15:09:48 <Kevin_Kofler> #startmeeting KDE SIG Meeting
15:09:48 <zodbot> Meeting started Tue Jul  8 15:09:48 2014 UTC.  The chair is Kevin_Kofler. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:09:48 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:09:52 <Kevin_Kofler> #meetingname kde-sig
15:09:52 <zodbot> The meeting name has been set to 'kde-sig'
15:09:55 <Kevin_Kofler> #topic Roll call
15:10:01 * Kevin_Kofler is present, who else?
15:10:06 * jreznik is here
15:10:08 <jreznik> still here
15:10:13 <Kevin_Kofler> :-)
15:10:20 <pino|work> here
15:13:30 <Kevin_Kofler> danofsatx, danofsatx-work, dvratil, mbriza, rdieter, rdieter_: ping?
15:13:38 <mbriza> hey
15:13:45 <danofsatx-work> aqui
15:14:40 <danofsatx-work> sorry, was catching up on email
15:16:33 <Kevin_Kofler> #chair jreznik pino|work mbriza danofsatx-work
15:16:33 <zodbot> Current chairs: Kevin_Kofler danofsatx-work jreznik mbriza pino|work
15:16:52 <Kevin_Kofler> #info Kevin_Kofler, jreznik, pino|work, mbriza, danofsatx-work present.
15:17:42 <Kevin_Kofler> #topic Agenda
15:19:28 <Kevin_Kofler> So, what to discuss today? (Attendance is sparse, sadly… I guess that's expected at this time of the year, but still.)
15:20:15 <mbriza> pinged jgrulich, he seemed pretty alive a while ago at table tennis
15:20:17 <danofsatx-work> are we participating in the F21 Alpha TC1?
15:20:52 <Kevin_Kofler> mbriza: :-)
15:20:56 <danofsatx-work> or entire release schedule, for that matter?
15:21:28 <Kevin_Kofler> danofsatx-work: As I understand it, yes, at least we didn't ask for anything different…
15:21:49 <Kevin_Kofler> FESCo OKed us as a release-blocking Spin for F21, which to me means releasing on the same schedule.
15:22:20 <danofsatx-work> ok
15:25:59 <Kevin_Kofler> So, if there isn't anything, I'm going to end the meeting right here…
15:26:05 <Kevin_Kofler> (Closing meeting in 60 seconds.)
15:26:40 <danofsatx-work> uh, ok...
15:27:05 <danofsatx-work> for the record, in case no one has been watching....Alpha TC1 release is today
15:27:54 <Kevin_Kofler> If they forget us @rel-eng, I'm going to yell loudly. ;-)
15:28:10 <danofsatx-work> heh
15:28:14 <Kevin_Kofler> #topic F21 KDE Schedule
15:28:17 <danofsatx-work> you? no, you don't say.....
15:28:27 <mbriza> alpha tc1 release? it was branched already?
15:28:31 <Kevin_Kofler> #info F21 KDE follows the general F21 Schedule.
15:28:35 <jreznik> it will probably take some time to produce tc1
15:28:48 <danofsatx-work> did I misunderstand?
15:29:09 <jreznik> mbriza: branch is today but not yet done, I was asking Dennis a few moments ago
15:29:28 <Kevin_Kofler> #info FESCo approved the KDE Spin as release-blocking for F21, which implies following the release schedule.
15:29:51 <mbriza> jreznik: ah, thanks
15:30:01 <danofsatx-work> ok, branch is today then.
15:30:06 * danofsatx-work is very confused
15:30:30 <Kevin_Kofler> #info F21 Alpha scheduled to branch today, F21 Alpha TC1 planned for right afterwards (but may take time).
15:30:40 <Kevin_Kofler> danofsatx-work: TC* releases are releases known to be broken.
15:30:50 <Kevin_Kofler> If they'd be shippable, they'd be called RC*, not TC*.
15:31:04 <Kevin_Kofler> So a TC can be scheduled right after branching, it's not a contradiction.
15:31:08 <danofsatx-work> oh, I know...I just didn't expect my knowledge of the process and schedule to be broken
15:31:18 <Kevin_Kofler> Of course, typically, things are so broken that the release doesn't compose at all, and so TC1 slips.
