KDE Frameworks 5.94.0

Sat May 21 22:47:54 BST 2022

El dissabte, 21 de maig de 2022, a les 12:02:29 (CEST), Ben Cooksley va escriure:
> On Sat, May 21, 2022 at 8:38 PM Albert Astals Cid <aacid at kde.org> wrote:
> 
> > El dimarts, 17 de maig de 2022, a les 8:53:16 (CEST), Ben Cooksley va
> > escriure:
> > > On Tue, May 17, 2022 at 10:10 AM Albert Astals Cid <aacid at kde.org>
> > wrote:
> > >
> > > > El dilluns, 16 de maig de 2022, a les 11:50:13 (CEST), Ben Cooksley va
> > > > escriure:
> > > > > On Mon, May 16, 2022 at 9:39 AM Albert Astals Cid <aacid at kde.org>
> > wrote:
> > > > >
> > > > > > El diumenge, 15 de maig de 2022, a les 23:03:27 (CEST), David
> > Faure va
> > > > > > escriure:
> > > > > > > On samedi 14 mai 2022 20:48:55 CEST Albert Astals Cid wrote:
> > > > > > > > El diumenge, 8 de maig de 2022, a les 0:11:35 (CEST), David
> > Faure
> > > > va
> > > > > > escriure:
> > > > > > > > > On samedi 7 mai 2022 18:38:44 CEST Antonio Rojas wrote:
> > > > > > > > > > El sábado, 7 de mayo de 2022 15:31:17 (CEST), David Faure
> > > > escribió:
> > > > > > > > > > > Dear packagers,
> > > > > > > > > > >
> > > > > > > > > > > KDE Frameworks 5.94.0 has been uploaded to the usual
> > place.
> > > > > > > > > > >
> > > > > > > > > > > New frameworks: none this time.
> > > > > > > > > >
> > > > > > > > > > This should have been a reply to this mail of course
> > > > > > > > > >
> > > > > > > > > > >  tr tt ug uk and uz translations are missing from all
> > > > tarballs.
> > > > > > > > >
> > > > > > > > > Thanks for noticing. My "update l10n" script did show an
> > error
> > > > about
> > > > > > that
> > > > > > > > > but I didn't see that error and moved on. I have now improved
> > > > things
> > > > > > so I
> > > > > > > > > can't launch the next step if this step failed.
> > > > > > > > >
> > > > > > > > > I now uploaded new tarballs (also including
> > > > > > > > > https://invent.kde.org/frameworks/
> > > > > > plasma-framework/-/merge_requests/523)
> > > > > > > > >
> > > > > > > > > I hope it's all fine now, let me know otherwise.
> > > > > > > >
> > > > > > > > I've just realized that the packages dependencies have not been
> > > > > > increased to
> > > > > > > > 5.94 and they are just at 5.93. I guess that's not that bad,
> > but
> > > > just
> > > > > > want
> > > > > > > > to make sure that doesn't mean some other change may be also
> > > > missing?
> > > > > > >
> > > > > > > Good catch. Fortunately nothing else should be missing. The
> > reason
> > > > this
> > > > > > didn't happen is that I run
> > > > > > >
> > > > > > > cd ../src && ./kdesrc-build --src-only && cd ../release-tools &&
> > > > > > ./list_frameworks.sh && ./update_l10n.sh &&
> > > > > > ./increase_frameworks_version.sh step2
> > > > > > >
> > > > > > > and update_l10n failed because anonsvn threw me out in the
> > middle of
> > > > it.
> > > > > > > I didn't notice, ran the next step (./make_rc_tag.sh) and this is
> > > > how we
> > > > > > got tags with missing languages.
> > > > > > > Then I fixed l10n and tagged again, completely missing that
> > > > > > "increase_frameworks_version.sh step2" was never done.
> > > > > > >
> > > > > > > Sigh, I wish svn.kde.org would work for the way we're using it
> > at
> > > > > > release time.
> > > > > >
> > > > > > One thing we may try is moving Frameworks to the "po injection"
> > test
> > > > group.
> > > > > >
> > > > > > "po injection" = scripty commits po files back to git
> > > > > >
> > > > > > Plasma Mobile Gear has been using it and as far as I understand it
> > is
> > > > > > working relatively well.
> > > > > >
> > > > > > Example:
> > > > > > https://invent.kde.org/plasma-mobile/alligator/-/commits/master/po
> > > > > >
> > > > > > This way you can remove any special handling of l10n altogether
> > from
> > > > the
> > > > > > KDE Frameworks release scripts
> > > > > >
> > > > >
> > > > > Is there any particular reason why we can't rollout po injection to
> > all
> > > > > repositories?
> > > >
> > > > We need to adapt the release scripts not to fetch l10n, we also have
> > > > various release scripts that inject ki18n_install/kdoctools_install
> > when
> > > > they do the fecth, so we need to properly add those calls to the root
> > > > CMakeLists.txt
> > >
> > >
> > > > Not a huge blocker, but it needs to be decided that we should do it and
> > > > then do it.
> > > >
> > >
> > > Given the number of benefits it has to release managers (no more tooling
> > > needed to download translations and integrate them), as well as build
> > > processes for nightly builds and for builds on developer systems I think
> > > we've a fairly compelling reason to switch it on globally.
> >
> > I am not disagreeing and that's always been the plan, but someone needs to
> > do the work for that to be possible :)
> >
> > I can only do so much, so help always welcome :)
> >
> 
> What sort of work is required here?

Copy and pasting from two messages ago in the thread

"
We need to adapt the release scripts not to fetch l10n, we also have various release scripts that inject ki18n_install/kdoctools_install when they do the fecth, so we need to properly add those calls to the root CMakeLists.txt
"
of each repository that has them missing <-- this part was missing from ↑

> 
> My understanding is everything on the scripty/infrastructure side is in
> place?

Yes, that seems to be functional.

Cheers,
  Albert

> 
> 
> >
> > Cheers,
> >   Albert
> >
> 
> Thanks,
> Ben
> 
> 
> >
> > > It will also
> > > ensure good CI coverage of the CMake files which should stop issues
> > around
> > > ki18n_install macros which have happened a few times.
> > >
> > > It'll also reduce the load on Leptone (the server for invent.kde.org and
> > > svn.kde.org) at release time.
> > >
> > > Cheers,
> > > Ben
> > >
> > >
> > > >
> > > > Cheers,
> > > >   Albert
> > > >
> > > > >
> > > > >
> > > > > >
> > > > > > Cheers,
> > > > > >   Albert
> > > > > >
> > > > >
> > > > > Cheers,
> > > > > Ben
> > > > >
> > > > >
> > > > > >
> > > > > > >
> > > > > > > I have now added more error checking, including at the beginning
> > of
> > > > > > make_rc_tag.sh,
> > > > > > > as well as a reminder to run "step2" after fixing the l10n
> > problems.
> > > > > > >
> > > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > > >
> > > > >
> > > >
> > > >
> > > >
> > > >
> > > >
> > >
> >
> >
> >
> >
> >
> 