kde-sig
LOGS
15:00:02 <jreznik> #startmeeting kde-sig -- https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2011-01-18
15:00:03 <zodbot> Meeting started Tue Jan 18 15:00:02 2011 UTC.  The chair is jreznik. Information about MeetBot at http://wiki.debian.org/MeetBot.
15:00:03 <zodbot> Useful Commands: #action #agreed #halp #info #idea #link #topic.
15:00:19 <jreznik> #meetingname kde-sig
15:00:19 <zodbot> The meeting name has been set to 'kde-sig'
15:00:39 <jreznik> #chair rdieter Kevin_Kofler than
15:00:39 <zodbot> Current chairs: Kevin_Kofler jreznik rdieter than
15:01:00 <jreznik> #topic roll call
15:01:06 <Kevin_Kofler> Present.
15:01:08 <jreznik> who's present today?
15:01:09 * rrix is present
15:01:11 <than> present
15:01:26 <jreznik> welcome back rrix!
15:01:37 <rrix> jreznik: back? this is my first sig meeting.. :)
15:01:38 * rdieter_work is occupied @ work, will be back hopefully within 5-10 minutes
15:02:50 <jreznik> ok, anyone else?
15:03:21 <jreznik> #info Kevin_Kofler rrix than jreznik rdieter_work present, nucleo listening
15:03:40 <jreznik> #topic Agenda
15:04:51 <Kevin_Kofler> Missing optional deps for 4.6 maybe?
15:05:23 <jreznik> Kevin_Kofler: it's done - it wasn't a big deal - lot of are disabled for a good reason or are not needed
15:05:36 * rnovacek is little bit late
15:05:59 <Kevin_Kofler> We're missing some stuff for routing in Marble (on the local machine, without relying on a web service which only supports Europe and which has usage restrictions in its TOS), at least.
15:06:16 <jreznik> Kevin_Kofler: but I'll recheck once 4.6 final is out - tomorrow?
15:06:52 <Kevin_Kofler> It's not NEEDED to not regress in functionality compared to 4.5, but it's useful and IMHO we should get the stuff packaged.
15:07:00 <Kevin_Kofler> We're also missing a bunch of optional KOffice deps.
15:07:05 <Kevin_Kofler> I haven't looked at other stuff.
15:07:39 <jreznik> Kevin_Kofler: ok, I'll add it to agenda
15:08:25 * ltinkl is here (a bit late)
15:08:46 <jreznik> #info rnovacek and ltinkl present too
15:08:50 <jreznik> #topic missing deps (4.6, koffice)...
15:09:20 <jreznik> for kdeedu I see only optional experimental Python Marble bindings in log
15:09:25 * SMParrish smparrish back from the abyss
15:09:31 <Kevin_Kofler> So I haven't looked at all packages, but for kdeedu (Marble), there are 2 routing tools that can be used to compute routes locally.
15:09:42 <Kevin_Kofler> Hmmm, have they dropped the routing stuff?!
15:09:52 <than__> do we have list of missing deps for 4,6?
15:09:55 <jreznik> Kevin_Kofler: I don't know or I'm blind
15:10:00 <Kevin_Kofler> It was there in earlier prereleases.
15:10:21 <jreznik> than__: I was checking it - mostly these deps are disabled because we want them disabled
15:10:27 <jreznik> Kevin_Kofler: http://kojipkgs.fedoraproject.org/packages/kdeedu/4.5.95/1.fc15/data/logs/i686/build.log
15:10:55 <jreznik> ops, it's not in optional part as usual
15:11:02 <jreznik> liblocation>=0.102
15:12:44 <jreznik> the question is - do we want to enable marble python bindings?
15:13:34 <Kevin_Kofler> Uh, no, liblocation is not what I was thinking of.
15:13:42 <Kevin_Kofler> liblocation is for GPS on the N900.
15:13:50 <Kevin_Kofler> We don't support the N900. ;-)
15:14:00 <than__> jreznik: i think yes, if it's stable
15:14:09 <jreznik> than__: it's experimental feature
15:14:23 <jreznik> Kevin_Kofler: it's only n900 one?
15:14:28 <than__> jreznik: ok, disable in this case
15:14:40 <jreznik> than__: it's disabled by default
15:14:58 <Kevin_Kofler> gosmore and routino are the offline routing things.
15:15:37 <Kevin_Kofler> Looks like it's a runtime dependency only, the stuff is always built.
