15:04:56 <rdieter> #startmeeting kde-sig
15:04:56 <zodbot> Meeting started Tue Dec 15 15:04:56 2015 UTC.  The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:04:56 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:04:56 <zodbot> The meeting name has been set to 'kde-sig'
15:04:57 <tosky> :)
15:04:59 <rdieter> #meetingname kde-sig
15:04:59 <zodbot> The meeting name has been set to 'kde-sig'
15:05:03 <rdieter> #topic roll call
15:05:12 <rdieter> hi all, friendly kde-sig meeting, who's present today?
15:05:12 <tosky> hi
15:05:22 * jgrulich is present
15:05:49 <than_> present
15:05:58 <heliocastro> present
15:07:10 <rdieter> #info rdieter tosky jgrulich than_ heliocastro present
15:07:17 <rdieter> #chair tosky jgrulich than_ heliocastro
15:07:17 <zodbot> Current chairs: heliocastro jgrulich rdieter than_ tosky
15:08:10 <dvratil> hi all
15:08:35 <rdieter> #info dvratil present
15:08:38 <rdieter> #chair dvratil
15:08:38 <zodbot> Current chairs: dvratil heliocastro jgrulich rdieter than_ tosky
15:08:39 <jreznik_> hi
15:08:44 <rdieter> #info jreznik_ present
15:08:50 <rdieter> #chair jreznik
15:08:50 <zodbot> Current chairs: dvratil heliocastro jgrulich jreznik rdieter than_ tosky
15:08:55 <rdieter> #topic agenda
15:09:00 <rdieter> ok, what to discuss today?
15:09:24 <rdieter> status updates for kf5-5.17/plasma-5.5 (dvratil,rdieter)
15:10:01 <rdieter> kf5 kdepim in rawhide (resolving conflicts, bug #1290143)
15:10:13 <rdieter> anything else?
15:11:12 <rdieter> heliocastro: worth giving update on qt-5.6/moc situation?
15:11:24 <heliocastro> Don't think so
15:12:05 <jreznik> heliocastro: any updates on qt 5.6 and multimonitor issue?
15:12:16 <heliocastro> Nope, didn't touch on that
15:12:23 <heliocastro> Ok, maybe a quick about moc
15:12:40 <rdieter> ok, maybe we can have a general "qt-5.6 status" topic
15:12:46 <rdieter> anything else to discuss?
15:12:47 <heliocastro> yes
15:13:28 <heliocastro> * yes for the general 5.6 topic
15:13:44 <rdieter> ok, let's get started
15:13:50 <rdieter> #topic kf5-5.17/plasma-5.5
15:14:12 <dvratil> 5.17 is in rawhide and Bodhi for F22 and F23
15:14:28 <rdieter> dvratil++
15:14:30 <rdieter> hero
15:14:36 <dvratil> I have temporarily patched out the #ifdefs in frameworkintegration that were breaking build due to the moc bug
15:15:01 <dvratil> for Plasma 5.5 I imported an upstream patch for the taskbar issue
15:15:20 <dvratil> and thanks to rdieter pulling in all the other upstream patches things should now be pretty solid
15:16:35 <dvratil> some karma for the pending updates would be nice
15:16:40 <dvratil> <EOR>
15:17:13 <rdieter> on kf5-5.17 gotcha, a few apps actually had bugs that previous versions of kf5 covered up, wrt application.setQuitOnLastWindowClosed(false); calls
15:17:45 <rdieter> this hit (at least): ark and kate
15:18:09 <rdieter> I put patched kate buids into the kf5-5.17 update
15:18:48 <rdieter> for ark, I was initially trying to update to ark-15.12.0, but then hit trouble with ark-unar-06.patch , it needs a bit of porting/fixing
15:19:14 <rdieter> anyone able/willing to help there (in ark master/ branch) would be lovely
15:19:37 <rdieter> else, I'll resort to backporting fix to ark-15.08.3
15:20:33 <rdieter> secondary topic:  looks like upstream is planning on releasing minor plasma releases... weekly (fast!)
15:20:37 <rdieter> and 5.5.1 came out today
15:20:48 <tosky> I think by-weekly, but still
15:20:56 <rdieter> we probably won't be able to track all of those releases for anything except maybe rawhide
15:21:09 <rdieter> tosky: oh ok, that's a little better
15:21:18 <rdieter> so just, fyi
15:21:35 <tosky> I stand corrected: weeky for 5.5.1 and 5.5.2, then more relaxed:
15:21:41 <tosky> https://techbase.kde.org/Schedules/Plasma_5
15:22:19 <rdieter> any other kf5/plasma5 comments ?
