15:04:09 <jreznik> #startmeeting kde-sig -- https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2011-06-07
15:04:09 <zodbot> Meeting started Tue Jun  7 15:04:09 2011 UTC.  The chair is jreznik. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:04:09 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:04:21 <jreznik> #meetingname kde-sig
15:04:21 <zodbot> The meeting name has been set to 'kde-sig'
15:04:32 <jreznik> #topic roll call
15:04:37 <Kevin_Kofler> Present.
15:04:38 <jreznik> who's present today?
15:04:45 * rnovacek is here
15:04:49 * jsmith lurks
15:05:55 <rdieter_work> yo
15:06:17 <than_home> present
15:06:58 <jreznik> #info jreznik Kevin_Kofler rnovacek rdieter_work than_home present, jsmith lurks ;-)
15:07:08 <jreznik> #chair Kevin_Kofler rnovacek rdieter_work than_home
15:07:08 <zodbot> Current chairs: Kevin_Kofler jreznik rdieter_work rnovacek than_home
15:07:26 <jreznik> #topic Agenda
15:07:56 <Kevin_Kofler> 4.7 packaging
15:08:01 <Kevin_Kofler> This Qt patch: https://bugzilla.redhat.com/show_bug.cgi?id=705348#c24
15:08:36 <zaniyah> jreznik: I'm also lurking
15:08:57 <jreznik> 4.6.4 packaging added
15:09:08 <jreznik> zaniyah: ;-)
15:09:28 <jreznik> anything else? a lot of stuff and I expect hot discussion :)
15:09:35 <jreznik> so let's start
15:09:50 <jreznik> easier one
15:09:55 <jreznik> #topic 4.6.4 status
15:10:15 <Kevin_Kofler> That one should be fairly straightforward.
15:10:16 <jreznik> than_home already started to work on 4.6.4, what's the status?
15:10:25 <Kevin_Kofler> Only possible painpoint: kdeedu is missing some .cmake files.
15:10:37 <Kevin_Kofler> So stuff building against kdeedu-devel, if any, might have trouble.
15:11:00 <Kevin_Kofler> If we're missing files, we'll have to add them from the 4.6.3 tarball.
15:11:14 <Kevin_Kofler> But by itself it's reported to build.
15:11:15 <than_home> kde-4.6.4 was commited in fedora git
15:11:44 <than_home> kdelibs, kdebase*, kde-l10n were built
15:12:04 <than_home> the rest are still building
15:12:18 * rdieter_work just re-enabled dist-f14-kde koji target
15:12:26 <jreznik> #info kde 4.6.4 commited to fedora git, kdelibs, kdebase*, kde-l10n are already built
15:12:40 <jreznik> #action rdieter_work to re-enable dist-f14-kde koji target
15:13:00 <jreznik> #info Kevin_Kofler reports missing some .cmake files in kdeedu
15:13:09 <Kevin_Kofler> (That was reported on kde-packager.)
15:13:24 <jreznik> someone should take a closer look on it
15:13:31 <than_home> Kevin_Kofler: do you mean kdeedu-4.6.4?
15:13:52 <Kevin_Kofler> Yes.
15:14:08 <Kevin_Kofler> It's now built from the git 4.6 branch, it used to be built from SVN.
15:14:35 <Kevin_Kofler> (They tried to build 4.6.3 from git and there were many issues, so they just used the last SVN revision. Now the git stuff is sorted out.)
15:14:44 <than_home> ok, i will take a look at this
15:14:56 <Kevin_Kofler> (but there was that report about .cmake files)
15:15:07 <jreznik> #action than_home to take a look at 4.6.4 kdeedu issues
15:15:39 <jreznik> anything else regarding 4.6.4?
15:16:34 <Kevin_Kofler> When we file the updates, we should mark https://bugzilla.redhat.com/show_bug.cgi?id=509645 as fixed.
15:16:48 <Kevin_Kofler> (unless we push my 4.6.3 builds with the fix first)
15:17:29 <Kevin_Kofler> See also https://git.reviewboard.kde.org/r/101513/
15:17:40 <than_home> Kevin_Kofler: please remind us if we forget ;-)
15:18:20 <jreznik> #info rhbz#509645 should be marked as fixed in 4.6.4 update (if not pushed as 4.6.3 one)
15:18:32 * thomasj is as well here
15:19:53 <jreznik> ok, let's move
15:20:18 <jreznik> #topic 4.6.80 aka 4.7 beta 1 packaging status
15:20:33 <jreznik> as you all know, 4.6.80 is a mess...
15:21:11 <jreznik> currently kdelibs, kdebase* and some other packages are built
15:21:34 <jreznik> #info currently kdelibs, kdebase* and some other packages are built
15:21:51 <Kevin_Kofler> So I think we should finish building things in the current structure.
15:21:57 <Kevin_Kofler> Quite some stuff is not split anyway.
15:22:01 <Kevin_Kofler> Let's start with that.
15:22:03 <jreznik> so the base splitted packages are now multisource hacked...
