14:02:16 #startmeeting KDE SIG -- https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2010-01-012 14:02:16 Meeting started Tue Jan 12 14:02:16 2010 UTC. The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot. 14:02:16 Useful Commands: #action #agreed #halp #info #idea #link #topic. 14:02:40 #chair Kevin_Kofler than jreznik svahl 14:02:40 Current chairs: Kevin_Kofler jreznik rdieter svahl than 14:03:22 #topic roll call 14:03:31 who's present today? 14:03:31 #meetingname kde-sig 14:03:31 The meeting name has been set to 'kde-sig' 14:03:35 Present. 14:03:45 here 14:04:05 present 14:04:17 ping: than, jreznik 14:04:39 hello 14:04:43 * than is present 14:04:52 I'm here 14:05:21 #topic agenda 14:05:33 agenda's pretty light today, any other topics to add ? 14:06:09 if there's time, we can continue discussion from last weeks "requirements for default applications" topic 14:06:13 status of plasma-netbook-tracker? 14:06:14 add enablinf JIT 14:07:04 added 14:07:42 #topic kde-l10n srpm split 14:07:48 than ? 14:08:19 the kde-ö10n srpm is too big 14:08:46 it's not good for the update 14:09:03 we should split the srpm out 14:09:09 It's better than having a bazillion SRPMs. 14:09:22 It's a PITA to file a Bodhi update with dozens of packages. 14:09:27 I agree, the current situation is a mess hard to maintain and especially to update 14:09:39 The current situation is the lesser evil. 14:09:40 Kevin_Kofler: it's more work in Bodhi update 14:09:43 how big is single srpm? and single update? 14:09:43 OTOH, Kevin is right with bodhi as well :) 14:09:55 there's obviously pros/cons here , but I tend to agree with than. 14:10:05 than: It's more work if split. 14:10:08 problem being is that a problem in any single locale kills the whole build 14:10:16 Currently we only need to list kde-l10n once. 14:10:25 and adding/removing locales is painful too 14:10:27 isn't there a possibility to have a meta-package in bodhi comprising all the kde-l10n? 14:10:31 No. 14:10:35 too bad 14:10:36 Bodhi doesn't support metapackages. 14:10:47 You have to list all the SRPM NVRs you want to update. 14:11:11 one easy script can help... 14:11:20 but it's going to be hell big update 14:11:35 Looks like the only con is with the bodhi updates. 14:11:36 so, downside is that we have more packages (obviously), upside is that we have far more flexibility for building, fixing things, adding (and removing) locales 14:11:53 I don't see the upside at all. 14:12:00 There'll be a lot of duplication in the specfiles. 14:12:16 And it'll just generally be more work for no benefit. 14:12:16 ye, I'd say having separate SRPMs is actually the lesser evil 14:12:26 One SRPM is the lesser evil. 14:12:33 Kevin_Kofler: there's befenit by update 14:12:36 IMO the pros outweigh the cons 14:12:38 All the tarballs are built the same way. 14:12:50 than: We won't be updating single locales. Why would we? 14:12:53 you don't have to update the whole kde.l10n 14:12:55 Upstream releases all of them at the same time. 14:12:58 or, there's a third solution 14:13:04 SUSE does it with a script 14:13:08 from one SRPM 14:13:08 If there's a new version for one locale, there's a new version for all locales. 14:13:22 ltinkl: That's what we're doing too, isn't it? 14:13:50 using a script/template is likely a good way forward (heck, that's precisely what I did with separate srpms in kde-redhat ages ago) 14:13:54 Kevin_Kofler: yep - all source tarballs are released together... 14:14:09 So why would we benefit from having separate SRPMs? 14:14:12 rdieter: scripting is the must 14:14:27 If a language doesn't build, we need to fix it anyway. 14:14:33 * ltinkl goign to look into OBS 14:14:34 And it's upstream's fault, really. 14:14:41 Kevin_Kofler: true, but at least this way we don't have to block on it 14:14:53 standalone packages gives us some flexibility but as Kevin_Kofler pointed I don't see that "easier updates" reason 14:14:58 That's just sweeping issues under the rug. 14:15:00 I hate that. 14:15:06 Just fix the bug. 14:15:15 it's not sweeping, it's prioritizing. 14:15:29 I can look into the current Catalan mess tonight or tomorrow if needed. 14:15:50 any other pros/cons, and does everyone understand them? 14:16:01 (But why the "translation coordinator" can't make his own language build is beyond me.) 14:16:57 Kevin_Kofler: because he's spanish, sorry catalan? :) 14:17:33 ie, are we ready to vote to enable this or not? 14:18:43 -1, splitting kde-l10n only gives us a lot more work with little to no practical benefit. 