14:08:02 #startmeeting KDE SIG Meeting - https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2009-09-29 14:08:02 Meeting started Tue Sep 29 14:08:02 2009 UTC. The chair is Kevin_Kofler. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:08:02 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:08:28 #chair than rdieter ltinkl jreznik SMParrish mathstuf 14:08:28 Current chairs: Kevin_Kofler SMParrish jreznik ltinkl mathstuf rdieter than 14:08:45 So, who's present? Me obviously. ;-) Who else? 14:08:52 * jreznik is here 14:08:53 * SMParrish here 14:08:58 * mefoster is here today ... 14:09:23 * ltinkl is here 14:09:39 #chair mefoster 14:09:39 Current chairs: Kevin_Kofler SMParrish jreznik ltinkl mathstuf mefoster rdieter than 14:09:41 * than is here 14:10:04 What happened to rdieter? He was here a few minutes ago... 14:10:24 dunno, I suppose some last-minute work issue 14:10:51 Let's get started then? 14:11:05 yes please 14:11:10 #topic kde-sig steering committee 14:11:34 * thomasj is here 14:12:13 So we had some discussion earlier on #fedora-kde about forming a formal steering committee to take contentious decisions so we know who exactly should be voting in such cases. 14:12:43 The suggestion was to have a committee of 7 people, and a list of 7 was circulated. 14:13:27 The problems are: how to legitimate the initial steering committee? And what should be the succession rules? 14:13:51 i don't see the need for this. 14:14:15 And should the initial people be allowed to keep their position as long as they want? What if we have many new contributors? 14:15:09 imo it just complicates the desition 14:15:40 than: Well, the reason this came up is that we have been unable to reach a decision on how to proceed with Phonon for something like 3 weeks, we're officially already frozen now, so we need a decision quickly. 14:15:51 We need a way to make timely decisions and not defer them forever. 14:16:52 But somehow I get a feeling that we won't be able to get a decision on this one either. 14:17:26 For legitimation, I supposed I could get the list in front of FESCo and see if they approve it. 14:17:27 I understand why, but I don't have any clue how to solve this issues 14:17:53 A committee of 7 should get a decision ;) 14:18:10 Dont think FESCo needs to get involved. we should be able to govern ourselves 14:18:13 maybe leader, who will be pushing to decide things in time would be enough 14:18:19 But going to FESCo is something we should only do once we can show that we all agree with the idea and just want it officialized. 14:18:37 * ltinkl doesn't see the need for FESCO's aproval either 14:19:07 Kevin_Kofler, FESCo doesn't have to do with the dicision in KD 14:19:12 s/KD/KDE 14:19:49 OK, so let's forget I brought up the F word. ;-) 14:20:12 did you? :) 14:20:20 fkdesco :D 14:21:03 i don't see the need to make the desition now, every user cann alway switch to xine-backend if the want 14:21:03 I think one person taking decisions isn't going to make people happy. 14:21:24 sorry, got roped into an emergency staff meeting, back now. 14:21:37 I mean, I'd volunteer to be the Kofler Dictator Emperor (KDE), but I think we need a more democratic approach. ;-) 14:21:48 Kevin_Kofler: not doing decision but pushing people to decide on time 14:21:53 I think a steering committee is a good idea, creates a stucture for deciding issues, and gives new contributors an idea of who does what 14:22:02 SMParrish: +1 14:22:39 SMParrish: but then you need rules for decision process 14:22:41 too 14:22:47 than: we're not talking only about the Phonon issue, such decision making needs to be standardized for the future as well 14:23:07 There's also the K3b/KOffice issue right now, and others may come up in the future. 14:23:13 ltinkl: indeed, phonon just highlights the need for it 14:23:28 yup 14:23:29 In the past, we had the default menu issue, that one was basically decided by an ad-hoc vote. 14:23:42 Which ended up pretty close (1 vote). 14:23:52 such issues that need decisions come pretty often imo 14:24:09 And Phonon is similar now, so it's important to know who should be allowed to vote if we go for voting. 