Description of problem: Gnome Software has had since a while ago a functionality on Fedora where it will automatically install the langpack for the system locale, but due to a recent bug for certain regional locales it will install the generic language langpacks instead. Current gnome-software 42.x (and IIRC 41.x on Fedora 35 too) will install the `langpacks-pt` package on pt_BR systems instead of the appropriate `langpacks-pt_BR` package. Not only this will install the wrong locale for those systems but, according to the upstream issue, for Chinese systems this will try to install `langpacks-zh` which does not exist. The fix was merged and will be available on Gnome Software 42.2, please backport it to Fedora's package.
Thanks for a bug report. > The fix was merged and will be available on Gnome Software 42.2, please backport it to Fedora's package. The 42.2 will be released in two weeks. Is this really in hurry?
(In reply to Milan Crha from comment #1) > Thanks for a bug report. > > > The fix was merged and will be available on Gnome Software 42.2, please backport it to Fedora's package. > > The 42.2 will be released in two weeks. Is this really in hurry? Actually, that's a good point. It might not be something urgent, but it's probably a bit of an annoyance in terms of I10n. Users of the affected locales will either get the langpacks for the generic language locale instead of the one for their region, which might not install their region-specific i10n packages via weak deps (this can be checked by running `dnf repoquery --whatsupplements langpacks-pt_BR` for example, do note that pt_BR is affected by bug 2084575 right now), so this might mean wrong dictionaries, hyphenation rules and thesaurus data due to wrong hunspell-*, hyphen-* and mythes-* packages being installed among other things, or no langpacks being installed at all in the specific case of Chinese locales (no langpacks-zh available). It's more of whether this internationalization-related bug can wait or should be fixed earlier. Of course, the correct langpacks can be installed manually by the user if they know how or want (probably even from Gnome Software if they look into it), but this is more about what GS does by default.
I'm not sure what I'm asked for here. My opinion is to wait. It's not about backporting upstream patch, it's simple and super fast, but getting the change to the users is the problem. The update can wait in the updates-testing for two weeks, then there will be the official upstream release and the users will install a new gnome-software, with more fixes. The official upstream release should get to the users earlier than in two weeks, because it'll be part of the larger GNOME update, which usually means more attention and more early testing by more users. That being said, if you think it'll help to some/many users, I can do the update, it's really trivial due to the existing upstream fix, I'm only the gain will be minimal due to the reasons above.
(In reply to Milan Crha from comment #3) > I'm not sure what I'm asked for here. My opinion is to wait. It's not about > backporting upstream patch, it's simple and super fast, but getting the > change to the users is the problem. The update can wait in the > updates-testing for two weeks, then there will be the official upstream > release and the users will install a new gnome-software, with more fixes. > The official upstream release should get to the users earlier than in two > weeks, because it'll be part of the larger GNOME update, which usually means > more attention and more early testing by more users. > > That being said, if you think it'll help to some/many users, I can do the > update, it's really trivial due to the existing upstream fix, I'm only the > gain will be minimal due to the reasons above. Yeah, the idea was to push the fix earlier. But since, as you mentioned, the time it hits stable will likely be around the same time the official release, then there would be little benefit. The bug is also very minor anyway, just a small annoyance, users who figured out what was wrong could already fix it by themselves. Feel free to close.
Okay, let's close this then. Thanks.