14:19:07 It's already split into subpackages. 14:19:14 I'm undecided, the status quo must be improved but I'm not sure splitting is the right way 14:19:45 0 (no really opinion here) 14:19:48 +1 for splitting 14:19:50 +1 14:20:08 Separate SRPMs mean we have to run make build 50+ times. 14:20:12 * thomasj_ late in, just back from kindergarten 14:20:22 And fill in 50+ names into Bodhi each time. 14:20:24 * rdieter forgets who's all vote eligible... :( who's #5 ? 14:20:28 +1 for split 14:20:37 ah. :) 14:20:47 jreznik: 14:20:49 ltinkl: What's wrong with the status quo? 14:20:52 sorry, my dad called me 14:21:11 I'm fed up of all that "the status quo is bad" complaining with no actual practical issue. 14:21:17 Kevin_Kofler: it's hard to update when a new version comes 14:21:27 Splitting only makes it even harder. 14:21:42 ye :/ 14:21:42 Kevin_Kofler: we described the practical issues, you just don't agree with them. there's a difference. :) 14:21:49 You have to do the same work plus do some stuff over 50 times you now have to do once. 14:21:54 it's not easily revertible thing 14:21:59 I agree on that 14:22:13 I propose - if someone prepare scripts 14:22:15 it's very revertable, why not ? 14:22:18 bodhi plus building will be worse 14:22:31 making updates/maintaining easier, I'm +1, otherwise I don't like it now 14:22:45 rdieter: Because we'll have tons of dead packages in CVS, pkgdb and Bugzilla. 14:23:01 We already have enough dinosaurs in Bugzilla confusing users (e.g. kdebase4). 14:23:02 yep 14:23:04 true, that's all part of "more packages" 14:23:30 I wonder if we can poke some folks to see if we can get bz components for EOL'd releases removed 14:23:38 so, can someone try to write the script - if it brings less overhead then, I'm for... 14:23:49 rdieter: good idea 14:24:08 well, looks like a majority is in favor of splitting, 14:24:15 #agreed kde-l10n srpm split 14:24:18 rdieter: I don't think that's possible, what will happen to existing (possibly closed) bugs? 14:24:35 Kevin_Kofler: good point, it's worth asking about though 14:24:46 rdieter: Uh, we're 9 people on the steering committee, aren't we? 14:24:51 anyone want to work on a template/script for this ? 14:24:52 Kevin_Kofler: that x4 could be changed to x component? 14:25:00 Kevin_Kofler: 7 14:25:16 Hmmm, who was on the list again? 14:25:18 I have some experience doing it back in the kde-redhat/kde3 days, it's not hard 14:25:48 In any case, I see only 3 +1 votes. 14:25:51 Kevin_Kofler: 3 people from RH - so all RH people working on KDE :D + you, Rex, SMParrish, svahl I think 14:26:15 I thought we had only 5, it's been awhile since we've had to vote 14:26:30 rdieter: no, 7 14:26:34 7 was the number 14:26:36 ok, sorry. 14:27:21 jreznik: Yeah, that was the list. 14:27:40 alright, no pass then. 14:27:45 So OK, we have +1 from than, rdieter, SMParrish, 0 from svahl, -1 from me. 14:28:01 and ltinkl: 0 (undecided, so far) 14:28:02 ltinkl, jreznik? 14:28:11 zarro 14:28:19 jreznik voted +1 14:28:33 [08:22] making updates/maintaining easier, I'm +1, otherwise I don't like it now 14:28:39 unless I'm misinterpreting that... 14:28:39 Kevin_Kofler: as I said - in case of good supporting script I'm +1 14:28:44 that's four +1's 14:28:44 Well, +1 contingent on scripts being there to make it easier. 14:28:55 There are no such scripts yet. 14:29:06 So that's not a vote for doing that split now. ;-) 14:29:19 I'll revive the sh script I had before. 14:29:21 I also don't see how scripts are going to solve the real issues. 14:29:31 I agree, having a template/script is a pre-req 14:29:33 Koji is going to be loaded a lot more with the many builds. 14:29:49 Also because it needs kdelibs-devel and its huge dependencies in the buildroot. 14:29:50 since we'll have to do a bunch of reviews for the new packages 14:29:57 (to build the translated docs) 14:30:06 Kevin_Kofler: at least - script would help us to maintain it - without it, it's unmaintainable 14:30:13 #action rdieter to work on kde-l10n template/script(s) 14:30:20 And for the Bodhi update, the issue is not just generating the list. 14:30:22 It 14:30:31 It's also that Bodhi gets slower the more SRPMs are in the update. 14:30:43 #topic plasma-netbook tracker 14:30:50 It's going to take forever and a year to process the update requests. 14:31:13 hrm... I don't recall this (netbook tracker) getting done, not by me anyway, anyone else ? 14:31:14 I think this is a huge step backwards and I don't see what we're gaining. 