Time to retire this package?

[xiota]

Chaotic AUR builders have been unable to clone mozilla-unified repo since Sep 2024.

Since Firefox supports wayland by default, there may no longer be any need to provide a wayland specific build?

If this package is not retired, it needs to be updated, including switch to mozilla-central repo.

[dr460nf1r3]

cc @thotypous

[PedroHLC]

Context: This went from "with-experimental-wayland" to a "wayland-only" build. It's interesting to exist, but the userbase is super niche.

I won't add my opinion on the answer itself, as I don't use Arch for ages.

[xiota]

Some more info: The aur/firefox-hg package also needs to be updated. If the -wayland-hg package is dropped, the -hg package could be simplified. It's currently used to build the firefox-nighty package on Chaotic AUR. Main difference from aur/firefox-nightly is that it tries to match the "actual" upstream firefox-nightly releases.

Metrics:

• firefox-wayland-hg # 4 (expected, since hasn't been rebuilt for over three months); notably, zero complaints that it hasn't been rebuilding.
• firefox-nightly # 699

Another option, retire both aur/firefox-hg and aur/firefox-wayland-hg. Revive (much simpler) firefox-nightly-hg. Or switch chaotic-aur/firefox-nightly to custom package.

[xiota]

I've dropped firefox-wayland-hg from chaotic-aur because it's unbuildable and affected by CVE-2024-9680. I also recommend retiring the AUR package because of limited usefulness. If no one argues for keeping it, will open a deletion request sometime next week.

[thotypous]

Another option, retire both aur/firefox-hg and aur/firefox-wayland-hg. Revive (much simpler) firefox-nightly-hg. Or switch chaotic-aur/firefox-nightly to custom package.

I agree with your proposal.

[Technetium1]

I don't know a good reason that this should still exist. Wayland support has been enabled by default since v121 was released on Dec 19, 2023. This makes any wayland exclusive build feel counterproductive at best. I find firefox-nightly to be satisfactory, and I don't know any current desires for firefox-nightly-hg, though I would not contest it being added.

[xiota]

Since firefox-wayland-hg is affected by aforementioned CVE, will use replaces in firefox-nightly to remove it from users computers.

Will also switch firefox-nightly to custom PKGBUILD. Previously, the switch to a different base package was to address some transient build problem and to build the "actual" upstream nightly releases. Could switch back to aur/firefox-nightly if it builds successfully and no one cares about the version mismatch. (IIRC It builds the most recent commit.)

Alternatively, could use aur/firefox-nightly-bin. (What are pros/cons of binaries vs rebuilding?)

After above is done, will open deletion requests against both aur/firefox-hg and aur/firefox-wayland-hg.

[PedroHLC]

Alternatively, could use aur/firefox-nightly-bin. (What are pros/cons of binaries vs rebuilding?)

Targeting the same libraries, headers, and compile flags the "end arch users" going to have in their system instead of Mozilla's ones.

This has rarely affected users in the last 5 years, but I recall some cases: webrtc, xdg portals and vaapi.

[xiota]

That reminds me, I have seen fairly frequent library mismatches with Firefox forks.

[PedroHLC]

That reminds me, I have seen fairly frequent library mismatches with Firefox forks.

But neither the -bin nor the native is a solution for that. It will happen less with -bin, but when it happens we can't do a thing. Meanwhile, even though you'll see it more with native, at least the fix is as easy as recompiling.