kde-sig
LOGS
15:01:53 <rdieter> #startmeeting kde-sig
15:01:53 <zodbot> Meeting started Tue Feb 20 15:01:53 2018 UTC.  The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:01:53 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:01:53 <zodbot> The meeting name has been set to 'kde-sig'
15:01:58 <rdieter> #topic roll call
15:02:07 <rdieter> hi all, friendly kde-sig meeting, who's present today?
15:02:15 <tosky> o/
15:03:13 <rdieter> #info rdieter tosky present
15:03:58 <than> present
15:04:36 <rdieter> #info than present
15:04:39 <rdieter> #chair tosky than
15:04:39 <zodbot> Current chairs: rdieter than tosky
15:05:57 <jgrulich> hello
15:06:26 <rdieter> #info jgrulich present
15:06:28 <rdieter> #chair jgrulich
15:06:28 <zodbot> Current chairs: jgrulich rdieter than tosky
15:06:32 <rdieter> #topic agenda
15:06:35 <rdieter> what to discuss today?
15:07:01 <rdieter> I'd like to give a brief update on Qt5/versioned-symbols in rawhide
15:07:24 <jgrulich> there will be Plasma 5.12.2 released today
15:07:32 <jgrulich> I can do that again
15:08:07 <rdieter> ok, lets start with that
15:08:11 <rdieter> #topic Plasma 5.12.2
15:08:31 <rdieter> jgrulich: for rawhide, go for it asap
15:08:55 <rdieter> now is probably a good time to consider building 5.12.x for f27 too
15:09:17 <jgrulich> rdieter: well, the tars will be available later today and tomorrow I'm on PTO so I will end up probably doing it on Thursday
15:09:33 <rdieter> jgrulich: ok, no problem
15:09:57 <tosky> they are trying to push to fix to 5.12.2, so yes, it may be really late today
15:10:13 <jgrulich> do we want it for F27 as well?
15:10:49 <rdieter> F27: +1
15:10:57 <jgrulich> F27: +1
15:11:39 <rdieter> sure it's another 5.x+1 update, but .12 is LTS, so is more appealing
15:11:58 <tosky> F27: +1
15:12:16 <tosky> Plasma 5.13, in case you didn't see it, will require Qt 5.10
15:12:50 <rdieter> on a related topic, F26, it may be time to freeze doing kf5 updates there (I've already had to backport several FTBFS fixes for plasma-5.10.x components)
15:13:25 <rdieter> tosky: thanks, worth noting here
15:13:35 <Kevin_Kofler> Another argument for getting Qt 5.10 out to our releases…
15:14:07 <rdieter> heh, we just considered 5.12 for stable releases, and you're already ahead to 5.134?
15:14:15 <rdieter> 5.13 ?
15:15:13 <rdieter> I would be perfectly content with the f27 staying on LTS qt5/plasma
15:15:45 <tosky> yep, that was the point: by the time that Plasma 5.13 is out, F27 will be more than half-way through its life and F28 will be out alread
15:15:48 <tosky> already*
15:18:47 <rdieter> Can certainly revist the topic later
15:18:55 <tosky> sure
15:19:06 <rdieter> anything else on 5.12.x ?
15:19:54 <rdieter> moving on...
15:20:14 <rdieter> #topic Qt5 and private api symbol versioning
15:20:40 <rdieter> So, now that we've had the opensuse private api symbol versioning patch in , and rebuild everything... it's not quite as rosy as I'd hoped
15:21:44 <rdieter> I'd hoped there'd be at least a one-way compatibility, that at least we could run binaries produced by unpatched Qt5
15:22:55 <rdieter> I'll try to look into the possibility of fixing that
15:23:23 <rdieter> previously, qtbase provided libQt5Core.so.5(Qt_5_PRIVATE_API)
15:23:32 <rdieter> and now provides only: libQt5Core.so.5(Qt_5.10.1_PRIVATE_API)
15:23:42 <rdieter> I'd like to look into providing both
15:23:46 <rdieter> if possible
15:24:52 <rdieter> any thoughts, comments ?
15:25:21 <tosky> did you ask openSUSE people if they thought about it, and if they did, whether they already tested and thrown some solutions?
15:27:15 <rdieter> tosky: good idea, no I did not
15:27:25 <rdieter> I'll do that
15:28:39 <rdieter> other than that, this is going to be nice not having to worry about tracking this stuff by hand anymore
15:29:42 <rdieter> #topic open discussion
15:29:50 <rdieter> moving on... anything else to discuss today?
15:30:28 <rdieter> over the next week or 2, I'd like to look into giving comps some more attention, Kevin_Kofler had filed some relevant bugs that should be fixable
15:30:38 <Kevin_Kofler> It is surely technically possible to support both symbol versions (and default to the specific one). It just has to be done.
15:31:06 <Kevin_Kofler> I don't really see the need though.
15:31:35 <Kevin_Kofler> Third-party binaries using private APIs have never been supported.
15:31:35 <rdieter> it's just one less incompatibility... I'd like to be able to run binaries produced by unpatched Qt5
15:32:18 <rdieter> there's a big difference between being unsupported and purposely breaking it
15:32:47 <Kevin_Kofler> IMHO, getting a ld.so error is better than getting a crash with no explanation.
15:34:19 <Kevin_Kofler> I can also give a status update on QtWebEngine 5.10.1: It is now built in Rawhide. I am having issues with the freeworld package again, I got it to build for Rawhide on x86_64 and i686, but aarch64 hit an error and I have no idea why because the Fedora build succeeded.
15:35:01 <rdieter> gremlins
15:35:03 <Kevin_Kofler> I want to get this out as a security update to F27 and F26 (that are currently on 5.10.0) ASAP, but first I need it building.
15:35:47 <Kevin_Kofler> https://koji.fedoraproject.org/koji/taskinfo?taskID=25128361
15:35:52 <Kevin_Kofler> http://koji.rpmfusion.org/koji/taskinfo?taskID=198537
15:36:10 <Kevin_Kofler> (Tags: f28 f29 – huh? It's branching time already?)
15:36:47 <rdieter> I think they're just getting ready for it
15:36:55 <rdieter> I noticed that earlier too
15:38:37 <rdieter> oh, and I think after today should have kde spin composable again
15:38:48 <rdieter> calligra finally rebuilt ok
15:40:15 * rdieter will close meeting shortly if there's no other comments
15:41:33 <rdieter> alright, thanks everyone!
15:41:36 <rdieter> #endmeeting