Bug#1107570: unblock (pre-approval): gnome-desktop/44.3-3
Control: tags -1 moreinfo confirmed
On 2025-06-09 23:18:27 +0100, Simon McVittie wrote:
> Package: release.debian.org
> Severity: normal
> X-Debbugs-Cc: gnome-desktop@packages.debian.org
> Control: affects -1 + src:gnome-desktop
> User: release.debian.org@packages.debian.org
> Usertags: unblock
>
> [ Reason ]
>
> New upstream bugfix release
Please go ahead and remove the moreinfo tag once the version is
available in unstable.
Cheers
>
> [ Impact ]
>
> Using U+2236 RATIO rather than U+003A COLON for the separator between
> hours, minutes and seconds can cause mis-display in RTL locales such as
> Arabic and Hebrew, which is why the GNOME HIG (human interface
> guidelines) stopped recommending it. This version goes back to using
> U+003A COLON for better RTL support.
>
> The most important upstream change (significantly improving the
> performance of gnome-control-center's backgrounds page) was already
> cherry-picked as a Debian patch; now it's using the unmodified upstream
> code.
>
> [ Tests ]
>
> A functionally equivalent package is available in experimental and seems
> to work fine on my laptop.
>
> There is some automated test coverage but it's rather basic. The
> "reftest" coverage for time formatting was removed in this update: it
> was only useful when gnome-desktop was replacing COLON with RATIO, which
> it no longer does.
>
> [ Risks ]
>
> As a core GNOME component this is a key package.
>
> All of the changes are narrowly targeted and would be easy to revert.
> I've preemptively reverted one change (swapping the default input method
> for the zh_HK locale) while asking advice from the -l10n-chinese list.
>
> [ Checklist ]
> [x] all changes are documented in the d/changelog
> [x] I reviewed all changes and I approve them
> [x] attach debdiff against the package in testing
> - this is for the package in experimental, what I'm proposing is
> an upload to unstable differing only in the changelog
>
> [ Other info ]
>
> If the -l10n-chinese team indicates that we should be following
> upstream's switch from ibus-table-cangjie5 to ibus-cangjie, that will
> most likely be a separate unblock request (dropping the patch that
> reverts it).
--
Sebastian Ramacher
Reply to: