Jun 16 18:06:46 --- Kevin_Kofler hat das Thema geändert auf: KDE SIG Meeting - https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2009-06-16 - Init Jun 16 18:06:49 Who's present? Jun 16 18:06:57 * ltinkl is present Jun 16 18:06:59 * SMParrish here Jun 16 18:07:02 here Jun 16 18:07:18 --> maxamillion (n=adam@li23-229.members.linode.com) hat #fedora-meeting betreten Jun 16 18:07:55 <-- nphilipp hat sich getrennt ("Leaving") Jun 16 18:08:29 --- Kevin_Kofler hat das Thema geändert auf: KDE SIG Meeting - https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2009-06-16 - kde-4.2.90 status Jun 16 18:08:35 First topic: KDE 4.2.90 status. Jun 16 18:08:41 Kevin_Kofler, let start the meeting Jun 16 18:08:47 And in particular, kdebindings-4.2.90 build failures for < f12. Jun 16 18:09:11 is there any build log file for this? Jun 16 18:09:27 there's a message to the kdebindings ML Jun 16 18:09:29 --> ltinkl_ (n=ltinkl@194.212.22.14) hat #fedora-meeting betreten Jun 16 18:09:38 i built fine with F12, Jun 16 18:09:41 http://apt.kde-redhat.org/apt/kde-redhat/mock/fedora-11-i386-kde/kdebindings/build.log Jun 16 18:10:42 fyi, fails similiarly for f-10 and f-11. Jun 16 18:11:02 What Phonon are you using for those builds? Jun 16 18:11:09 In F12 we have Qt's Phonon. Jun 16 18:11:17 This may be dependent on the Phonon version. Jun 16 18:11:31 <-- iarlyy (i=c8a46b5e@gateway/web/ajax/mibbit.com/x-eb751d22f0951f4d) hat #fedora-meeting verlassen Jun 16 18:11:31 <-- hanthana hat sich getrennt (Read error: 110 (Connection timed out)) Jun 16 18:11:46 yes it looks like issue in phonon Jun 16 18:12:06 we use the phonon in qt in f11 Jun 16 18:12:17 s/f11/f12 Jun 16 18:12:18 Hmmm, actually, wait! Jun 16 18:12:33 This is the issue with the old PyQt4 which we already had a patch for before we upgraded PyQt4 in Rawhide. Jun 16 18:12:39 My memory just "clicked". :-) Jun 16 18:12:53 let me check Jun 16 18:12:57 The syntax to disable the meta object is different in the old and new PyQt4. Jun 16 18:13:20 So we need that old-PyQt4 patch if we want to build this for kde-redhat unstable without updating the PyQt stack. Jun 16 18:13:23 --> phunster_ (n=lanny@pool-74-105-133-148.nwrknj.fios.verizon.net) hat #fedora-meeting betreten Jun 16 18:13:44 Note that the patch needs to be conditionally applied only where PyQt4 is old. Jun 16 18:13:51 <-- mcepl hat sich getrennt (Read error: 110 (Connection timed out)) Jun 16 18:14:00 It would be possible to support both by just giving both the old and new spelling of the keyword. Jun 16 18:14:24 I can do that if you prefer it over conditionalizing the patch. Jun 16 18:14:26 <-- warren hat sich getrennt ("Leaving") Jun 16 18:16:14 <-- inode0 hat sich getrennt ("Leaving.") Jun 16 18:16:18 Kevin_Kofler, Hm i couldn't find the patch Jun 16 18:16:22 <-- comphappy hat sich getrennt ("Quit") Jun 16 18:17:06 Look in the CVS attic for kdebindings. :-) Jun 16 18:17:20 --> tatica (n=tatica@nelug/designer/tatica) hat #fedora-meeting betreten Jun 16 18:17:31 fyi, I was using latest qt/phonon doing these guilds Jun 16 18:17:33 builds Jun 16 18:17:49 --> fengshaun (n=armin@blk-224-129-27.eastlink.ca) hat #fedora-meeting betreten Jun 16 18:18:04 --> ltinkl__ (n=ltinkl@194.212.22.14) hat #fedora-meeting betreten Jun 16 18:18:07 <-- ltinkl hat sich getrennt (Nick collision from services.) Jun 16 18:18:07 Actually it's not in the attic, it's still there... Jun 16 18:18:10 