15:22:25 <tosky> for ark, I think that simply contacting upstream would help
15:22:36 <tosky> now that there are active maintainers again
15:23:04 <rdieter> tosky: about unar?  there's an old/inactive review request for that, fyi, but it's been stale for ~2 years :(
15:23:26 <tosky> rdieter: maybe it was not seen by the new maintainers
15:23:42 <tosky> what is the big advantage? The usage of unar for unrar v3?
15:23:55 <rdieter> unar is free, unrar is not
15:24:15 <rdieter> ie, nice to be able to ship ark that works with free implementation
15:24:19 <tosky> yes, I see
15:25:27 <rdieter> anyway, if no one steps up to help, I'll go with the backporting plan in the meantime
15:25:46 <rdieter> (no one steps up asap, today'ish that is)
15:26:03 <dvratil> regarding the Plasma update - I propose we do 5.5.1 update (I can do it later this week), I doubt anyone will be able to get 5.5.2 in if it's released just before Christmas
15:26:04 <rdieter> moving on...
15:26:21 <rdieter> dvratil: good plan, +1
15:26:43 <rdieter> I can help too, maybe work to relax BuildDeps so we can build most of it in parallel too
15:26:52 <rdieter> (if that's not already possible)
15:27:01 <dvratil> that should be checked by the script on update
15:27:11 <rdieter> ie, BuildRequiers: plasma-foo >= 5.5 , intead of BR: plasma-foo >= %{version}
15:27:43 <dvratil> ah, well,  I think we have only 5 build groups now, we just quite good
15:27:54 <rdieter> unless upstream really enforces that in CMakeLists.txt
15:27:54 <dvratil> better than the dozen or so for frameworks :)
15:28:30 <rdieter> dvratil: so you'd rather not relax the build deps
15:28:31 <rdieter> ?
15:28:54 <dvratil> if it does not make my automation checks harder, I'm OK with it
15:29:02 <rdieter> (relaxing will help others too, like seconary archs)
15:29:03 <dvratil> but let's solve that another time
15:29:08 <rdieter> ok,
15:29:18 <rdieter> moving on...
15:29:32 <rdieter> #topic kf5 kdepim in rawhide (resolving conflicts, bug #1290143)
15:29:37 <rdieter> .bug 1290143
15:29:42 <zodbot> rdieter: Bug 1290143 kdepim dependency issue - https://bugzilla.redhat.com/1290143
15:29:54 <rdieter> the conflicts in ^^ (I think) are currently causing live image compose failures
15:30:28 <rdieter> I think we can remove all (most!?) of the Obsoletes/Conflicts: akonadi/kdepimlibs
15:30:48 <rdieter> now that we've made (qt4) akonadi parallel-installable with kf5-akonadi
15:31:13 <rdieter> (it is libraries-only now, it doesn't actually work, just a compat hack to be able build stuff that depends on it)
15:31:25 <dvratil> +1, the libs should be coinstallable
15:31:34 <jgrulich> rdieter: all kf5-pim stuff has now removed all obsoletes/conflicts with kdepimlibs
15:31:41 <jgrulich> missing kdepim and kdepim-runtime
15:31:48 <rdieter> jgrulich: and akonadi ?
15:32:11 <rdieter> I guess most of the conflicts was indeed just kdepimlibs
15:32:16 <jgrulich> kf5-akonadi, kf5-akonadi-calendar and kf5-akonadi-search
15:32:27 <rdieter> jgrulich: thanks
15:32:43 <rdieter> that was fast!
15:32:51 <rdieter> I guess nothing else to discuss then :)
15:32:54 <rdieter> moving on...
15:33:04 <rdieter> #topic Qt 5.6.0 status
15:33:14 <heliocastro> ok
15:33:30 <rdieter> .bug 1291040
15:33:30 <heliocastro> So, i really believe the moc issue is ours
15:33:31 <zodbot> rdieter: Bug 1291040 qt-5.6.0 tracker - https://bugzilla.redhat.com/1291040
15:33:38 <rdieter> fyi, bug tracking qt-5.6 related issues
15:34:41 <heliocastro> qt 5.6 in general is fine, even with gold linker
15:35:08 <heliocastro> moc is including all subdirectories * should not
15:35:37 <heliocastro> Reaching qtglobal.h and then creating invalid QT_VERSION
15:36:02 <rdieter> how is that not an upstream issue ?
15:36:05 <heliocastro> Qt trolls say moc sohuld follow directory chain
15:36:19 <heliocastro> Which make me more confuse
15:36:35 <heliocastro> So i need now other eyes on this issue to help
15:37:02 <rdieter> possibly because our qconfig.h is custom ?
