[Development] Meeting minutes from Qt Release Team meeting 21.11.2023

[Jani Heikkinen]

Qt 6.6 status:

  *   Qt 6.6.1 release content is frozen and package creation is ongoing
  *   Qt 6.6.1 will be released when package testing is done
     *   Most probably this will happen at the beginning of next week

Qt 6.7 status:

  *   Platform and module freeze -milestone will be in effect Friday 24th of November
  *   The target is to update Qt 6.7.0 snapshot to the production later this week


Next meeting Tue 5th of December 16:00 CET

br,
Jani Heikkinen
Release Manager

irc log below:
[17:00:12] <+jaheikki3> ablasche: akseli: carewolf: lars_: mapaaso: The-Compiler: thiago: vohi: ping
[17:00:53] <vohi> jaheikki3: pong
[17:01:11] <+jaheikki3> time to start qt release team meeting
[17:01:17] <+jaheikki3> On agenda today:
[17:01:20] <+jaheikki3> Qt 6.6 status
[17:01:26] <+jaheikki3> Qt 6.7 status
[17:01:33] <thiago> jaheikki3: pong
[17:01:35] <+jaheikki3> Any additional item to the agenda?
[17:02:40] <+jaheikki3> Let's start from Qt 6.6 status
[17:03:03] <+jaheikki3> Branching from '6.6' to '6.6.1' done
[17:03:23] <carewolf> pong
[17:03:32] <+jaheikki3> All Qt 6.6.1 release blockers are fixed and Qt 6.6.1 release content is frozen
[17:03:44] <+jaheikki3> Package creation is ongoing
[17:03:46] <thiago> there's the discussion on IPC
[17:04:01] <+jaheikki3> thiago: where?
[17:04:29] <thiago> https://codereview.qt-project.org/c/qt/qtbase/+/518039
[17:05:20] <vohi> thiago: is 6.6 worse than any release before it was? My interpretation of your email is that things are as bad as before right now.
[17:05:30] <thiago> yes
[17:05:36] <thiago> it was bad, I made worse
[17:06:57] <vohi> ok, so the unavoidable memory leak in 6.6, also after the recent fixes following the refactoring, are new in 6.6
[17:07:01] <+jaheikki3> But is this really critical enough to delay Qt 6.6.1 release? Qt 6.6.2 will be released at the beginning of next year
[17:07:06] <vohi> and not fixable for 6.6.1
[17:08:09] <thiago> 6.6.0 switched the POSIX backend, which leaks memory 100% of the tmie
[17:08:16] <thiago> the SysV one at least tried to delete
[17:08:19] <thiago> it would be status quo
[17:08:29] <thiago> the switch only happens for the new API
[17:08:42] <thiago> so 6.6.0 is telling people to switch from the old (status quo) to the new, worse one
[17:08:50] <thiago> I want people *not* to port
[17:10:25] <+jaheikki3> vohi: ?
[17:10:45] <vohi> thiago: ok; we could make the change that deprecates those APIs for 6.7, and we can publish the online documentation with patches on top of 6.6.1 without having to reopen the package. Local documentation would not include those changes of course. But maybe a blog post or otherwise head-up about the situation would anyway be a good idea; people won't see the doc changes unless they are actively looking
[17:11:25] <thiago> I think 6.6.1 needs a change that un-deprecates the old API
[17:11:31] <thiago> that's the bare minimum
[17:13:08] <+jaheikki3> thiago: do we have that change in place already?
[17:13:09] <vohi> those deprecations are from 6.10 on for most people, pointing at "Native IPC Key" documentation; can we update that documentation, and remove the deprecations for 6.6.2?
[17:13:12] <thiago> no
[17:13:27] <thiago> the change I have undeprecates the old API but deprecates the entire class
[17:13:42] <thiago> we could split it in two
[17:14:06] <thiago> youŕ e misunderstanding
[17:14:12] <thiago> ah, wait, no, I am
[17:14:18] <thiago> those are 6.10 deprecations
[17:15:00] <thiago> but people who enable deprecated_since=6.0 will get it and will begin porting to worse code
[17:15:07] <thiago> I still suggest we remove the deprecations, completely
[17:16:25] <vohi> I agree that we should remove the deprecations in 6.6.2 at the latest. If we have that patch in 6.6, then we can generate the online documentation with that change. The question is then if we need to stop the 6.6.1 show for that patch. deprecated_since=6.0 is an opt-in, right?
[17:17:12] <thiago> it is
[17:18:01] <vohi> Ok; I think with updated online documentation and a blog and/or mail to interest@ telling people to ignore the deprecation warning and not port to the new API will be acceptable then
[17:18:11] <thiago> ok
[17:18:21] <thiago> we can discuss further next week too
[17:18:43] <thiago> but 6.6.1 doesn't need to wait for that
[17:18:59] <vohi> 👍
[17:19:01] <+jaheikki3> Great!
[17:19:50] <+jaheikki3> Who does the online documentation change in 6.6? I'll make then sure that it will be taken in the documentation update for Qt 6.6.1
[17:20:35] <+jaheikki3> and what about that blog/interest ml?
[17:20:53] <thiago> I can do the interest ML
[17:21:12] <+jaheikki3> thanks
[17:21:15] <thiago> but both I'd like to suggest we discuss what the long-term will be, frst
[17:21:23] <thiago> so, next Friday at the earliest
[17:22:48] <+jaheikki3> That's OK. It just means we don't have those ready with Qt 6.6.1 release but I think that's OK; we can update the online documentation whenever we have consensus & needed changes in place
[17:23:02] <vohi> I'll ping Topi about including the patch when generating the online docs (or regenerting the online docs when the patch is merged)
[17:24:08] <+jaheikki3> Great! So we will release Qt 6.6.1 from the current content when packages are tested and ready to be released
[17:24:25] <+jaheikki3> Most probably that will happen at the beginning of next week
[17:25:03] <+jaheikki3> that's all about qt 6.6.1 at this time. Any comments or questions?
[17:26:00] <+jaheikki3> Ok, then quick Qt 6.7 status update
[17:26:23] <+jaheikki3> Platform and module freeze -milestone will be in effect this Friday
[17:27:07] <+jaheikki3> Based on my understanding most important changes are already in
[17:27:25] <+jaheikki3> It seems there won't be new modules in Qt 6.7
[17:28:35] <+jaheikki3> Dependency updates has been working pretty well recently and the target is to update Qt 6.7.0 snapshot to the production later this week
[17:29:11] <+jaheikki3> That's actually all about Qt 6.7 status at this time. Any comments or questions?
[17:31:15] <+jaheikki3> Ok, let's then end this meeting now and have new one Tue 5th December at this same time; we can skip next weeks meeting & see in QtCS
[17:31:35] <+jaheikki3> Thanks for your participation, bye!
[17:31:41] <+jaheikki3> See you in QtCS
[17:32:12] <vohi> thanks, see you next week! :)
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.qt-project.org/pipermail/development/attachments/20231122/2e5af361/attachment-0001.htm>