kde-sig
LOGS
15:03:08 <jreznik> #startmeeting kde-sig -- http://fedoraproject.org/wiki/SIGs/KDE/Meetings/2012-02-07
15:03:08 <zodbot> Meeting started Tue Feb  7 15:03:08 2012 UTC.  The chair is jreznik. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:03:08 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:03:17 <jreznik> #meetingname kde-sig
15:03:17 <zodbot> The meeting name has been set to 'kde-sig'
15:03:27 <jreznik> #topic roll call
15:03:40 * ltinkl is present
15:03:42 <Kevin_Kofler> Present.
15:04:03 <rdieter> yo
15:04:05 * than is present
15:04:41 * rnovacek is here
15:04:49 * jreznik is obviously here too :)
15:05:03 <jreznik> #chair ltinkl Kevin_Kofler rdieter than rnovacek
15:05:03 <zodbot> Current chairs: Kevin_Kofler jreznik ltinkl rdieter rnovacek than
15:05:17 <jreznik> #info jreznik ltinkl Kevin_Kofler rdieter than rnovacek present
15:05:27 <jreznik> #topic agenda
15:07:29 <jreznik> any topic to be added to agenda?
15:07:49 <Kevin_Kofler> I see 4.8 is already on it…
15:08:10 <rdieter> talk about qt/kde stuff in comps (been meaning to do that for awhile, keep forgetting)
15:08:23 <MarkDude> \o
15:08:29 <rdieter> err... qt/kde critpath I mean
15:08:51 * jreznik wanted to ask what does it mean, got it
15:08:53 <Kevin_Kofler> We also need to fix the KDM error on live image bootup (that "ConsoleKit blablabla" error message).
15:09:48 <rdieter> Kevin_Kofler: just need to add Requires: ConsoleKit somewhere, right?
15:10:33 <rdieter> .bug 786157
15:10:35 <zodbot> rdieter: Bug 786157 kde ConsoleKit support and/or systemd session support - https://bugzilla.redhat.com/show_bug.cgi?id=786157
15:10:42 <rdieter> probably no need to spend meeting time on that, if so.
15:10:50 <Kevin_Kofler> AIUI, that might not be the proper fix, but it'll work as a quick hack, sure…
15:11:16 <jreznik> seems we will have a time today (- critpath :)
15:11:20 <jreznik> anything else?
15:12:05 <jreznik> with these new features, would be great to have a wiki to track, what we are planning for next releases - and we have to support
15:12:57 * ltinkl adds udisks2 support to that list
15:13:24 <Kevin_Kofler> AIUI, the proper fix is that 1. kde-workspace needs support for yet another shutdown method and 2. KDM needs to fail silently if ConsoleKit is not there, unless we can drop ConsoleKit support entirely.
15:13:28 <jreznik> ok, let's add it to meeting
15:13:40 <jreznik> #topic 4.8.0 status
15:14:05 <Kevin_Kofler> The quick hack is to add Requires: ConsoleKit to kde-workspace (for shutdown/restart with GDM) and kdm (to get rid of the error).
15:14:40 <Kevin_Kofler> So one 4.8 blocker should be fixed now: https://bugzilla.redhat.com/show_bug.cgi?id=771053
15:14:49 <Kevin_Kofler> .bug 771053
15:14:54 <zodbot> Kevin_Kofler: Bug 771053 new nepomuk config group - https://bugzilla.redhat.com/show_bug.cgi?id=771053
15:15:21 <jreznik> #info KDE 4.8.0 blockers: new nepomuk config group (#771053) is fixed
15:15:23 <Kevin_Kofler> I fixed that now, in upstream master and 4.8 and in kde-runtime-4.8.0-3.fc17 in Rawhide.
15:15:37 <rdieter> yay, thanks!
15:15:40 <Kevin_Kofler> Please verify that the fix works.
15:15:48 <Kevin_Kofler> So, what about the other 2 blockers?
15:16:29 <Kevin_Kofler> Is the PowerDevil one still relevant?
15:16:53 <Kevin_Kofler> And has anybody looked into the GwenviewPart issue?
15:17:04 <rdieter> powerdevil thing is still plaguing some people, the most notable symptom is that laptop lid close => sleep events still sometimes fail
15:17:38 <Kevin_Kofler> Is that really a profile migration issue as the bug claims? It looks more like a regression in the actual code to me.
15:17:48 <rdieter> as for the gwenview part thing... hrm... workarounds exist, so I'm not too concerned about it, may even that we shouldn't consider it a blocker really
15:18:16 <rdieter> Kevin_Kofler: the summary is probably misleading, true.