14:24:19 And if not, I'd like to hear your suggestions for an alternate procedure. 14:24:39 And I really don't think one of us deciding alone is going to be the solution. 14:24:51 We all make mistakes. 14:24:57 there is no better alternative imho :) 14:25:13 Kevin_Kofler, i agree with you :) 14:25:26 if we really need a steering committee, we should limt to 5 users on the list 14:25:34 and we cant have a dictatorship or anarchy, so voting is the only way 14:26:24 i like Kevin's proposal of 7 3 RH 4 Community 14:27:00 seems we've mostly agreed for the need here, bikeshedding on size next. either 5 or 7 is ok with me, honestly, 7 seems a better fit, based on who we have as good candidates right now 14:27:47 who will be on the list? 14:28:05 SMParrish: hh, there are even not 3 RH people working officially on KDE :) 14:28:05 The 3 RH members are kinda obvious: than, ltinkl, jreznik 14:28:19 me, lukas, jaroslav, rex, kevin and ? 14:28:30 For the 4 community ones, I've suggested rdieter, svahl, SMParrish and myself. 14:29:02 svahl as the live CD maintainer and SMParrish as the main triager and maintainer of kpackagekit and a few other packages. 14:29:24 +1 for svahl and SMParrish 14:29:42 I agree on this list 14:29:57 agreed here too 14:30:19 I do as well. its a good well rounded list 14:30:51 than, ltinkl, jreznik, svahl , SMParrish, kevin and rex are on the lis 14:30:52 I hope mathstuf won't be angry for being left out. rdieter suggested him too, but 8 members would mean we could end up with 4:4 stalemates, so that's no good, and I think those 7 really need to be on there. 14:30:53 the list was fairly easy - but now issues - what next - how to nominate new member? 14:31:19 Kevin_Kofler: no hard feelings; time squeeze lately anyways 14:31:42 jreznik: in the packaging committee we pretty much self-nominate (remaining committee members vote), we could do similar 14:31:47 mathstuf: OK :-) 14:32:32 rdieter: Agreed, I think an FPC-style approach is going to be the best solution. 14:32:50 Elections aren't going to work without a clear way to define who's allowed to vote. 14:33:26 So the FPC's self-renewing approach is probably the best at this point. 14:33:34 and how to make free space for nominated member? or we're going to grow to even numbers? 14:34:39 we can tackle that when the time comes. For now, I'd recommend sticking with 7 and not growing. meaning, someone would have to resign (or get kicked out) in order to add a new member 14:34:43 jreznik, do we really need it? 14:35:06 more people on the list will complicate the decition 14:35:20 imo it doesn't make sense 14:36:49 so we have the list of "deputies", let's move on? :) 14:36:51 how many votes do we need to accept decision? 14:36:59 simple majority 14:37:09 ltinkl, yes 14:37:13 7 members, so we need 4 +1 votes to pass a decision. 14:37:35 but 14:37:50 I guess we'll want to make things like in most Fedora committees where you need 4 votes even if people are missing. 14:37:58 and how many votes to be valid? at least 2/3 of members? 14:37:59 what if there are only 4 members present in the voting? would 3 votes still be sufficient? 14:38:08 ltinkl: no 14:38:17 ok :) that was my main concern 14:38:26 well, we could consider that, but I'd recommend against it, imo 14:38:33 People who are not present can vote per e-mail, preferably before the meeting. 14:38:35 maybe we should vote by email - so even people not attending meeting can vote 14:38:59 either way is fine, let's be flexible 14:39:01 Just do it like in FESCo: folks normally vote in the meeting, but you can vote per e-mail beforehand if you know you'll be missing. 14:39:21 we're trying to make decision making easier, not harder. :) 14:39:26 Of course if we can get to +4 by e-mails already, we don't have to bring the issue up in the meeting. :-) 14:39:26 rdieter, +1 14:40:10 For a vote to be valid, there need to be 4 +1 votes, period. If only 4 people are there, they can only pass decisions unanimously. If only 3 people are there, no voting can be held. 14:40:21 Kevin_Kofler: so that's I don't like voting before - the meeting is to talk about issue 14:40:46 off to class 14:40:54 But of course that's only where voting is actually necessary. We don't want to make decision making harder than now, we can still have our usual consensus decisions. 