http://cvs.fedoraproject.org/viewvc/rpms/kdebindings/devel/kdebindings-4.2.85-old-PyQt4.patch?revision=1.1&view=markup Jun 16 18:18:12 <-- ltinkl_ hat sich getrennt (Nick collision from services.) Jun 16 18:18:16 --- ltinkl__ ist jetzt bekannt als ltinkl Jun 16 18:18:19 rdieter: If you have the latest PyQt4, you DON'T want that patch. Jun 16 18:18:23 it's still there Jun 16 18:18:27 If you have the old PyQt4, you need it. Jun 16 18:18:28 * ltinkl is back (hopefully) Jun 16 18:18:37 Check your specfile conditionals. Jun 16 18:18:58 Alternatively, we can change the patch to use: Jun 16 18:19:05 class Notifier : QObject /NoDefaultCtors,PyQt4NoQMetaObject,NoQMetaObject/ Jun 16 18:19:07 * rdieter checks Jun 16 18:19:14 i.e. both the old and new keyword. Jun 16 18:19:14 if that's it, *so* sorry. Jun 16 18:19:45 yep, that's it, damn. my bad, move on nothing to see here. :) Jun 16 18:20:00 Kevin_Kofler, this change is good Jun 16 18:20:59 <-- ltinkl hat sich getrennt (Read error: 60 (Operation timed out)) Jun 16 18:21:00 rdieter, could you please try with this patch? Jun 16 18:21:16 and see if it really fixes the issue Jun 16 18:21:34 I was just going to not use the old-PyQt4 patch at all. Jun 16 18:21:37 --> jreznik_mobile (n=jreznik_@89-24-6-101.i4g.tmcz.cz) hat #fedora-meeting betreten Jun 16 18:21:58 or does the new patch work either way? Jun 16 18:22:17 rdieter, it should Jun 16 18:22:20 If you change it to use both keywords, it'll work with both the old and new version (or at least it should). Jun 16 18:22:21 cool, will do Jun 16 18:22:28 <-- rishi hat sich getrennt ("Ex-Chat") Jun 16 18:22:31 It's a bit ugly, but it avoids having to conditionalize. Jun 16 18:22:57 So we don't have to worry about when we upgrade PyQt4 in the official F10 and F11 repos. Jun 16 18:23:42 --- fengshaun ist jetzt bekannt als fengshaun|away Jun 16 18:24:29 ok, doing test builds, move on Jun 16 18:24:31 --> oshan (n=DarkM@124.43.61.30) hat #fedora-meeting betreten Jun 16 18:24:59 --- Kevin_Kofler hat das Thema geändert auf: KDE SIG Meeting - https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2009-06-16 - OO.o KDE 4 integration Jun 16 18:25:04 Next topic: OO.o KDE 4 integration Jun 16 18:25:10 See also: http://lists.freedesktop.org/archives/ooo-build/2009-June/000036.html Jun 16 18:25:17 It seems some patches recently landed in oo.o-build. Jun 16 18:25:33 i asked dtardon for help Jun 16 18:25:39 I haven't checked how many changes to the core OO.o packaging is needed and how much of it can be done in a separate SRPM. Jun 16 18:25:52 What we definitely need is a maintainer (this is why we don't have an oo.o-kde right now). Jun 16 18:26:13 The OO.o maintainer provided a specfile somewhere in Bugzilla, but didn't want to maintain the package, he wants a KDE guy to handle it. Jun 16 18:26:27 Note that the specfile was for the old KDE 3 integration. Jun 16 18:27:00 how stable is oo.o-kde? Jun 16 18:27:06 And only the parts in upstream OO.o, some stuff was only in oo.o-build even for KDE 3, he said it can be added later, but whoever wants to maintain it should sort it out (and also sort out what OO.o patches are needed). Jun 16 18:27:10 ill check tomorrow Jun 16 18:27:35 jreznik_mobile: If you're volunteering to maintain it, that'd be great. Jun 16 18:28:11 I can sign up for comaintaining, but I'd like somebody else to handle the initial packaging and review submission. Jun 16 18:28:15 im not sure but