15:33:06 <Kevin_Kofler> Even if it's only a TC for an Alpha release, there have to be at least some ISOs getting spat out of the compose process. :-)
15:33:52 <jreznik> mbriza: dgilmore will start it soon
15:34:18 <jreznik> btw. adamw drafted release criteria for workstation and kde, take a look pls
15:34:32 <mbriza> yeah, i was just wondering, i knew the branch should have been somewhen around today
15:36:19 <mbriza> wow, arm test cases
15:36:40 <Kevin_Kofler> jreznik: URL please…
15:36:48 <mbriza> https://fedoraproject.org/wiki/QA:Desktop_validation_results_template
15:37:15 <mbriza> https://fedoraproject.org/wiki/User:Adamwill/Draft_workstation_release_criteria
15:39:04 <Kevin_Kofler> ARM is only getting tested with KDE? Wow…
15:39:12 <Kevin_Kofler> GNOME's Hell still doesn't work properly on ARM?
15:39:30 <danofsatx-work> heh.....poetic justice
15:39:43 <adamw> it's not release blocking on ARM, just because of GPU support
15:39:56 <jreznik> Kevin_Kofler: kde 5 is around, so...
15:40:11 <adamw> things are basically as-they-were in f20 for KDE
15:40:22 <Kevin_Kofler> Is llvmpipe still so broken on ARM that it crashes gnome-shell?
15:40:29 <adamw> but do point out any places where there are inadvertent logic fails or anything in the process due to extensive mucking around
15:40:37 <pwhalen> gnome does work on one arm target (at least), but KDE was selected as release blocking
15:40:47 <adamw> Kevin_Kofler: i don't know about that, but ARM hardware generally isn't powerful enough that it'd give a plausible experience with llvmpipe, aiui.
15:41:29 <Kevin_Kofler> GPU support is going to be an issue for Plasma 5, i.e. probably starting from Fedora 22, too, by the way.
15:41:56 <jreznik> yep
15:41:59 <adamw> mbriza: the workstation criteria are Workstation-specific, to be clear. they don't apply to KDE.
15:41:59 <Kevin_Kofler> Qt 5 requires OpenGL 2, you have to have at least software rendering (llvmpipe) working to use Qt 5 GUI apps.
15:42:19 <adamw> mbriza: the stuff in the 'generic' page (current draft https://fedoraproject.org/wiki/User:Adamwill/Draft_F21_Alpha_criteria) is what's applicable to the KDE spin.
15:43:06 <mbriza> ah, okay, thanks
15:43:44 <danofsatx-work> but, but, but......we can do everything the Workstation product can do, and do it better!
15:43:57 * danofsatx-work sulks
15:44:01 <Kevin_Kofler> danofsatx-work: But we can't use the testcases 1:1.
15:44:24 <Kevin_Kofler> There's a test case that says that Evince etc. must be installed, of course we will "fail" that, by design. ;-)
15:44:31 * danofsatx-work actually reads the criteria
15:44:41 <Kevin_Kofler> (this one: https://fedoraproject.org/wiki/QA:Testcase_workstation_core_applications )
15:45:14 <Kevin_Kofler> If we want something like that, we need our own version with our own list.
15:45:51 <Kevin_Kofler> Actually, Evince is not on the list, but e.g. gedit is.
15:46:24 <Kevin_Kofler> The list makes sense only for Workstation. So in that case, it's normal that the test case is only applicable to Workstation.
15:46:48 * jreznik has to run now
15:47:19 <Kevin_Kofler> Some other test cases would make sense though…
15:47:30 <Kevin_Kofler> E.g. the theming one: https://fedoraproject.org/wiki/QA:Testcase_workstation_theming
15:47:57 <Kevin_Kofler> Of course, applications will not look the same as on the Workstation image, but they should look consistent here too.
15:48:11 <Kevin_Kofler> (all using Oxygen theming, vs. Adwaita on GNOME/Workstation)
15:48:26 <mbriza> Qt uses QtGTK by default in gnome3
15:48:47 <Kevin_Kofler> mbriza: Yes.