14:41:14 Only if somebody calls for a vote, we need to have one. 14:41:29 I think we have some good groundwork set, can we move on now? 14:41:57 jreznik: Voting before the meeting is only a fallback if you know you'll be missing. 14:42:11 We can also take votes afterwards if we don't have enough votes during the meeting to decide either way. 14:42:24 So you can read the log and base your mail-in vote on that. 14:43:43 We just need to get 4 votes for one or the other solution in in some way. 14:44:02 ok, could someone summarize it somewhere to accept rules and let's move 14:44:19 jreznik: after meeting ok? 14:44:33 or did you want the summary now? 14:44:57 rdieter: probably after meeting - mailing list? 14:45:42 #agreed The KDE SIG Steering Committee will be formed by (in alphabetical order): jreznik, Kevin_Kofler, ltinkl, rdieter, SMParrish, svahl, than. 14:46:03 cool, I can do that (eventually I can, unless someone beats me to it) 14:46:26 #agreed 4 votes will be required to pass decisions where a vote is called for. 14:46:37 rdieter: yes please, some skilled guidelines writer should take care :D 14:46:59 #action rdieter will summarize the exact rules. 14:47:09 * rdieter looks around for the mysterious referenced skilled writer :) 14:47:30 * ltinkl stares back at rdieter :)) 14:47:50 * rdieter concedes defeat 14:47:55 * Kevin_Kofler thinks we really need to move on. :-) 14:48:01 moveon++ 14:48:05 #topic k3b/koffice reverts, recommended by upstreams 14:48:25 #link http://rdieter.livejournal.com/15770.html 14:48:33 Some background there. 14:49:01 This is a bit of a contentious decision, and it's needed really soon as F12 is theoretically already frozen. 14:49:20 In the comments of that blog post, I brought up some reasons not to revert. 14:49:21 (sorry I blogged before having a thorough discussion) 14:49:43 But there are also arguments in favor of a reversion. 14:49:48 having to make last minute decisions always suck 14:50:40 * jreznik does not use k3b anymore, so does not know the current state of it 14:50:48 For K3b, Kubuntu appears to be shipping 1.66 alpha2 in Karmic too, and it seems to be fairly stable. There's a crash with ripping, but I think I already have a fix for that. 14:51:37 For KOffice 2.1, it's already beta, not alpha, we've shipped other betas in the past, and in fact we're only on a beta in the first place because we decided not to ship 2.0.x (and that was probably the right decision). 14:51:54 Kevin_Kofler, we are nit sure how stable the 1.66 really 14:52:01 KOffice 2 is missing some apps, but we can ship a koffice1 SRPM with just those apps (which wouldn't be on the live CD). 14:52:16 the k3b is not very stable from what I saw 14:52:19 KOffice developers are still saying that it's not ready for serious work 14:52:26 same for KOffice 14:52:31 indeed 14:52:31 The main KOffice apps are all in 2.1 and we'd ship the latest version. 14:53:03 * ltinkl will brb 14:53:27 * thomasj uses the beta for work.. cant say if it's serious work.. but for his business 14:53:30 All the work on things like better ODF compatibility across all programs, improving KFormula etc. is all focused on 2.x, 1.6 is a dead codebase. 14:54:36 there's one reason why to ship ko2 for me - krita 14:54:56 On the other hand, I'll admit I did exactly zero testing of the new K3b and KOffice. 14:55:16 jreznik: Yeah, that one has improved a lot too. 14:55:43 running short on time, I'd call for voting on these items (separately?) 14:55:57 separately 14:56:04 but I'm not sure how to vote 14:56:24 FYI 3 k3b bugs in rawhide but 2 are related to genisoimage and k3b itself 14:56:33 The problem with separate votes is that the benefit of not reverting is lower if we don't keep both KDE 4 versions. 14:56:45 sebastian knows very well the current state of k3b if he means k3b-1.66 is not ready for consumption 14:56:45 (We'd still drag in the KDE 3 stuff onto the live image.) 