AUR (en) - rav1e-git

Git Clone URL:	https://aur.archlinux.org/rav1e-git.git (read-only, click to copy)
Package Base:	rav1e-git
Description:	The fastest and safest AV1 encoder
Upstream URL:	https://github.com/xiph/rav1e
Licenses:	BSD-2-Clause, LicenseRef-custom
Conflicts:	rav1e
Provides:	librav1e.so, rav1e
Submitter:	Chocobo1
Maintainer:	Chocobo1
Last Packager:	Chocobo1
Votes:	7
Popularity:	0.000000
First Submitted:	2018-02-25 07:31 (UTC)
Last Updated:	2024-08-02 07:05 (UTC)

Latest Comments

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

Chocobo1 commented on 2020-01-23 03:54 (UTC) (edited on 2020-01-23 04:03 (UTC) by Chocobo1)

/usr/.crates2.json should be removed from the package.

I see no such file on my builds, where does it come from?

gmes78 commented on 2020-01-23 00:29 (UTC)

/usr/.crates2.json should be removed from the package. I also suggest passing the -f flag to rm, so that if /usr/.crates2.json (or /usr/.crates.toml) isn't there, the build doesn't fail.

gmes78 commented on 2019-11-22 18:21 (UTC)

OK, updated although I haven't really compared the run time differences. A quick glance reveals that cargo-c is spending most of its time waiting for rustc to complete so I guess time difference won't be significant.

I made a comparison just now:

$ rm -rf src/ pkg/ && makepkg -o --noprepare && time makepkg --noarchive
makepkg --noarchive  906,78s user 14,30s system 215% cpu 7:07,42 total
$ git reset --hard PKGBUILD && git pull
$ rm -rf src/ pkg/ && makepkg -o --noprepare && time makepkg --noarchive
makepkg --noarchive  630,62s user 14,11s system 177% cpu 6:03,62 total

The difference between the former (cargo-c with optimizations) and the latter (cargo-c without optimizations) is that the latter is 64 seconds faster. Which is the exact same difference as my other comparison.

In other words, disabling optimizations for cargo-c doesn't impact its runtime speed in any noticeable way.

Chocobo1 commented on 2019-11-22 03:24 (UTC)

May I suggest building cargo-c in debug mode?

OK, updated although I haven't really compared the run time differences. A quick glance reveals that cargo-c is spending most of its time waiting for rustc to complete so I guess time difference won't be significant.

gmes78 commented on 2019-11-21 19:00 (UTC)

May I suggest building cargo-c in debug mode? It cuts off a lot of compile time (and we don't need a high level of optimization for a build tool):

$ time cargo install --root /tmp/cinstall-release cargo-c
cargo install --root /tmp/cinstall-release cargo-c  384,17s user 6,11s system 354% cpu 1:50,10 total

$ time cargo install --debug --root /tmp/cinstall-debug cargo-c
cargo install --debug --root /tmp/cinstall-debug cargo-c  127,79s user 5,57s system 288% cpu 46,145 total