15:18:32 <rdieter> that was my initial guess (and most likely wrong at this point)
15:18:52 <Kevin_Kofler> Re Gwenview, what workaround? Ark has no concept of "Open with".
15:19:10 <Kevin_Kofler> So if the default doesn't work, there's no obvious workaround at all.
15:19:11 <rdieter> workaround is to configure a different kpart to handle those particular image types
15:19:19 <rdieter> admittedly, not pretty
15:20:00 <Kevin_Kofler> ltinkl: You said last week that you might be able to look into this, any updates?
15:20:15 <ltinkl> Kevin_Kofler: sry, didn't find anything relevant
15:21:55 <Kevin_Kofler> Upstream is also very unhelpful. :-(
15:22:13 <Kevin_Kofler> I hate it when developers do major changes and are then unable to fix the stuff they broke.
15:23:00 <rdieter> Kevin_Kofler: doesn't even appear upstream is tried to reproduce the problem either.
15:24:14 <Kevin_Kofler> KDE needs a stricter regression policy… If you break stuff and you don't care about fixing it, you should be banned from doing any further major change.
15:24:19 <rdieter> so, I'm ok with not considering either of these blockers anymore (esp if they are still present after 4.8.1 lands)
15:24:43 <Kevin_Kofler> I updated the summary of #771054.
15:24:48 <Kevin_Kofler> .bug 771054
15:24:50 <zodbot> Kevin_Kofler: Bug 771054 PowerDevil lid close 4.8 regression - https://bugzilla.redhat.com/show_bug.cgi?id=771054
15:25:15 <rdieter> thx
15:25:26 <rdieter> that's all I have wrt 4.8.0, move on?
15:25:28 <jreznik> #info  PowerDevil  is actually a regression (lid close) (#771054)
15:25:45 <jreznik> I'm ok with the resolution
15:26:11 <jreznik> #agreed to no longer consider these two bugs (PowerDevil, Gwenview) as a blocker bug
15:26:33 <jreznik> #topic qt/kde in critpath
15:26:53 <Kevin_Kofler> What sucks about the Gwenview issue is that so far only the Ark maintainer has answered, the Gwenview maintainer who caused all the mess has been dead silent.
15:27:34 <rdieter> so, my rough proposal is: 1. remove qt/kde related items from critpath immediately in comps/git 2.  discuss @ devconf what, if anything, we should add back.
15:27:57 <jreznik> rdieter: wouldn't it better to first talk about it and then act?
15:28:13 <jreznik> (and devconf is in a week...)
15:28:26 * ltinkl takes the lid/close bug
15:28:37 <rdieter> maybe. :)  I know Kevin_Kofler's been wanting to drop critpath for awhile...
15:28:47 <Kevin_Kofler> My proposal is: do only 1., discuss more important things than 2. in DevConf. :-p
15:29:22 <jreznik> the thing is - I still believe that idea of critpath is not a bad one - just implementation sucks
15:29:24 <Kevin_Kofler> But if we're going to discuss this anyway, we shouldn't do any changes in comps right now.
15:29:42 <jreznik> and we suck in QA too at all... (sometimes)
15:29:43 <Kevin_Kofler> I also wonder what FESCo will say if we just go and change this.
15:30:02 * Kevin_Kofler worries that we'll get lynched. :-(
15:30:34 <Kevin_Kofler> But in principle I say that all our stuff should be delisted from critpath indeed.
15:30:49 <rdieter> ok, let's wait for devconf then.
15:31:11 <Kevin_Kofler> Not only because it's a PITA for us, but also because it drags in "half of the distro", like qrencode, mysql-server etc.
15:32:10 <jreznik> that's one question - what would say FESCo
15:32:42 <jreznik> one thing - we expected better behaviour from other teams if we join it and show we're Fedora... but you know... never happened
15:33:51 <jreznik> on the other hand - I'd like to be more constructive, so let's try to find the issues we have with critpath, we can try to propose changes and in case FESCo will refuse it - we can ask to be removed from critpath... that's my proposal
15:34:28 <Kevin_Kofler> The main issue is that it's dep-transitive.
15:34:35 <Kevin_Kofler> That just doesn't work for us.
15:34:42 <jreznik> Kevin_Kofler: yep
15:35:00 <Kevin_Kofler> KDE has many dependencies, most of which are very much not critical.
15:35:01 <jreznik> #info the main issue for KDE is that critpath is dep-transitive
15:35:45 <Kevin_Kofler> The next issue is that the definition of critpath got extended over time adding more and more stuff for no good reason.
15:36:05 <Kevin_Kofler> It used to be only the stuff critical enough for the system to boot and update. Then somebody went and decided e-mail is critical too, huh?
