Package Details: firefox-vaapi 132.0-2

Git Clone URL: https://aur.archlinux.org/firefox-vaapi.git (read-only, click to copy)
Package Base: firefox-vaapi
Description: Fast, Private & Safe Web Browser (with VA-API patches)
Upstream URL: https://www.mozilla.org/firefox/
Keywords: firefox libva va-api vaapi video
Licenses: MPL-2.0
Conflicts: firefox
Provides: firefox
Submitter: ewout
Maintainer: dcelasun
Last Packager: dcelasun
Votes: 11
Popularity: 0.062831
First Submitted: 2023-02-24 13:48 (UTC)
Last Updated: 2024-11-05 19:01 (UTC)

Required by (167)

Sources (7)

Latest Comments

1 2 3 4 5 6 7 Next › Last »

deemon commented on 2024-11-24 10:16 (UTC) (edited on 2024-11-24 10:24 (UTC) by deemon)

For me the build failed on Manjaro, because it complained that my nss is not >=3.105 or something (Manjaro has still 3.104) -- np, will wait until manjaro gets the correct version also sometime later. But it's not even listed as dependency and started building regardless and failed later, when the build process itself didn't have correct version.

kikislater commented on 2024-11-23 05:30 (UTC)

Thank you broller, this is what we need !

broller commented on 2024-11-23 04:45 (UTC)

@dcelasun Oh okay, I understand. Thank you for explaining, appreciate the clarifications a lot!

dcelasun commented on 2024-11-22 22:53 (UTC) (edited on 2024-11-22 22:53 (UTC) by dcelasun)

@broller: It would, but only if you are reusing the same build directory between versions, and only if you use makepkg --noextract with the second build onwards. Basically, it's something for individual users to do, not for me in the PKGBUILD. At most I can add a comment in the code.

broller commented on 2024-11-22 22:50 (UTC)

@dcelasun I understand and am fully aware the first build becomes a lot slower. Would it still not help speeding up compilation whenever this package gets an update, eg. from upstream firefox? That was my initial intent, sorry if I worded it wrong.

dcelasun commented on 2024-11-22 20:48 (UTC)

sccache is great for speeding up subsequent builds, but it actually slows down the first build. It's for Firefox developers who change a few lines of code and don't want to rebuild the entire thing.

This package follows the Arch PKGBUILD as closely as possible, so anyone building it in a clean chroot (i.e "the Arch way") would actually have worse performance. If this doesn't apply to you, try adding that option to anywhere in mozconfig, so maybe around line 145 and give it a shot. To use the cache, make sure your second build doesn't overwrite src/, so use makepkg --noextract.

broller commented on 2024-11-22 20:19 (UTC)

During my build of firefox, I got invited to check this link out: https://firefox-source-docs.mozilla.org/setup/configuring_build_options.html#sccache

ive taken a look at the PKGBUILD file but couldnt find any form of ac_add_options --with-ccache=sccache in the prepare() function that is recommended to speed incremental builds up.

Would it be possible to add it to the PKGBUILD or is it intentionally omitted?

hsantos92 commented on 2024-11-06 18:08 (UTC)

@dcelasun Thank you! That works.

dcelasun commented on 2024-11-05 19:02 (UTC)

@hsantos92 my bad, should be fixed now. It seems GfxInfo.cpp has changed since 131.x.

hsantos92 commented on 2024-11-05 18:43 (UTC) (edited on 2024-11-05 18:44 (UTC) by hsantos92)

How can I fix this?

-> Patching VAAPI

patching file widget/gtk/GfxInfo.cpp

Hunk #1 succeeded at 1094 (offset -18 lines).

-> Patching Nvidia blocklist

patching file widget/gtk/GfxInfo.cpp

Hunk #2 succeeded at 991 with fuzz 2.

Hunk #3 succeeded at 1023 (offset 8 lines).

Hunk #4 FAILED at 1065.

Hunk #5 succeeded at 1089 (offset 8 lines).

1 out of 5 hunks FAILED -- saving rejects to file widget/gtk/GfxInfo.cpp.rej

==> ERROR: A failure occurred in prepare().

Aborting...

-> error making: firefox-vaapi-exit status 4

-> Failed to install the following packages. Manual intervention is required:

firefox-vaapi - exit status 4

1 2 3 4 5 6 7 Next › Last »