Arch Linux User Repository

Please discuss bugs here before reporting them upstream, especially if they can't be reproduced with the original tarballs. Some bugs are caused by packaging issues and need to be fixed here. We can always escalate upstream if necessary.

Make sure to fetch the developers' signing key before building the package.

The official instructions are here. The link is also mentioned in the PKGBUILD.

If the site is blocked in your region, you'll have to work around it or trust me that this is what it says:

% gpg --auto-key-locate nodefault,wkd --locate-keys torbrowser@torproject.org

which (as of 2024-06-26) should also show this fingerprint:

EF6E 286D DA85 EA2A 4BA7  DE68 4E2C 6E87 9329 8290

If your gpg says otherwise, you may have been fed garbage.

Looks line it's a known issue: https://github.com/mullvad/mullvad-browser/issues/422

Launching mullvad-browser with MOZ_ENABLE_WAYLAND=1 environmental variable solved it for me.

Launching Mullvad browser from the terminal with MOZ_ENABLE_WAYLAND=0 variable still crashes with the same errors:

Crash Annotation GraphicsCriticalError: |[0][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=2.15888) [GFX1]: Error in eglSetDamageRegion: 0x3009 Crash Annotation GraphicsCriticalError: |[0][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=2.15888) |[1][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=2.17088) [GFX1]: Error in eglSetDamageRegion: 0x3009

I ran the flatpak versions of Mullvad browser and Firefox ESR forcing x11-only and wayland-only sockets separately. I also ran the regular versions of both the browsers from terminal with MOZ_ENABLE_WAYLAND=1 to make sure they're running under wayland.None of the six scenarios generate either the error or the crash.

Does it happen with the upstream tarball? If so, there's probably not much we can do about it, except…

Mullvad flatpak works fine. So does Firefox ESR.

I'd first see if all three were using x11 or wayland (about:support → Window Protocol), and if mullvad-browser is the only one running through wayland, force it in line with MOZ_ENABLE_WAYLAND=0 /usr/bin/mullvad-browser.

The browser keeps crashing after a few minutes of usage. Launching from terminal shows this error continuously repeating:

Crash Annotation GraphicsCriticalError: |[0][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7507) |[1][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7637) [GFX1]: Error in eglSetDamageRegion: 0x3009 Crash Annotation GraphicsCriticalError: |[0][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7507) |[1][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7637) |[2][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.8937) [GFX1]: Error in eglSetDamageRegion: 0x3009 Crash Annotation GraphicsCriticalError: |[0][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7507) |[1][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7637) |[2][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.8937) |[3][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.8977) [GFX1]: Error in eglSetDamageRegion: 0x3009 Crash Annotation GraphicsCriticalError: |[0][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7507) |[1][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7637) |[2][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.8937) |[3][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.8977) |[4][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=4.0257) [GFX1]: Error in eglSetDamageRegion: 0x3009 Crash Annotation GraphicsCriticalError: |[0][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7507) |[1][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.7637) |[2][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.8937) |[3][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=3.8977) |[4][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=4.0257) |[5][GFX1]: Error in eglSetDamageRegion: 0x3009 (t=4.0267) [GFX1]: Error in eglSetDamageRegion: 0x3009

I'm using Plasma. The behavior is the same under the LTS kernel. Mullvad flatpak works fine. So does Firefox ESR.

Tested it with a fresh clone under a throwaway unix user, it builds fine.

Please try cleaning your AUR helper caches or build the package manually.

Tried multiple times, checksum validation is failing

Validating source files with sha256sums... mullvad-browser-linux-x86_64-14.0.9.tar.xz ... FAILED

Edit - Sorry, my bad, manually cleaning the cache helped. I though using clean build would delete the existing file, but apparently I should have tried manually cleaning the file. Thanks and sorry for the trouble.

Thanks, and sorry about that. Not sure how that happened, I'll add a separate git clean && makepkg step to my update routine to prevent this from happening in the future.

Should be fixed now.

sha256sum for the main tar ball was updated in the .SRCINFO, but not the PKGBUILD[1], leading to a checksum failure during build.

[1]: https://aur.archlinux.org/cgit/aur.git/commit/?h=mullvad-browser-bin&id=65506a5c4bfcb73f99772bc02b1446a7c4f825fa

Please discuss bugs here before reporting them upstream, especially if they can't be reproduced with the original tarballs. Some bugs are caused by packaging issues and need to be fixed here. We can always escalate upstream if necessary.

It might be better to replace the pipewire-pulse optional dependency with strict dependency of pulse-native-provider meta package. That package pulls in pipewire-pulse automatically according to whether the user is using pulseaudio or pipewire.