14:57:01 it's unstable 14:57:03 Proposal for vote: revert to (kde3) k3b-1.0.x and koffice-1.6.x for F-12 14:57:24 * rdieter doubts anyone voting would do so differently for each, but if so, please comment 14:57:28 it's not good to ship this version in F12 release 14:57:31 -1, for the reasons already detailed 14:57:35 +1 14:57:47 +1 (revert 14:57:50 than: I assume that's a +1 to reverting? 14:58:03 +1 to reverting 14:58:35 SMParrish, jreznik: ? 14:59:01 -1 if there are issues no one is filing bugs about them 14:59:31 jreznik: it's in your hands :) 14:59:43 Yeah, why should we revert based on fictional bugs nobody runs into? 15:00:04 And do we really want to ship an older K3b than Kubuntu? 15:00:05 * jreznik is not sure - does not use it... 15:00:31 Kevin_Kofler, the most users still use the stable k3b 15:00:50 I'd like to see new KO2 but as developers says it's not ready... last time I had something in my hands - I've tested it... 15:00:51 * rdieter thinks upstream developers know best, better than we can conjecture about 15:01:00 it's no wonder why there's no report in bugzilla 15:01:02 Can i open files created with KO2 in KO1? 15:01:03 Kubuntu Karmic will apparently be shipping 1.66 only, Mandriva seems to be already shipping it. 15:01:22 but with policy - we do what upstream says +1 to revert 15:01:27 thomasj: For the stuff which uses ODF, yes, but formatting may be broken. 15:01:29 thomasj: good question 15:01:31 For Krita, I don't know. 15:01:37 I use KO2 for my business 15:01:47 Well, no F12 for me then 15:01:56 Kevin_Kofler, we don't have to do what Mandrake does 15:01:56 No big deal 15:01:57 thomasj: don't worry,we'll continue to provide ko2 at least unofficially 15:02:05 rdieter, ok, thanks 15:02:06 thomasj: redhat-kde repo 15:02:30 rdieter: we have a winrar :) 15:02:33 Kevin_Kofler, It's not what sebasian wants 15:02:40 ok, looks like proposal passes, move on? (though we're close to out of time) 15:03:13 Well, I don't think we have any intention to change our votes, so that'd mean it passes 4:2. 15:03:27 We could solicit svahl's opinion by e-mail for the record, but it won't change the outcome. 15:04:26 But he told us he already has a plan to make room on the live image for kdelibs3, so we need not worry about that. 15:04:39 bugzappers, yell at us to clear out if you're ready (sorry) 15:04:41 I still think it's a missed opportunity not to go pure KDE 4 now, but the majority has spoken. 15:04:46 * jreznik would like to see kdelibs3 free image... 15:05:07 #agreed F12 will revert to (kde3) k3b-1.0.x and koffice-1.6.x for F-12 (passed 4:2). 15:05:12 yup me too but ppl should realize there is life after F12 15:05:19 jreznik: Are you changing your vote then? 15:05:25 Kevin_Kofler: :))) 15:05:30 Kevin_Kofler, :) 15:05:43 Kevin_Kofler: but we have fedora-kde repo, we can concentrate on more stable system - even with kdelibs3 15:05:43 he isn't changing his vote, move on :DD 15:06:04 sorry redhat-kde repo 15:06:08 #topic future of Phonon 15:06:16 #info upstream (sandsmark) recommends building/packaging phonon from qt, and building/packaging backends separately 15:06:20 do we have time for this discussion here? :) 15:06:27 #info mandriva developments integrating pulseaudio support (and improving gstreamer backend) 15:06:32 #link http://mail.kde.org/pipermail/phonon-backends/2009-September/000304.html 15:06:34 Kevin_Kofler: -> fedora-devel mailing list 15:06:35 or should we move back to our channel 15:06:50 #link http://colin.guthr.ie/git/phonon/log/?h=pulse 15:06:57 I've just sent it there (after rdieter approval) 15:07:06 * ltinkl somehow feels this will be lengthy and painful 15:07:24 there's really no time to wait, honestly, we need a decision, like yesterday 15:07:52 I'd propose a slightly different tack, esp based on recent phonon list developments, 15:08:48 well, I was about to suggest going back to kde/phonon completely, would allow us to more easily benefit from and cherry pick from mandriva's ongoing work 15:08:49 Proposal: revert to the F11 style of Phonon packaging: standalone Phonon, phonon-backend-xine by default, but build Qt against GStreamer so qtconfig-qt4 can set up Phonon-GStreamer. (We have both GStreamer and xine-lib on the live image anyway.) 