15:36:15 <Kevin_Kofler> (That's how kdepim got added.)
15:36:43 <jreznik> #info critpath goes beyond "system boots and updates"
15:36:57 <jreznik> good point too
15:38:41 <Kevin_Kofler> These issues together lead to a HUGE number of packages being unnecessarily critpath, due to being dependencies of the KDE workspace or other "critical" KDE packages.
15:39:40 <Kevin_Kofler> For the record, I also think that the whole concept of critpath is just broken and unnecessary, and in fact I voted against it when I was in FESCo…
15:41:21 * jreznik thinks critpath makes sense but agree with these two issues
15:44:54 <jreznik> so we should write a ticket with proposal for changes to FESCo, #1 is must, #2 would be greate to solve somehow - more critpath levels?
15:45:37 <Kevin_Kofler> Hell no! Not even MORE of a mess!
15:45:47 <Kevin_Kofler> Critpath should just go back to be only really critical stuff.
15:45:48 * rdieter mentions we have complete control over what is considered critpath
15:46:00 <Kevin_Kofler> E-mail is NOT critical.
15:46:02 <rdieter> we just need to decide what that is
15:46:06 <rdieter> and make it so
15:46:20 <jreznik> Kevin_Kofler: makes sense
15:46:43 <jreznik> rdieter: but we don't have control over the deps that's are brought to critpath
15:47:10 <rdieter> jreznik: um, then reduce what's mentioned in comps until it fits exactly what we want
15:48:10 <Kevin_Kofler> I don't think this can be done. We can get a subset of what is really critical, but the subset will probably be empty. ;-)
15:48:25 <rdieter> for example, we don't have to mention anything directly qt/kde
15:48:27 <Kevin_Kofler> The dependencies drag in all sorts of non-critical crap.
15:48:34 <rdieter> and include stuff like:  ConsoleKit, udisks, etc...
15:48:48 <Kevin_Kofler> Those are all critical through GNOME anyway.
15:49:07 <Kevin_Kofler> Well, ConsoleKit probably not anymore, but you get the idea.
15:49:20 <rdieter> Kevin_Kofler: now maybe, but I'd rather not risk depending on someone else's critpath
15:49:21 <jreznik> qt/kdelibs/kde-base* are critical for us, otherwise it doesn't make sense to play the critpath game at all
15:49:48 * rdieter votes to postpone further discussino until devconf, we've other stuff to get to today
15:50:06 <Kevin_Kofler> defer +1
15:50:07 <jreznik> ok for me, I'll add it to FAD schedule
15:50:25 <jreznik> #agreed to postpone critpath to DevConf
15:50:47 <jreznik> there will be also several FESCo members, we can talk to them (marcela, t8m, mitr)...
15:51:01 <Kevin_Kofler> For the Gwenview issue: Can we simply reduce the InitialPreference of the GwenviewPart?
15:51:19 <ltinkl> Kevin_Kofler: yup, seems like a good workaround
15:51:41 <ltinkl> what other parts will come "higher" in the preference list?
15:51:49 * ltinkl thinking of the khtml image part
15:52:01 <rdieter> okular provides a kpart I tested
15:52:10 <ltinkl> for images?
15:52:42 <rdieter> okularkimgio
15:53:03 <rdieter> khtmlimage probably does make more sense
15:53:25 <ltinkl> yup since you don't need any extra app installed
15:54:51 <rdieter> move on++ , hour running out. :)
15:55:23 <Kevin_Kofler> Yeah, move on, we can take care of the tech details later.
15:55:51 <rdieter> I guess there's not much else either, already have a plan wrt ConsoleKit.  planning wiki?
15:56:02 <jreznik> #topic ConsoleKit breakage
15:57:36 <Kevin_Kofler> So what should we do? Add the Requires for now to fix the immediate breakage, look into proper fix later?
15:57:51 <jreznik> Kevin_Kofler: +1
15:57:55 <Kevin_Kofler> We'd only need to update the kde-workspace package, both affected packages are subpackages of it.
15:59:04 <jreznik> I think it's easiest short term solution
15:59:15 <jreznik> the longterm is get rid of CK
15:59:47 <Kevin_Kofler> Right.
16:00:13 <Kevin_Kofler> I can have a look at what's needed to fix this properly, I'll add the Requires for now.
16:00:59 <jreznik> let's add it to our fresh list
16:01:08 <jreznik> planning list
16:02:27 <jreznik> so anything else? we're running out of time (we already ran)
16:03:48 <rdieter> nothing now, we can discuss planning back in #fedora-kde
16:03:57 <jreznik> ok
16:03:59 <jreznik> thanks guys
16:04:04 <jreznik> #endmeeting