15:37:13 <heliocastro> Big change
15:37:17 <heliocastro> ĉhance
15:37:20 <heliocastro> arghh
15:37:28 <rdieter> we patch it to avoid multilib conflicts
15:37:31 <rdieter> http://paste.fedoraproject.org/301142/93838145
15:38:04 <rdieter> does moc try to parse qconfig.h directly or something ?
15:38:13 <heliocastro> The fact is that some include is killing the VERSION variable
15:38:20 <heliocastro> moc read the includes
15:38:33 <heliocastro> if includes point to qconfig.h, then moc reads it
15:38:48 <dvratil> hmm, I guess MOC chokes on the  __WORDSIZE
15:38:53 <dvratil> check that we added
15:39:29 <rdieter> we've been doing it that way for a *long* time, fwiw
15:39:42 <rdieter> (same hack is used in qt(4) packaging)
15:39:51 <heliocastro> dvratil: Nope, otherwise just remove the QT_VERSION macro would not work
15:40:26 <rdieter> anyway, I can test omitting the qconfig.h multilib hack to see if that helps or not after meeting
15:41:01 <dvratil> I have to run now, sorry guys
15:41:04 <rdieter> dvratil: thanks
15:41:07 <heliocastro> see you
15:41:23 <heliocastro> But from one side is a relief that is not a build issue
15:41:37 <heliocastro> And is really a config/system issue
15:41:48 <rdieter> other qt-5.6.0 issue may or may not be related to moc too, is that some apps now FTBFS due to Q_IMPORT_PLUGIN use and now getting undefined symbols when linking
15:42:12 <rdieter> apps affected include: qupzilla, trojita
15:42:55 <RaphGro> trojita upstream builds from git with Qt5.6, so I expect they do not suffer from that.
15:43:08 <jgrulich> I also need to leave → boring christmas shopping
15:43:23 <heliocastro> rdieter: I beleieve is moc issue as well
15:43:32 <rdieter> RaphGro: ok, I'll include those when I test the moc after meeting
15:43:46 <RaphGro> rdieter, include what?
15:43:47 <rdieter> heliocastro: question now is: what to do about it?  Any ideas yet?
15:44:19 <rdieter> RaphGro: I'm going to revert our qconfig.h hacks, to see if that helps fix moc related issues.  so part of my test will include testing qupzilla/trojita builds too
15:44:40 <RaphGro> rdieter, ok.
15:45:08 <heliocastro> rdieter: I will take another look tonight
15:45:28 <rdieter> heliocastro: giant hammer would be to move -headerdir %{_includedir}/qt5 to -headerdir %_libdir/qt5/include
15:45:39 <rdieter> but we can't realistically do that in any stable release
15:45:56 <heliocastro> rdieter: No, this would be hackish and we would be putting everything under carpet
15:46:01 <rdieter> then we wouldn't have to worry about multilib conflicts
15:46:24 <rdieter> we'd only be putting headers in a multilib-safe location
15:46:44 <heliocastro> But doesn't matter where tehy are
15:46:53 <heliocastro> still need to be included in some way
15:47:00 <rdieter> if our qconfig.h hacks are the cause, then yes it matters
15:47:21 * heliocastro trusts rdieter
15:47:24 <rdieter> if we moved them, then we would no longer need the qconfig.h hacks, and stuff would 'just work'
15:47:59 <rdieter> otherwise, we need to find a multilib-safe qconfig.h what also works with moc
15:48:07 <rdieter> s/what/that/
15:48:16 <heliocastro> *if* they are the cause
15:48:20 <rdieter> true
15:48:30 <rdieter> anything else on qt-5.6 ?
15:48:46 <heliocastro> jreznik mentioned about 3 screens +
15:48:56 <heliocastro> But is not 5.6 at all
15:49:00 <heliocastro> Is plasma related
15:49:31 <rdieter> fyi, qt5-qtbase-5.6.0-0.8 and newer includes the latest committed/upstream fixes related to screen connect/disconnect
15:49:57 <heliocastro> rdieter: I know, i even compiled here
15:50:00 <heliocastro> Still crashing
15:50:05 <heliocastro> lxqt runs fine
15:50:16 <heliocastro> plasmashell only crash, kde itself works
15:50:51 <rdieter> seems no upstream kde.org/plasma devs use multiple screens?
15:51:00 <heliocastro> 2 screens == fine
15:51:06 <heliocastro> 3+ kabumm
15:51:09 <RaphGro> same question here all the time …
15:51:19 <heliocastro> Guys, i need to go
15:51:24 <rdieter> heliocastro: thx
15:51:33 <heliocastro> o/
15:51:40 <rdieter> moving on...
15:51:44 <rdieter> #topic open discussion
15:51:48 <rdieter> anything else for today?
15:55:24 <rdieter> sounds quiet, ok, thanks everyone for coming today
15:55:28 <rdieter> #endmeeting