david sits near to me Jun 16 18:28:29 --> ldimaggi_ (n=ldimaggi@nat/redhat/x-237fde210adbcf60) hat #fedora-meeting betreten Jun 16 18:28:44 so we can talk on lunch Jun 16 18:29:13 hes our oo guy in brno Jun 16 18:29:39 --> itbegins (n=Simon@nat-studcudn-172-24-62-0.fitz.cam.ac.uk) hat #fedora-meeting betreten Jun 16 18:30:04 I hope we can get this sorted out. We probably still won't ship OO.o on the live image for space reasons, but having the KDE integration available for those who need OO.o would be great. Jun 16 18:30:23 <-- oshan hat sich getrennt (Read error: 104 (Connection reset by peer)) Jun 16 18:30:40 indeed Jun 16 18:30:47 We also need to check what's available. Full integration would be NWF (native widget framework), dialogs and URL handlers. Jun 16 18:31:02 its nice to have it Jun 16 18:31:11 agreed, I am curious to see how the new koffice is going to be though, the current one looks strange being qt3 and all Jun 16 18:31:23 For URL handlers, last I checked, ooo-build had a trivial patch to enable the gnome-vfs stuff on KDE, it could be enabled, but ideally we'd have KIO integration. Jun 16 18:31:32 before ko2 gets ready Jun 16 18:31:49 ok Jun 16 18:31:52 Even after KOffice 2, I think there will still be a use for OO.o-kde. Jun 16 18:32:07 how big is the oo.o-kde? Jun 16 18:32:11 <-- phunster_ hat sich getrennt ("Ex-Chat") Jun 16 18:32:11 <-- phunster hat sich getrennt ("Ex-Chat") Jun 16 18:32:18 <-- ldimagg__ hat sich getrennt (Read error: 113 (No route to host)) Jun 16 18:32:22 yes, i agree that there will still be a user case for oo.o-kde Jun 16 18:32:44 than: I don't know, but the big thing is OO.o itself, not the KDE integration. Jun 16 18:32:45 than: as I understand it, stock ooo + kde additions Jun 16 18:32:59 -kde would just be the KDE integration and require the core OO.o. Jun 16 18:33:07 ah ok Jun 16 18:33:25 imho few so Jun 16 18:33:29 --> openpercept (n=openperc@fedora/openpercept) hat #fedora-meeting betreten Jun 16 18:33:44 Most likely we'd have to get -gnome split out too at some point, but first we need -kde packaged in the first place. Jun 16 18:35:03 So, jreznik_mobile, are you going to check ooo-build for KDE-related patches? We need to see what can be put in a -kde SRPM, what should go to the main OO.o package (like that patch to enable the GNOME vfs stuff if there's still no KIO support, so URLs work in KDE too) etc.? Jun 16 18:35:04 Kevin_Kofler, yes it does make sense to split -gnome Jun 16 18:35:25 so kde user can only have -kde stuff Jun 16 18:35:30 Of course, splitting -gnome makes less sense if we drag it in for the VFS stuff anyway. So that'd be an argument for not doing that. Jun 16 18:35:32 ok Jun 16 18:35:54 --> phunster (n=lanny@pool-74-105-133-148.nwrknj.fios.verizon.net) hat #fedora-meeting betreten Jun 16 18:36:50 Though if they support the current gvfs stuff, that one only needs glib2 AFAIK, not all of GTK+ etc. Jun 16 18:37:15 But gvfs itself would still be a dep. Jun 16 18:38:10 Anything else about OO.o? Jun 16 18:38:40 all Jun 16 18:38:53 move on Jun 16 18:39:35 --- mchua ist jetzt bekannt als mchua_lunch Jun 16 18:39:57 move on please Jun 16 18:40:51 --- Kevin_Kofler hat das Thema geändert auf: KDE SIG Meeting - https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2009-06-16 - Selinux/knotify issues Jun 16 18:40:54 Some recent bugs: Jun 16 18:40:58 http://bugzilla.redhat.com/506126 Jun 16 18:41:09 