15:15:43 <Kevin_Kofler> That's why it doesn't show up in the build.log.
15:15:48 <jreznik> Kevin_Kofler: makes sense
15:16:38 <jreznik> runtime deps are not easy to catch :(
15:16:52 <Kevin_Kofler> For KOffice, there's a bunch of compile-time deps missing, they're already in a specfile comment.
15:17:41 <Kevin_Kofler> Looks like lcms 2 is already available (lcms2 package).
15:17:55 <Kevin_Kofler> I don't know whether we should build against that or stick with lcms 1.
15:18:04 <Kevin_Kofler> The other stuff is probably all not packaged (yet).
15:18:07 <jreznik> Kevin_Kofler: I can take a look tmrw
15:18:35 <jreznik> #info KOffice: there's a bunch of compile-time deps missing, they're already in a specfile comment
15:18:54 <jreznik> #info kdeedu: gosmore and routino are the offline routing, runtime deps
15:19:18 <jreznik> #info kdeedu: we don't want to enable experimental Python bindings for Marble now
15:19:51 <rdieter_work> back, hi.
15:19:51 <jreznik> #action jreznik to look on missing KO deps
15:20:03 <jreznik> rdieter_work: heya!
15:20:12 <Kevin_Kofler> jreznik: liblocation is part of Maemo.
15:20:29 <Kevin_Kofler> I think it doesn't support any hardware other than the N900's.
15:20:41 <jreznik> Kevin_Kofler: I know, I just wasn't sure if it is generic one or not
15:20:53 <Kevin_Kofler> We have gpsd support already.
15:21:28 <jreznik> it's probably ment only for marble n900 port
15:21:52 <Kevin_Kofler> Right.
15:22:05 <Kevin_Kofler> http://websvn.kde.org/?view=revision&revision=1111815
15:23:41 <Kevin_Kofler> I can have a look at gosmore and routino, but somehow they both look like a mess to package to me. :-(
15:24:12 <Kevin_Kofler> Maybe I can also convince the local GIS geek (volter) to look at that stuff. ;-)
15:24:31 <Kevin_Kofler> I need to look closer before I can promise anything.
15:24:45 <jreznik> #action Kevin_Kofler to look at gosmore and routino (and convince local GIS geek - volter - to look at too)
15:24:57 <jreznik> Kevin_Kofler: "to look" :)
15:25:38 <jreznik> my 3G modem should be GPS too but I don't know it's supported in Linux (or even it's really there)
15:26:03 <ltinkl> jreznik: same as mine, it has GPS, but only works under Windows
15:27:07 <jreznik> ok then - I think, I've checked all build time deps but I'll recheck once 4.6 final is out again (should be the same)
15:27:12 <jreznik> we can move on
15:27:34 <jreznik> (or the question is - how to deal with runtime ones?)
15:29:38 <jreznik> #topic keep kdepim-4.6 or revert to kdepim-4.4
15:29:52 * ltinkl expects a heated discussion :)
15:30:33 <rdieter_work> I'm mostly interested in hearing from anyone who's tested kdepim-4.5.94.1 , kmail in particular
15:31:02 <rdieter_work> I had a tentatively positive experience setting it up against my gmail imap account.
15:31:22 * jreznik is not sure what version tested... I don't have my 4.6 netbook around...
15:31:48 <rdieter_work> dgilmore reported testing it against his ~100gb email store.  summary: worked better than any previous releases, got further, still chewed cpu, never finished, and akonadi crashed
15:31:53 <jreznik> really huge 1gb+ imap - initial import worked very well - much more better than 4.4 kmail
15:32:21 <jreznik> (ah, 1gb is not a huge one :)
15:32:53 <jreznik> but after reboot it started to eat my cpu - nepomuk, akonadi, virtuoso, mysql fighting 100% of my cpu
15:32:58 <Kevin_Kofler> Well, if it now works for rdieter, ship it! :-)
15:33:11 <Kevin_Kofler> (The previous versions didn't.)
15:33:27 <ltinkl> ...still chewed cpu, never finished, and akonadi crashed...
15:33:34 <ltinkl> doesn't sound like it works
15:33:47 <Kevin_Kofler> Yeah, well, 100 GB of IMAP data… :-/
15:33:52 <Kevin_Kofler> Does the average user have that? ;-)
15:34:40 <ltinkl> Kevin_Kofler: same for jreznik with 1GB
15:34:54 <jreznik> ltinkl: no crash - I had to kill it manually
15:35:04 <jreznik> but yes - it's much more better than before!