15:22:25 <jreznik> Kevin_Kofler: yep, now the question is how to finish kdeedu, kdegraphics and kdebindings
15:22:37 <Kevin_Kofler> But I think it should be possible to get kdeedu and kdegraphics going at least, kdebindings is missing pieces so that one is going to be the worst mess.
15:22:57 <rnovacek> I started creating new packages for kdeedu apps
15:23:03 <jreznik> I tend to agree with Kevin_Kofler, even it's more work - to finish it with current structure and the slowly start splitting (package reviews etc)
15:23:05 <Kevin_Kofler> kdebindings is always broken.
15:23:12 <than_home> jreznik: we have clear state for kdeedu now, we follow the kde upstream splittings
15:23:18 <Kevin_Kofler> Actually, my plan is to not split at all.
15:23:35 <Kevin_Kofler> I actually want to undo the kdeedu-math split for which the use case went away eons ago.
15:23:36 <jreznik> Kevin_Kofler: that's another question, let's fight current situation
15:23:45 <Kevin_Kofler> (and just merge it into kdeedu and Obsoletes: kdeedu-math there)
15:24:04 <rnovacek> I think splitting is cleaner solution for kdeedu
15:24:32 <rnovacek> those apps are independent (most of them even don't depend on libkdeedu)
15:24:34 <Kevin_Kofler> If you think doing all those reviews can be done in time for Fedora 16…
15:24:36 <jreznik> kdeedu splits are final, annma proved that today
15:25:00 <Kevin_Kofler> Well, there's also the release-team which has a say in what gets released…
15:25:12 * rdieter_work thinks we should follow upstream splits with kdeedu as a test-case
15:25:13 <Kevin_Kofler> The kdeedu developers don't necessarily have the final word.
15:25:15 <jreznik> Kevin_Kofler: it can be done but it will take some time, so I'd like to have kdeedu in monolithic package too
15:25:35 <rdieter_work> see how it goes, possibly follow suit with kdegraphics, kdebindings too
15:25:38 <rnovacek> I put specfiles for kdeedu apps here: http://fedoraproject.org/wiki/SIGs/KDE/KDE47Packaging#kdeedu
15:25:48 <rnovacek> few of them so far
15:25:51 <Kevin_Kofler> Too bad it's only Eric and me fighting the crowd. :-(
15:27:00 <Kevin_Kofler> Doing split SRPMs also means more buildtime dependencies to deal with.
15:27:08 <Kevin_Kofler> (i.e. building libkdeedu before the rest)
15:27:16 <jreznik> Kevin_Kofler: yes, it means...
15:27:18 <rdieter_work> besides Kevin_Kofler, whose opinion is fairly clear, anyone else object to packaging split kdeedu ?
15:27:20 <Kevin_Kofler> And also verifying the BRs for each single package.
15:27:29 <Kevin_Kofler> And issuing many more builds at each update.
15:27:34 <jreznik> for us as packagers - it's more difficult
15:28:01 <jreznik> the real question is if we want to follow our policy to stick as close to upstream
15:28:13 <Kevin_Kofler> And also polluting the update metadata (because there's at least one binary package per SRPM).
15:28:44 <Kevin_Kofler> jreznik: We have some unsplit packages of modular upstreams in Fedora already!
15:28:50 <Kevin_Kofler> See e.g. xorg-x11-apps or how it's called.
15:29:12 <jreznik> Kevin_Kofler: I'm talking about our packages
15:29:14 <Kevin_Kofler> Yes, that's the name.
15:29:16 <Kevin_Kofler> http://pkgs.fedoraproject.org/gitweb/?p=xorg-x11-apps.git;a=blob;f=xorg-x11-apps.spec;h=b16da448804957111d7a1d3871a54cde201d4327;hb=HEAD
15:29:24 <jreznik> we used upstream as excuse not to split
15:29:29 <Kevin_Kofler> There are 21 source tarballs in there.
15:29:49 <jreznik> no when they split we are arguing oposite way... :)
15:30:15 <Kevin_Kofler> X.Org also has split tarballs and the Fedora packaging is not always split.
15:30:38 <Kevin_Kofler> In our packaging, there's kde-l10n and kde-i18n, which have split binary packages, but unsplit source packages.
15:30:44 <Kevin_Kofler> And of course the Marble 1.1 hack.
15:31:01 <Kevin_Kofler> (xorg-x11-apps is entirely unsplit, no subpackage at all.)
15:31:09 <zodbot> Announcement from my owner (jsmith): Fedora Board public IRC meeting in approximately 30 minutes in #fedora-board-meeting
15:31:11 <rnovacek> Now when upstream splitted packages most distros will split too
15:31:20 * rdieter_work doesn't think providing examples of non-split packages should influence this decision either way, imo
15:31:30 <rnovacek> and then more users will ask us to split
15:31:40 <Kevin_Kofler> Most distros already split stuff.
15:31:42 <rdieter_work> rnovacek: that too, +1
15:32:43 <Kevin_Kofler> Representative line of xorg-x11-apps.spec: for app in * ; do
15:33:58 * jreznik tends to stay with current structure at least for 4.7, then try what can we do with kdeedu for F16...
15:34:29 <jreznik> for example - it makes sense for kdeedu as it's a big package of unrelated education apps...