15:09:05 rdieter: Yeah, that's my proposal too. 15:09:09 ok 15:09:26 A standalone Phonon SRPM allows us to ship whatever Phonon we want or need to ship, without relying on what Qt ships. 15:09:32 +1 15:09:38 -1 15:09:39 +1 from me obviously 15:10:02 +1 15:10:24 a standalone phonon seems ok to me, esp, given the benefits of putting in mandriva's work 15:10:35 but I'd go with gstreamer as the default backend 15:10:46 (I've stated the reasons many times before) 15:10:51 * jreznik agrees with ltinkl 15:11:07 That makes +3 -3. 15:11:27 Ask svahl by mail for the deciding vote? 15:11:42 Kevin_Kofler: where do you take those numbers from :) 15:11:43 ok, I think we can consider the standalone part of the proposal a success, the default backend hinging on svahl's vote 15:12:00 yup 15:12:02 rdieter: Right. 15:12:32 #agreed We will move back to building a standalone phonon SRPM. 15:13:01 #info The vote for the default backend is split 3:3, needs the 7th vote from svahl. 15:13:28 Kevin_Kofler, does it we drop phonon part from qt? 15:13:40 s/does it/ does it mean 15:13:50 Yes, like we did in the past. 15:13:56 The real upstream for Phonon is not Qt. 15:13:58 it's bad 15:14:18 It's good because it allows us to switch to any Phonon branch, like the Mandriva one, or upstream trunk. 15:14:28 We don't have to wait for Qt to switch or use a huge patch. 15:14:39 the qt/phonon upstreamer recommend to use the part in qt 15:14:52 why do we drop it? 15:15:02 than: that was the lesser of his recommendations honestly (xine being the stronger one) 15:15:35 And he also recommended building both backends from the Phonon tarball, even the GStreamer one. 15:16:19 more alive stack is better - question is backward compatibility for Qt only apps 15:16:21 At that point, I think we can just as well use the Phonon tarball throughout, that way the library will always match the backends and we can more easily switch to another branch. 15:16:26 rdieter, which GStreamer-backend will be dropped? 15:16:42 nothing will be dropped 15:17:01 we wil have 2 phonon gstreamer-backend? 15:17:05 The GStreamer backend in the Phonon tarball is the same one as the one in qt. 15:17:26 Except it seems to be a newer version, at least in phonon trunk. 15:17:28 oh, phonon-backend-gstreamer will be dropped from qt packaging, yes 15:17:35 added back to phonon packaging 15:17:40 We'll just ship the version from phonon. 15:17:43 Kevin_Kofler, i know, why do we have to use the one in phonon/KDE 15:17:47 let's clear out , we're over time 15:17:58 Because that's what upstream recommended. 15:18:07 it's easier to add the gstreamer patches to Qt 15:18:07 And because it's newer and has features KDE 4.4 will need. 15:18:32 but not Trolltech 15:18:32 Why would we patch the copy in Qt with a huge patch as opposed to just packaging the tarball which already has the correct version? 15:19:25 I think it's really easier to take the Phonon tarballs from KDE, esp. since that's where the development happens 15:19:45 Kevin_Kofler, using the phonon snapshot is not good idea 15:19:53 it could break the Qt 15:20:29 We don't use a snapshot in F12, we use the stable tarball. 15:20:45 But for F13, we're going to consider the Mandriva branch for sure. 15:21:06 Kevin_Kofler, it's better to use Qt part in this case 15:21:12 They're working on PulseAudio integration in both backends and they may also have some GStreamer backend bugfixes there. 15:21:46 any end in sight, guys? btw, I vote for whatever colin guthrie does =) 15:21:54 lol 15:22:04 dude's good 15:22:27 ye, it seems he's done a pile of impressive work there 15:22:29 adamw: We really want to ship his Phonon branch, but not in F12 as it's under heavy development and F12 is in feature freeze. ;-) 15:23:16 He's working on having PulseAudio sources/sinks show up in Phonon as devices instead of the one "PulseAudio" device which goes to the default sink. 15:23:22 ah, right. 15:23:30 but yeah, bz is waiting for the channel...sorry to rush you 15:23:54 yes, we're *way* over time, we should clear out, Kevin_Kofler, end meeting please 15:24:03 I think we have made our decisions, for any followups, please see #fedora-kde! 15:24:07 Or the fedora-kde ML. 15:24:14 #endmeeting