15:35:09 <than__> i tested new kdepim today, got several crashes
15:35:11 <ltinkl> ye but eating the CPU is no good either
15:35:12 <Kevin_Kofler> But it does indeed sound that it's still a scalability regression compared to the old stuff.
15:35:30 <Kevin_Kofler> Despite the improvements over previous releases which didn't scale at all.
15:35:36 <than__> i haven't taken a look where it crashed
15:35:39 <ltinkl> nepomuk and akonadi are still very much unfinished projects imho
15:36:01 <Kevin_Kofler> The question is: can we support 4.4 forever?
15:36:10 <ltinkl> and for a crucial stuff like email, contacts or calendarm we should not ship it
15:36:20 <Kevin_Kofler> Upstream right now plans to discontinue kdepim 4.4 support after 4.4.10.
15:36:21 <rdieter_work> forever, of course not.  for at least one more fedora release, certainly.
15:36:27 <ltinkl> I mean, if it was a game or some minor app, no problem
15:36:50 <ltinkl> Kevin_Kofler: they will certainly not discontinue that
15:37:09 <rdieter_work> I could also look into shipping kdepim-enterprise branch
15:37:10 <jreznik> Kevin_Kofler: I'm not sure upstream can abandon 4.4 release...
15:37:22 <ltinkl> I don't think SUSE or other distros plan to switch to the new kdepim anytime soon
15:37:23 <rdieter_work> err,  s/I/we/
15:37:26 <Kevin_Kofler> Of course, if we stick strictly to Ubuntu-style updates, we'd ship 4.4.10 and just not update it, ever.
15:37:45 <jreznik> quest for kde siggers - try to import your mailboxes to new kdepim...
15:37:47 <ltinkl> no, not ever, I'd just reevaluate that for F16
15:38:01 <Kevin_Kofler> rdieter_work: The problem is, there are 2 branches, enterprise4 and enterprise5. Enterprise5 is basically 4.5. Enterprise4 has no Akonadi at all.
15:38:07 <ltinkl> cmon, life doesn't end after f15
15:38:13 <Kevin_Kofler> So it doesn't have the Akonadi KAddressBook introduced in 4.4.
15:38:22 <Kevin_Kofler> So it'd mean downgrading KAddressBook.
15:38:46 <Kevin_Kofler> I don't think that's possible without causing data migration issues.
15:38:50 <rdieter_work> fine by me, the akonadi integration in 4.4 is mildly painful anyway, but... handling the 'downgrade' could be toublesome, nevermind
15:39:15 <rdieter_work> only practical option then is 4.4.x
15:39:27 <Kevin_Kofler> (FWIW, 4.4's KAddressBook is incomplete crap which should never have been shipped. It's supposedly better in 4.6, but now there's other messy stuff.)
15:39:45 <rdieter_work> so, is it worth continuing testing and asking for more feedback on kdepim-4.6-beta(s)?
15:40:09 <Kevin_Kofler> They should really have made 4.4 equivalent to enterprise4. I guess if they'd known they'd be supporting it for so long, they'd have done it.
15:40:14 <rdieter_work> I'd say it has become at least testable recently
15:40:16 <ltinkl> rdieter_work: ye, I'd just ship it in kde-unstable for the brave souls, for Fedora I would stick with kdepim4.4
15:40:29 <rdieter_work> Kevin_Kofler: agreed
15:40:39 * ltinkl nods too
15:40:44 <Kevin_Kofler> The halfway-done Akonadi migration in 4.4 is really lame, it was supposed to be an interim release to be replaced ASAP.
15:40:56 <ltinkl> yup
15:41:56 <ltinkl> so the resolution?
15:41:57 <Kevin_Kofler> I'd stick to 4.6 in Rawhide and make a final reversion decision by final freeze (i.e. beta freeze).
15:42:06 <jreznik> Kevin_Kofler: +1
15:42:17 <rdieter_work> I think that's another ~3 weeks or so
15:42:20 <ltinkl> ok
15:42:31 <ltinkl> gives us more time to test
15:42:41 <Kevin_Kofler> If by then it's still broken, Epoch it out of F15 and keep it in Rawhide for F16.
