name of variable to accurately reflect its purpose: it is a character count, so name it CHAR-COUNT instead of BYTE-COUNT. (312ab5c)
Merge branch 'master' of ssh://git.code.sf.net/p/maxima/code (45f2b63)
In MAXIMA-DUMP, distinguish 64-bit CCL from non-64-bit. (8be65b5)
Fixed a typo. (285d013)
Merge branch 'master' into rtoy-li-eval (71a1422)
Crosscompiling: Re-enable the tex fonts (now installed in a private subdirectory below c:\maxima-..., that works even when installing as unprivileged user). Enable out-of-source builds for wxmaxima. (8294b2b)
revert change to %itopot accidentally introduced in [09d160] (e10e3e0)
Fixed bug #3109 - is(sin(x) <= 1) returns "unknown", is(sin(x) <= 1.00001) returns "true". Also added test cases. "sign(sin(x)-1)" now returns "nz" instead of "pnz", similar cases are also handled. (0bcccbe)
Return noun instead of spurious result for integrands containing signum (5a300aa)
freel second element of pair fixes mistake in commit [691cbc] (09d1603)
First cut at evaluation of li[s](z). Not working. (788782a)
Announce that ccl64 has been configure. (1eb49d1)
Relax threshold slightly for test 534 for ecl. (3a20a64)
Mark test 50 as a known failure for ccl. (c49bc5e)
Fix up dependencies to errset in generr.lisp. (73adac4)
Added CMakelists (did forget the git add in the last commit) Updated Makefile.am to include all required files (at least I hope so) (ea52114)
Crosscompiling: Allow installation as regular user (without admin privileges) Do not install jsmath tex fonts to enable this (patch for wxmaxima sumitted, that the fonts may be included in future again) Remove our own copy of NSIS.template.in and using and patching the one shipped with cmake instead. (3d6e8dd)
Add support for testing ccl64. (ff56bf7)
Add support for ccl64 (beb1aa2)
Added missing images to distribution archive. (73c832b)
draw_graph_list: make sure that wxplot_size does not contain rationals. (2a5e26a)
Crosscompiling: Use wxmaxima-16.04.1 release (1513697)
Add a little more detail to INSTALL.lisp about arguments to CONFIGURE (a0cbda6)
Update INSTALL.lisp to mention that dumping an image works on CCL (c67d9a2)
Print a message in MAXIMA-DUMP if dumping an image is not supported (930bff6)
Update decfp to compile and add tests. (c406bb6)
Describe the key for *bf-sincos-table*. (6dbc8e6)
Add Richard Fateman's decimal bigfloat package. (432712e)
Fix integrate((x*sqrt(x^3-4*x))/(x^2-4),x) Fixes bugs #3114, #3115 introduced by [647516] Fix integrate(x=0,x) (037ea4a)
Fix taylor testsuite failure Bug introduced by [22f4fd] freel check first argument (691cbc4)
5.38post: updated changelogs, updated maxima.spec to resolve a build error. (f738aff)
Set version to 5.38post (c8646e2)
|