14:31:24 (the kde-l10n split) 14:31:59 bug #549746 14:32:01 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=549746 medium, low, ---, than, ASSIGNED, Tracker for plasma-netbook related issues 14:32:11 * rdieter yays 14:32:23 I just see my serious technical concerns having been ignored and arguments having been used despite me basically proving them void. 14:33:09 We already take too long to get a new KDE out.# 14:33:18 (Upstream usually already has the next version at that point.) 14:33:26 (or almost) 14:33:48 Kevin_Kofler: no one is discounting your concerns, they are indeed valid, but we also have to weight that against the other issues raised (whether you agree with them or not) 14:34:13 Kevin_Kofler: that's another issue we should solve - we need update process, to know who's working on what, not just mess someone is building that, someone committing there... 14:34:37 Re plasma-netbook, I think we need someone to volunteer to test and fix things if we want to support it officially. 14:34:48 Kevin_Kofler: indeed 14:35:08 Basically a spin maintainer for KDE Netbook, just like svahl signed up for the KDE Desktop spin. 14:35:23 I don 14:35:26 #help need volunteers to lead plasma-netbook feature, testing, offer feedback 14:35:26 * jreznik volunteers as he is netbook user 14:35:34 I don't think a dual-use spin makes that much sense. 14:35:37 * mefoster too 14:35:44 (helping, not leading) 14:35:45 I could help there too 14:36:04 for netbook we need lightweight spin 14:36:32 it does not target common uses but really constrained uses - web, mail... 14:36:33 OpenOffice anyone? :) 14:36:43 yeah, I'd propose baby step #1, perhaps create a separate comps group for kde-netbook , then eventually consider a separate spin 14:36:50 The question is, do you want that netbook interface or do you want a "netbook as mini computer" interface (i.e. plasma-desktop)? 14:36:53 rdieter: +1 14:36:54 BTW: is plasma-netbook considered "official" or is it just a tech preview? 14:37:01 I have actually used eclipse and openoffice.org on my Aspire One -- although with an external monitor :) 14:37:30 Did Eclipse allow you to enter more than a character per minute? ;-D 14:37:53 Kevin_Kofler: netbook interface... if I want to use it as netbook - small computer to access net 14:38:05 so, is jreznik volunteering to lead the kde-netbook efforts? 14:38:13 that was the first EEE and then they killed it with WXP 14:38:48 rdieter: I'm interested in and with help from svahl, mefoster we can have nice KDE based system for netbooks 14:38:59 ok, fair enough. 14:39:24 #action jreznik, svahl, mefoster to help work on kde-netbook support, testing 14:39:45 jreznik: So start by having a look at the issues currently on the tracker? :-) 14:39:47 #topic re-test enabling qt/webkit jit 14:39:48 I have promise to accept it as spin from spins people :D 14:39:50 btw: plasma-netbook is just useless without kdeplasma-addons. but the latter has kdeedu-marble as a requirement which is quite big 14:39:59 Some might just be missing packages (e.g. kdeplasma-addons). 14:40:18 svahl: sure,that's an issue in need of resolving somehow 14:40:22 svahl: Hmmm, kdeplasma-addons-netbook subpkg? 14:40:54 Kevin_Kofler: would be a start 14:41:12 netbook spin would need some "intelligent" cuts/splits/reorganization of packages 14:41:21 svahl: much of the netbook efforts will be around where/how to do prudent package splitting 14:41:32 jreznik: exactly 14:41:33 That said, Marble is cool to have on a netbook, with mobile Internet it allows you to use the netbook as a universal city map. 14:41:55 But whether that's enough to warrant it being installed by default is another question. 14:42:21 The default will have to be minimal for those devices with extremely small SDDs. 14:43:01 Re the QtWebKit JIT: 14:43:03 https://bugzilla.redhat.com/show_bug.cgi?id=549994 14:43:04 Bug 549994: medium, low, ---, than, CLOSED RAWHIDE, kwin crashes when enabling plasma-netbook workspace and selinux=enforcing 14:43:05 back to /topic , from our discussions around qtwebkit/jit , it appears to indeed still be an issue. 14:43:13 This is why I redisabled it. 14:43:14 Kevin_Kofler: it was problem with first netbooks but indeed - it should be less space consuming... 14:43:23 It turned out to be due to QtWebKit. 14:43:45 we should contact upstream what they thought 14:43:54 (plasma-netbook uses something WebKit-based by default) 14:43:55 I'm surprised that only happens on rawhide, so could be due to it's (rawhide's) more restrictive selinux policy (vs. more-relaxed policy of releases) 14:44:16 Maybe it's arch-specific? 