http://bugzilla.redhat.com/506122 Jun 16 18:41:30 bug #506126 Jun 16 18:41:32 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=506126 medium, low, ---, than, NEW, SELinux is preventing knotify4 from making the program stack executable. Jun 16 18:41:39 bug #506122 Jun 16 18:41:41 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=506122 medium, low, ---, than, ASSIGNED, SELinux is preventing knotify4 from changing a writable memory segment executable. Jun 16 18:41:43 yay Jun 16 18:41:48 Most likely 2 instances of the same bug. Jun 16 18:42:03 by all accounts so far, induced by rpmfusion stuff somewhere Jun 16 18:42:04 The submitter claims all he uses is Fedora and some codecs from RPM Fusion Free (list provided). Jun 16 18:42:23 So it looks like a bug in one of the Fedora or RPM Fusion packages. Jun 16 18:42:39 Unfortunately, SELinux never tells us which shared object is actually at fault. :-/ Jun 16 18:42:45 <-- a hat sich getrennt (Remote closed the connection) Jun 16 18:42:46 The error reporting really needs to be improved there. Jun 16 18:43:16 AFAIK, SELinux actually knows which .so requires exec* (as the labels can be set per shared object), it just doesn't tell us. Jun 16 18:44:41 <-- fengshaun|away hat sich getrennt (Read error: 104 (Connection reset by peer)) Jun 16 18:46:37 --> abadger1999 (n=abadger1@65.78.187.8) hat #fedora-meeting betreten Jun 16 18:46:51 I guess what's needed is to go through all shared libs around and readelf them all to check for exec* requirements. :-( Jun 16 18:47:13 Unless someone here knows how to get SELinux to spit out the required information. Jun 16 18:47:45 Might be a good idea to cc dwalsh on those and get his comments. Jun 16 18:47:49 Kevin_Kofler, add the selinux maintainer to cc Jun 16 18:48:10 he could help us to debug the problem Jun 16 18:49:01 OK, I'm doing that. Jun 16 18:49:40 Kevin_Kofler, it's great if you could take care of this issue Jun 16 18:50:42 I see dwalsh is already CCed on #506122. Jun 16 18:51:06 I'll ping him to ask if he can help us debug it. Jun 16 18:51:22 (in Bugzilla at first, I may try to catch him on IRC if he doesn't reply) Jun 16 18:51:47 Kevin_Kofler, thanks Jun 16 18:54:15 Done. Jun 16 18:54:27 --- Kevin_Kofler hat das Thema geändert auf: KDE SIG Meeting - https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2009-06-16 - Open Discussion Jun 16 18:54:30 Anything else to discuss? Jun 16 18:54:35 We have 6 minutes left. Jun 16 18:54:37 any thing else? if not we can close the meeting and continue to discuss on fedore-kde Jun 16 18:54:43 fyi, kdebindings built fine Jun 16 18:54:47 woo Jun 16 18:54:57 rdieter, goog news :) Jun 16 18:57:41 --> inode0 (n=inode0@fedora/inode0) hat #fedora-meeting betreten Jun 16 18:57:50 <-- sseiersen hat sich getrennt (Read error: 104 (Connection reset by peer)) Jun 16 18:59:23 jreznik_mobile: Are you doing the summary today? svahl is not here... Jun 16 18:59:58 --- mchua_lunch ist jetzt bekannt als mchua Jun 16 19:00:18 I can e-mail you the log if your mobile doesn't log. :-) Jun 16 19:00:50 --- Kevin_Kofler hat das Thema geändert auf: Channel is used by various Fedora groups and committees for their regular meetings | Note that meetings often get logged | For questions about using Fedora please ask in #fedora | See http://fedoraproject.org/wiki/Meeting_channel for meeting schedule. Jun 16 19:01:04 OK, time's up, we can discuss the logging stuff on #fedora-kde. :-)