14:13:31 <Kevin_Kofler> #startmeeting KDE SIG Meeting - https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2009-12-01
14:13:31 <zodbot> Meeting started Tue Dec  1 14:13:31 2009 UTC.  The chair is Kevin_Kofler. Information about MeetBot at http://wiki.debian.org/MeetBot.
14:13:31 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
14:13:36 <Kevin_Kofler> #meetingname kde-sig
14:13:36 <zodbot> The meeting name has been set to 'kde-sig'
14:14:20 <Kevin_Kofler> #chair than ltinkl rdieter_work svahl SMParrish
14:14:20 <zodbot> Current chairs: Kevin_Kofler SMParrish ltinkl rdieter_work svahl than
14:14:26 <Kevin_Kofler> So who's present?
14:14:33 <svahl> present
14:14:56 <ltinkl> present
14:14:56 <mathstuf> here
14:16:40 <mefoster> here
14:16:45 * than is present
14:16:45 <Kevin_Kofler> #topic Qt 4.6
14:16:54 <Kevin_Kofler> OK, first topic: Qt 4.6. than?
14:16:58 <SMParrish_mobile> Here
14:17:15 <than> Qt.4.6.0 is now available in rawhide
14:18:12 <than> i will add a regression fix in next qt build
14:18:20 <than> it fixes the kmail issue that rex mentioned on fedora-kde
14:20:44 <mathstuf> builds for F12 for kde-redhat underway? (testing and such)
14:21:10 <Kevin_Kofler> rdieter_work is probably the only one who knows that and he seems to be busy with work.
14:21:15 <than> mathstuf: it's not on my plan
14:21:22 <mathstuf> ok
14:21:52 <Kevin_Kofler> For releases, I guess we'll only push this with KDE 4.4 final?
14:22:01 <Kevin_Kofler> Or do we have any plans to provide Qt 4.6 before?
14:22:09 <Kevin_Kofler> It'd cause chaos for kdebindings, at least. :-(
14:22:25 <than> mathstuf: of course we will build qt-4.6.x with 4.4.0 release for F12
14:22:33 <mathstuf> yeah
14:22:40 <mathstuf> i was thinking of testing
14:22:53 <Kevin_Kofler> Smoke, and thus the Ruby and C# bindings, doesn't build with kdebindings 4.3 and Qt 4.6, and trunk uses a completely new Smoke generator, so fixes can't be easily backported.
14:22:55 <mathstuf> since rawhide is, as yet, quite fun
14:23:23 <mathstuf> id say wait for 4.4 due to the kdebindings mess
14:23:37 <Kevin_Kofler> Same here.
14:23:55 <Kevin_Kofler> I didn't get kdebindings 4.3.3 to build against Qt 4.6 in Rawhide and 4.3.4 didn't change anything there.
14:24:03 <Kevin_Kofler> I'd have to hack things to build, I don't think it's worth it.
14:24:54 <Kevin_Kofler> OK, since time is running away, let's move on...
14:25:01 <Kevin_Kofler> #topic KDE 4.4 Beta 1
14:25:21 <ltinkl> working on it
14:25:44 <ltinkl> I have the CVS done locally, uploading and committing one by one now
14:25:57 * than is working on 4.3.4
14:25:59 <Kevin_Kofler> This should also solve the webkitkde issues.
14:26:08 <Kevin_Kofler> than: That's the next topic. :-)
14:26:44 <Kevin_Kofler> ltinkl: WebKit support is currently commented out in the kdenetwork specfile, it needs to be reenabled.
14:26:55 <ltinkl> Kevin_Kofler: ok
14:27:04 <Kevin_Kofler> Any issues with rebasing patches?
14:27:40 <than> the KDE-4.3.4 upload will finish in next hour
14:27:56 <mathstuf> there are a few that won't apply
14:28:07 <mathstuf> kdebindings comes to mind from some backported stuff
14:28:25 <Kevin_Kofler> Backported stuff from pre-4.3.80 trunk won't apply of course, it's already applied. :-)
14:28:30 <Kevin_Kofler> So those can just be dropped.
14:28:43 <mathstuf> i can run through a mock run if we want to avoid spamming koji to fail in %setup
14:29:03 <Kevin_Kofler> Please tell me about everything that's NOT backported from 4.3.80 and that doesn't apply, I'll fix it ASAP (probably tomorrow).
14:32:31 <Kevin_Kofler> #topic KDE 4.3.4
14:32:45 <Kevin_Kofler> OK, I take it that was all about 4.3.80, so let's discuss 4.3.4.