14:44:21 who wants to own this issue, and work on poking upstream ? 14:44:27 Is the JIT already enabled on x86_64? 14:44:32 Kevin_Kofler: it's arch specific, JIT is enabled only on few archs 14:44:38 Kevin_Kofler: ah, could be. 14:45:04 If it's always disabled on x86_64, then there'd be no wonder you can't reproduce it (and it'd just be another reason to keep it off, we don't want 32-bit to be faster than 64-bit ;-) ). 14:45:44 4.6 has some code for x86_64 JIT, but I don't know if it's enabled yet. 14:46:08 #help need to test/evaluate qtwebkit/jit, and poke upstream as necessary 14:46:28 FWIW, this is a general WebKit issue and also hits webkitgtk. 14:46:48 (Chromium does its own thing for JS, see V8.) 14:47:22 (And I don't know how the V8 JIT reacts to SELinux, but that's not our problem.) 14:47:31 moving on... 14:47:34 it's enabled only on 32 bit, my home laptop is 32 bit 14:47:44 #topic #554713: Only "Internal audio analog stereo" present after login 14:47:49 bug #554713 14:47:51 Bug https://bugzilla.redhat.com/bugzilla/show_bug.cgi?id=554713 medium, low, ---, than, NEW, Only "Internal audio analog stereo" present after login 14:48:15 I asked there what version of kdemultimedia was used, the updated kmix/pa patch in 4.3.90-2 is a bit better 14:48:29 I've tested this during the meeting with kdemultimedia-4.3.90-2 on anohter machine: same behaviour 14:48:48 so... it's picking the wrong pa device? 14:48:55 ie, doesn't work ? 14:48:59 Uh, this has nothing to do with KMix at all. 14:49:06 It's the Phonon KCM at work there. 14:49:09 yes, only the internal audio device i shown 14:49:19 i=is 14:49:26 oh, I thought the report mentioned what's seen in kmix (sorry) 14:49:45 Systemsettings → Multimedia is the Phonon KCM, not KMix. 14:50:30 duh, nevermind me then. -ENOTENOUGHCOFFEE 14:51:11 svahl: does it happen in F12? 14:51:31 than: I haven't testet that, yet 14:51:59 could it be a wrong starting order of pulseaudio and start-pulseaudio-kde (due to the log entries in messages)? 14:54:18 looks like we're stumpted for now, until we can do some more debugging and reproduce it I guess 14:54:49 #topic open floor 14:55:02 we've about 5 min, left, any parting words ? 14:55:13 rdieter: it should be reproducible with the next nightly images 14:55:33 could be rawhide specific, I don't see that on my f12+kde-4.3.90 box 14:55:36 f12, missing my intel hda sound chip. It offers only internal sound and my ati sound chip. If that helps. 14:55:45 kde 4.3.4 14:56:32 I'll take the default apps requirements, http://fedoraproject.org/wiki/User:Rdieter/FUDCon_Toronto_KDE_F13_TODOs , topic to ml for further discussion 14:56:41 ok 14:56:46 I think this voting procedure is actually encouraging rushed and bad decisions. Previously we tried to rediscuss things until we got a full consensus, now we just vote and rush things through with too little discussion. 14:57:01 Kevin_Kofler: +1 14:57:03 :) 14:57:25 oh, I thought we had clearly heard everyone's point of view, and opinion already. Did we miss something ? 14:57:31 There's no need to rush changes like the kde-l10n split and I think the unwanted side effects will bite us in the rear. 14:57:55 tell ya what, let's reserve judgement until we have the template/scripts, then we can rediscuss and vote again 14:58:16 I think if we want vote it should be 2 rounded (one meeting propose, then discuss on ml - and get more replies from community, and second meeting to vote) 14:58:51 FYI: progress continues on sesame2 (although it'll soon be obsolete by virtuoso, I fear ...) 14:59:00 akurtakov is being very helpful 14:59:12 rdieter: yep - prepare scripts, prove it's possible, it's a good solution, everyone can think about it in quiet place (toilet :D) and then vote :) 14:59:14 mefoster: thanks, tis true that kde-4.4 nepomuk will use virtuoso by default. 14:59:50 Well, this process is shaking out a lot of old Java packaging stuff anyway ... 14:59:54 mefoster: how many reviews remain? (rough estimate) 15:00:19 Roughly? Lots ... 15:00:24 heh, ok. 15:00:56 mefoster: I can help with reviews but I'm not java expert 15:00:56 #help reviewers needed for remaining java/soprano-sesame stack 15:01:00 Although once one more package gets through (akurtakov currently reviewing), everything is pretty small and the specs are nearly cut-and-paste identical so should go fast 15:01:23 thanks everyone, time's up. 15:01:26 #endmeeting