15:43:13 <than__> Kevin_Kofler: agreed
15:43:14 <rdieter_work> but is anyone actually going to test it?  (our testing feedback so far hasn't been very good)
15:43:23 <Kevin_Kofler> https://fedoraproject.org/wiki/Releases/15/Schedule
15:43:24 <rdieter_work> mostly because it's been so broken
15:43:38 <than__> rdieter_work: it's the problem if it's not well tested
15:43:39 <ltinkl> rdieter_work: yup, I have 4.6 built locally anyway, I'll compile kdepim as well
15:43:44 <Kevin_Kofler> Final freeze is March 22.
15:43:52 <ltinkl> (on my other testing user account)
15:44:08 <rdieter_work> Kevin_Kofler: I'd propose making the decision at feature freeze instead
15:44:27 <than__> which kdepim version  does kde upstream recommend?
15:44:29 <Kevin_Kofler> Or rather "beta change deadline" as it's called now, used to be "final devel freeze".
15:44:34 <rdieter_work> that's when features are supposed to be testable anyway.
15:44:44 <Kevin_Kofler> rdieter_work: It's testable.
15:44:47 <ltinkl> than__: 4.4 for now
15:44:47 <Kevin_Kofler> In fact we want it tested!
15:44:48 <jreznik> rdieter_work: as I said - kde siggers should be the brave ones now...
15:44:58 <rdieter_work> Kevin_Kofler: barely testable now.  we'll see
15:45:12 <Kevin_Kofler> It does not need to be 100% complete until the beta change deadline.
15:45:14 <rdieter_work> like I said, if it's actually testable by feature freeze, fine.  If not, revert
15:45:56 <rdieter_work> no point in contiuing the charade after feature freeze if it's not realistically testable
15:46:01 <rdieter_work> by then...
15:46:36 <Kevin_Kofler> So make 2 go/no-go decisions for kdepim 4.6? 1. Is it testable by Feb 8 (feature freeze)? 2. Do the tests look good by 100% complete deadline (March 22)?
15:46:50 <rdieter_work> sounds better to me, yes.
15:47:17 <Kevin_Kofler> Of course the decision needs to be done BEFORE March 22, as the reversion would have to be already pushed at that point.
15:47:37 <rdieter_work> of course, that's an implementation detail. :)
15:48:05 <Kevin_Kofler> The Beta should really have the kdepim we're going to ship.
15:49:08 * jreznik agrees
15:49:09 <rdieter_work> indeed.  anyway, that's a plan I can support 100%.
15:49:49 <jreznik> than__, ltinkl?
15:49:59 * ltinkl agrees
15:50:06 <rnovacek> +1
15:50:09 <rdieter_work> putting off the decision means more work for me, I hadn't finished working on the -mobile splits for 4.6 yet.  underhandedly hoped to be able to put that off until f16. :)
15:50:29 <than__> +1
15:51:46 <rdieter_work> rrix: I forget, your plasma'ification of kontact/summary didn't make it into 4.6 or did it?
15:52:12 <Kevin_Kofler> rdieter_work: Comment out the stuff in CMake?
15:52:18 <Kevin_Kofler> It doesn't work anyway.
15:52:22 <Kevin_Kofler> (Or does it now?)
15:52:31 <Kevin_Kofler> (It was basically unusable in the betas.)
15:53:05 <rdieter_work> I guess it doesn't matter too much, was just curious.
15:53:21 <rdieter_work> looks like we agree on the plan, we can probably move on
15:56:40 <jreznik> #agreed to make 2 go/no-go decisions for kdepim 4.6? 1. Is it testable by Feb 8 (feature freeze)? 2. Do the tests look good by 100% complete deadline (March 22)?
15:56:48 <jreznik> #topic some keyboard options/layouts (set via system-config-keyboard) make kdm input non-functional
15:57:02 <jreznik> .bug 661395
15:57:04 <zodbot> jreznik: Bug 661395 some keyboard options/layouts (set via system-config-keyboard) make kdm input non-functional - https://bugzilla.redhat.com/show_bug.cgi?id=661395
15:58:03 <rdieter_work> this one's a bit out of my ignorant-monoligual-american league.
15:58:52 <rdieter_work> If folks could test for reproducibility, and offer any clues, insight in the bug, would be much appreciated
15:59:23 * jreznik can install some other keyboard layouts :) using only en_us one :D
15:59:50 <rdieter_work> fwiw, I marked it a f15 blocker
16:00:17 <jreznik> ok
16:00:22 <jreznik> time is over now...
16:00:29 <jreznik> thanks all!
16:00:34 <jreznik> #endmeeting