14:32:54 <Kevin_Kofler> So than started importing this, that's good.
14:33:05 <Kevin_Kofler> The big question: can we get this out to F10?
14:33:35 <Kevin_Kofler> We have a bit less than 2 weeks to get it out to stable if we want it in F10 stable updates.
14:33:50 <Kevin_Kofler> I don't think letting it rot in testing at EOL time is a good plan.
14:34:06 <jreznik> it's bug fix release so it's should be sane but you never know...
14:34:33 <than> Kevin_Kofler: i doesn't make sense to build it for F-10
14:35:19 <SMParrish_mobile> -1 for pushing to f10
14:35:33 <than> -1
14:35:36 <Kevin_Kofler> What if we start building it now and queue it directly to stable once we're done?
14:35:50 <mefoster> Kevin_Kofler: What would be the point?
14:35:59 <Kevin_Kofler> There might even be time for it to sit in testing for 1 or 2 pushes.
14:36:20 <than> Kevin_Kofler: we don't know whether there's any regression in 4.3.4
14:36:21 <ltinkl> -1
14:36:21 <Kevin_Kofler> mefoster: Fixing bugs before F10 goes EOL?
14:36:35 <mathstuf> i don't think we should rush it
14:36:36 <ltinkl> if we screw up something, we won't be able to fix it for F10
14:36:57 <mathstuf> if it makes it, it makes it, but testing would be needed first
14:36:58 <Kevin_Kofler> If we go directly to stable, we should have 1 or 2 pushes left. :-p
14:37:11 <svahl> What about leaving it in updates-testing? So it would be up to the users if they would update KDE to 4.3.4 or be safe with a widely used 4.3.3?
14:37:24 <svahl> or am I missing sth?
14:37:50 <Kevin_Kofler> Well, that's not really what updates-testing is for.
14:38:31 <SMParrish_mobile> F10 is stable and about to eol, we should not push anything to it that could cause issues
14:38:49 <jreznik> but it's only way with schedules as are... I don't like pushing it to stable unless it fixes some important issue (and I'm not aware about any)
14:38:53 * thomasj here, but late as always
14:38:56 <Kevin_Kofler> I'd be very tempted to hit that "Push to stable" button just before the deadline. :-)
14:39:27 <mathstuf> wouldn't make friends im afraid :(
14:39:41 <jreznik> Kevin_Kofler: but that's you ;-) it's like red button :D missile launched
14:39:51 <Kevin_Kofler> FC6 got a KDE point release update in the very last push.
14:40:11 <Kevin_Kofler> There was even a last FE6 push just for the KDE -extras packages.
14:41:12 <jreznik> we don't even know how "stable" 4.3.4 is now without F11/12 builds
14:42:26 <Kevin_Kofler> Well, I guess I'm the only one who wants to push that stuff out to F10.
14:43:08 <jreznik> I'm not strongly against but it should be done really carefully...
14:43:08 <Kevin_Kofler> #topic Phonon/PulseAudio issues due to the F12 Live CD missing xine-lib-pulseaudio
14:43:38 <Kevin_Kofler> So this is a big mess we got ourselves into.
14:44:02 <Kevin_Kofler> We failed to notice that the live image doesn't have the xine-lib PulseAudio plugin on it (nor the GStreamer one for that matter).
14:44:03 <mefoster> There also seems to be a separate, recent PulseAudio/Phonon breakage: http://lists.fedoraproject.org/pipermail/fedora-kde/2009-December/004766.html
14:44:21 <Kevin_Kofler> So Phonon isn't using PulseAudio there.
14:44:39 <Kevin_Kofler> And installing the plugin in an update is not enough to solve the issue for existing users.
14:44:47 <thomasj> ouch
14:44:51 <Kevin_Kofler> The problem is that Phonon remembers incorrect device settings.
14:44:54 <Kevin_Kofler> They need to be reset.
14:45:32 <Kevin_Kofler> That issue from the ML is also "interesting", it's not related to this mess.
14:46:06 <Kevin_Kofler> My big question is, how do I trigger a Phonon setting reset for all the users who ended up with bad settings.
14:46:38 <Kevin_Kofler> I can't even use kconf_update as Phonon stores its settings in QSettings, not KConfig.
14:47:26 <Kevin_Kofler> We need to think of some sort of kludgy hack and I don't know how to do it, especially not without breaking things for people who intentionally turned PA's priority down.