15:48:48 <mbriza> which then in turn uses gtk2
15:49:00 <Kevin_Kofler> Indeed, it's not ported to GTK+ 3 yet.
15:49:20 <Kevin_Kofler> But if GTK+ 2 apps pass the test, Qt apps will, too.
15:49:39 <Kevin_Kofler> It just tests that QGtkStyle works. :-)
15:51:11 <Kevin_Kofler> #topic Open discussion
15:51:14 <Kevin_Kofler> So, anything else?
15:52:17 <kalev> https://bugzilla.redhat.com/show_bug.cgi?id=1098735 maybe? might need some developer attention.
15:52:43 <Kevin_Kofler> .bug 1098735
15:52:46 <zodbot> Kevin_Kofler: Bug 1098735 apper: hawkey backend missing features - https://bugzilla.redhat.com/1098735
15:54:26 <Kevin_Kofler> Uhm, yeah…
15:54:40 <Kevin_Kofler> It's sad that Richard Hughes doesn't feel responsible.
15:55:09 <Kevin_Kofler> It's really his code that is not implementing the required interfaces.
15:55:33 <Kevin_Kofler> Just because GNOME Software doesn't use them doesn't mean they can just throw an error.
15:58:52 <kalev> I heard there was a plan to have Apper use appstream data too, maybe it doesn't need the comps support then?
15:59:45 <Kevin_Kofler> I guess we need to discuss this with dantti, the upstream Apper developer.
16:00:33 <Kevin_Kofler> That said, there are still some blockers preventing AppStream support for working in Fedora.
16:00:49 <Kevin_Kofler> The metadata being generated is invalid XML that GNOME's tools happen to accept.
16:01:19 <Kevin_Kofler> Hughsie didn't test any of his new stuff outside of GNOME. This really sucks.
16:01:57 <Kevin_Kofler> We also don't have AppData for KDE apps in yet.
16:02:15 <Kevin_Kofler> So the AppStream data would show only GNOME apps, which is also not acceptable of course.
16:02:25 <kalev> can you file tickets about the invalid XML issues?
16:02:28 <kalev> https://github.com/hughsie/appstream-glib
16:02:54 <Kevin_Kofler> The invalid XML from the Fedora generation process is already filed somewhere in RH Bugzilla (IIRC, rdieter filed it).
16:03:39 <Kevin_Kofler> The fact that appstream-glib accepts it, well, as long as it also accepts valid XML, I wouldn't bother complaining…
16:04:25 <Kevin_Kofler> There's also:
16:04:26 <Kevin_Kofler> .bug 1057843
16:04:29 <zodbot> Kevin_Kofler: Bug 1057843 [abrt] PackageKit: pk_backend_cancel(): packagekitd killed by SIGSEGV - https://bugzilla.redhat.com/1057843
16:04:38 <Kevin_Kofler> that didn't get any response from Richard Hughes.
16:05:09 <kalev> No, I mean, the program that generates the data that's shipped in the appstream-data package lives in https://github.com/hughsie/appstream-glib
16:05:23 <kalev> so it's the place to fix issues with the XML generation
16:08:51 <Kevin_Kofler> By the way, I also dislike how policy decisions on what applications are listed are imposed at metadata generation time, so Apper has no possibility to list anything GNOME doesn't like.
16:08:56 <Kevin_Kofler> This is very broken.
16:08:56 <kalev> I commented in bug 1057843
16:09:43 <Kevin_Kofler> E.g., they require Comment= in the .desktop file and a 48×48 or larger icon, which are things only GNOME needs.
16:10:15 <Kevin_Kofler> They also decided not to list Qt 3 (and GTK+ 1) apps by policy.
16:10:48 <Kevin_Kofler> All these are things that don't belong in the metadata generator.
16:13:44 <kalev> anyway, just wanted to bring up this to make sure KDE has a working software installer
16:13:51 <kalev> nothing else from me, thanks
16:15:54 <danofsatx-work> I have a hard stop in 5 mins, so I'm done....
16:17:56 <Kevin_Kofler> Looks like that's all, thanks all for coming and see you next week!
16:17:59 <Kevin_Kofler> #endmeeting