Package Details: gcc47 4.7.4-2

Git Clone URL: https://aur.archlinux.org/gcc47.git (read-only, click to copy)
Package Base: gcc47
Description: The GNU Compiler Collection (4.7.x)
Upstream URL: http://gcc.gnu.org
Licenses: GPL, custom, LGPL, FDL
Submitter: mosra
Maintainer: severach
Last Packager: severach
Votes: 21
Popularity: 0.000000
First Submitted: 2013-06-06 16:22 (UTC)
Last Updated: 2021-09-28 19:22 (UTC)

Latest Comments

« First ‹ Previous 1 2 3 4 5 Next › Last »

Pilsbert commented on 2016-01-10 14:35 (UTC)

It seems that the problem is that ppl, cloog, and LLVM (i.e., gold) are compiled using the system compiler (GCC 5.3) and hence the system libstdc++ is used. However, when now bulding GCC 4.7 this will run into an issue due to the different (ABI?) versions (I get errors where GLIBCXX_3.4.20 and GLIBCXX_3.4.21 cannot be found). For me, the following works with GCC 5.3: * Remove 'libmpc' 'cloog' 'ppl' from depends and makedepends. * To automatically download the dependencies MPFR, GMP, and MPC (as described in the GCC manual) change build() { cd ${srcdir}/${_basedir} to: build() { cd ${srcdir}/${_basedir} ./contrib/download_prerequisites * Remove the configure flags: --with-ppl --enable-cloog-backend=isl --disable-ppl-version-check --disable-cloog-version-check --enable-gold --enable-ld=default --enable-plugin --with-plugin-ld=ld.gold * The patch suggested by rubenvb is NOT necessary here! That's all. As I understand it, removing the above configure flags will remove support to use the gold linker. However, for me this is no problem as I only need GCC 4.7 to compile Matlab MEX files. Hence, the remaining problem will be that now GCC 4.7 won't use some memory optimizations, which require cloog/ppl. I guess to reenable that we have to compile ppl and cloog with the just compiled GCC 4.7 (and not with GCC 5.3!) and than recompile the GCC 4.7 with those libs. However, I'm not familiar with that so maybe someone else want to add this feature :)

severach commented on 2016-01-05 02:27 (UTC)

checking for C compiler default output file name... configure: error: in `/home/chris/build/gcc47/src/gcc-build/x86_64-unknown-linux-gnu/libmudflap': configure: error: C compiler cannot create executables See `config.log' for more details. Makefile:11502: recipe for target 'configure-target-libmudflap' failed

nexero commented on 2015-11-17 07:08 (UTC)

@rubenvb Thanks, without this patch this package wouldn't compile..

rubenvb commented on 2015-09-29 12:39 (UTC)

This patch is needed to compile with GCC 5: diff --git a/gcc/cp/cfns.h b/gcc/cp/cfns.h index 62cdfab..4f63cc4 100644 --- a/gcc/cp/cfns.h +++ b/gcc/cp/cfns.h @@ -53,6 +53,9 @@ __inline static unsigned int hash (const char *, unsigned int); #ifdef __GNUC__ __inline +#ifdef __GNUC_STDC_INLINE__ +__attribute__ ((__gnu_inline__)) +#endif #endif const char * libc_name_p (const char *, unsigned int); /* maximum key range = 391, duplicates = 0 */ @@ -96,7 +99,7 @@ hash (register const char *str, register unsigned int len) 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400, 400 }; - register int hval = len; + register int hval = (int)len; switch (hval) { (remember to put a newline at the end of the patch file)

petric789 commented on 2015-07-10 07:46 (UTC)

mosra: My apologies, error was due to my MAKEFLAGS tweaks. With the default MAKEFLAGS it builds flawlessly. Although the workaround from the gcc48 comments is needed. Thanks! I will remove my other comment, so others don't get confused.

mosra commented on 2015-07-09 20:51 (UTC)

huh. now please tell me what *I* am doing wrong. I tried to build the package four times in last six months, always completely from scratch and it always built successfully. I'm on latest [extra], nothing from [testing], x86_64, gcc 5.1 and clang 3.6.1 installed. Could MAKEFLAGS etc. affect this?

mosra commented on 2015-07-09 08:57 (UTC)

petric789: can you try the workarounds mentioned in the gcc48 package? the problems seem to be similar: https://aur.archlinux.org/packages/gcc48/

petric789 commented on 2015-07-09 08:46 (UTC)

It still fails to build at this point. I have the exact same problem as simgunz.

« First ‹ Previous 1 2 3 4 5 Next › Last »