14:48:24 <jreznik> no idea :(
14:48:34 <Kevin_Kofler> I guess something like this could work:
14:48:56 <Kevin_Kofler> if (Fedora12PulseAudioKludge not set in Phonon configuration)
14:49:02 <ltinkl> Kevin_Kofler: installing a system wide config, which overrides the user settings
14:49:03 <Kevin_Kofler> reset Phonon settings
14:49:09 <Kevin_Kofler> set Fedora12PulseAudioKludge
14:49:18 <Kevin_Kofler> ltinkl: That does not work.
14:49:52 <Kevin_Kofler> User settings take precedence and Phonon always saves them even if the user didn't actually set anything.
14:51:37 <Kevin_Kofler> The Mandriva Phonon PulseAudio patches would probably also solve our issue as they disable Phonon's device preferences completely for PA and use PA's instead, but the problem is that they need a new PA with the device-manager module.
14:51:43 <Kevin_Kofler> So we'd need a PA update first.
14:52:01 <Kevin_Kofler> See also https://bugzilla.redhat.com/show_bug.cgi?id=541419
14:52:03 <bugbot_> Bug 541419: medium, low, ---, lpoetter, NEW, pulseaudio device-manager module missing
14:53:06 <than> Kevin_Kofler: perhaps put a script in /etc/kde/env, that fixes the settings by startup kde
14:54:02 <Kevin_Kofler> But that will always reset the settings, even if the user set something intentionally.
14:54:09 <Kevin_Kofler> It's not a very nice solution.
14:55:12 <nucleo> Is KDE in Fedora will be named now "KDE Software Compilation"?
14:56:49 <nucleo> like in announce http://www.kde.org/announcements/announce-4.3.4.php
14:58:25 <SMParrish_mobile> Kevin_Kofler: Can you make the script delete itself the first time it's run
14:58:46 <rjune> SMParrish_mobile, you *can* but it's not advised
14:59:04 <Kevin_Kofler> I think we should just get the Phonon PulseAudio integration in ASAP.
14:59:12 <Kevin_Kofler> That'll make this issue obsolete.
14:59:14 <rjune> if it's part of RPM, then you break the RPM. if it's not, then you install outside of an RPM.
14:59:24 <Kevin_Kofler> We already have the requisite PA in F12.
14:59:32 <Kevin_Kofler> We just need the device-manager module enabled!
14:59:47 <Kevin_Kofler> Then we can just push a Phonon from trunk, or a patched one.
14:59:59 <mathstuf> SMParrish_mobile: instead of deletion, just have it blank itself
15:00:04 <mathstuf> hrm
15:00:12 <rjune> Kevin_Kofler, I don't usually hang out for this, but if that's all you you need, why the discussion on how to solve it?
15:00:15 <mathstuf> that breaks if there are 2 users that need the fix
15:00:30 <SMParrish_mobile> mathstuf: Thats just what I ws thinking
15:00:42 <rjune> mathstuf, lock file in $HOME, then make it run on login
15:01:00 <rjune> it's a simple solution, bus.t it's inelegant, especially if the actual fix is as simple as Kevin_Kofler say
15:01:58 <Kevin_Kofler> Well, it's not really a fix for the underlying issue, but neither is the "reset settings" kludge.
15:02:23 <Kevin_Kofler> The real issue is that Phonon keeps settings in a way which breaks if the available devices change.
15:02:37 <Kevin_Kofler> But fixing that is not really something we can do.
15:02:53 <Kevin_Kofler> I'm going to discuss with lennart about the PA device-manager module.
15:03:03 <rjune> Ahh.
15:03:16 <Kevin_Kofler> If he agrees to shipping that, we can push a new Phonon snapshot to testing and see if that helps.
15:03:37 <jreznik> ok, sounds like a plan
15:03:41 <Kevin_Kofler> But it only really "fixes" the issue by disabling Phonon's own device management entirely when PA is running.
15:03:41 <killefiz> c
15:04:10 <Kevin_Kofler> #action Kevin_Kofler is going to discuss shipping the device-manager module and Phonon PA integration with lennart.
15:04:48 <Kevin_Kofler> If we can't pull that off, I'll see about alternative (hack) solutions.
15:04:55 <Kevin_Kofler> #topic Open Discussion
15:05:04 <Kevin_Kofler> Anything else? We're already over time...
15:05:30 <jreznik> let's move open discussion to #fedora-kde
15:05:41 <Kevin_Kofler> OK
15:05:43 <Kevin_Kofler> #endmeeting