15:34:57 * Kevin_Kofler suggests a kde-sc package with everything. ;-)
15:35:13 <Kevin_Kofler> (Well, not seriously, but it'd be better than the opposite extreme. ^^)
15:37:16 <Kevin_Kofler> But I don't think this discussion is going anywhere.
15:37:40 <jreznik> Kevin_Kofler: looks like you're the only one against, I'm somewhere in the middle
15:38:03 * jreznik understands more work for us but also sees an advantage for users...
15:39:24 <Kevin_Kofler> Let's move on…
15:39:54 <jreznik> so the decision?
15:40:21 <than_home> jreznik: we follow the kde upstream splitting
15:40:26 <rnovacek> stick with current schema and split for f16?
15:40:39 <than_home> split for f16
15:40:42 <jreznik> rnovacek: it's for f16 now in rawhide
15:41:03 <Kevin_Kofler> I guess the idea is that we try both approaches for now to see what gets ready ASAP, but target splitting for this release cycle (i.e. F16).
15:41:13 <rnovacek> but f15 is gonna get 4.7 too, right?
15:41:15 <Kevin_Kofler> (i.e. split ASAP, since everyone except me wants it)
15:41:49 <Kevin_Kofler> You'll see the mess you're creating. ;-)
15:42:00 <rdieter_work> rnovacek: what to do wrt f15/kde47 is a bit unknown at this point, considering the significant change in packaging
15:42:12 <Kevin_Kofler> (but I blame upstream most of all)
15:42:15 <jreznik> rnovacek: I'm not sure I'll vote for that - it depends on the mess created
15:42:36 <than_home> rnovacek: with splitting it's not easy for update in f15
15:42:41 <jreznik> and that's why I propose - let's to first classic kdeedu, graphics, bindings
15:42:48 <Kevin_Kofler> That's another argument for not splitting at all!
15:43:23 <jreznik> then let's play with splitting... but that really means no update of splitted version for older releases
15:43:23 <rdieter_work> Kevin_Kofler: that's a bit extreme, with the same logic we'd never backport anything that involved packaging changes
15:43:56 <rdieter_work> but, change could be mitigated with prudent metapackaging
15:44:04 <than_home> Kevin_Kofler: we will have same problems if we splitt it in the future
15:44:23 <Kevin_Kofler> That's why I said "not at all", i.e. never…
15:44:43 <than_home> it's better to change it now for f16
15:44:57 <rnovacek> Kevin_Kofler: those packages will become hardly maintainable. What if separate releases (like Marble) will hapen more often?
15:45:24 <Kevin_Kofler> xorg-x11-apps (and a few other packages like that) manage that kind of mess just fine.
15:45:49 <Kevin_Kofler> If we're already hacking together multiple tarballs, doing a kdeedu build with a newer Marble hacked in will be even easier.
15:45:58 <Kevin_Kofler> It was already not that hard to do 4.6.2-2.
15:46:04 * rdieter_work doesn't think anyone's going to change their minds with ongoing discussion, and repeated arguments
15:46:53 <Kevin_Kofler> I've done worse hacks, like packaging koffice-kivio to coexist with KOffice 2.
15:47:15 <Kevin_Kofler> (Hopefully Calligra Flow will let me retire that mess…)
15:48:35 <Kevin_Kofler> But I think rdieter_work is right, we should move on.
15:49:22 <jreznik> ok, let's move
15:49:44 <jreznik> #topic Lohit fonts accidentally disable the bytecode interpreter for Qt [patch]
15:49:57 <jreznik> .bug 705348
15:49:59 <zodbot> jreznik: Bug 705348 Lohit fonts accidentally disable the bytecode interpreter for Qt - https://bugzilla.redhat.com/show_bug.cgi?id=705348
15:51:02 <than_home> i will rebuilt qt with the patch soon so we can test
15:51:07 <Kevin_Kofler> So Behdad provided a patch to Qt's fontconfig code which should fix this mess once and for all: https://bugzilla.redhat.com/show_bug.cgi?id=705348#c24
15:51:17 <Kevin_Kofler> than_home: Great, please do.
15:52:09 <jreznik> #info Behdad provided a patch to Qt's fontconfig code
15:52:44 <jreznik> #action than_home to rebuild Qt with the patch soon to test
15:53:15 <rdieter_work> kudos to behdad and his font-fu
15:54:23 <rdieter_work> based on his comments, qt's fontconfig/freetype related code was a bit on the sad side, a little surprising it's worked as well as it has for so long.
15:55:30 <Kevin_Kofler> Hardly anything used fontconfig settings before the new BCI vs. autohint tradeoffs.
15:55:51 <Kevin_Kofler> (per-font settings, in particular)
15:56:08 <rdieter_work> good point, ie, no one ever noticed.... till now.
15:57:09 <rdieter_work> anything else for today, else, jreznik and I have a #fedora-board-meeting to run to here in a few minutes.
15:57:28 <Kevin_Kofler> It's all from me…
15:58:29 <jreznik> ok, thanks guy
15:58:44 <jreznik> I'll prepare summary after board meeting
15:58:49 <jreznik> #endmeeting