15:04:20 #startmeeting KDE SIG Meeting 15:04:20 Meeting started Tue Jul 29 15:04:20 2014 UTC. The chair is Kevin_Kofler. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:04:20 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:04:24 #meetingname kde-sig 15:04:24 The meeting name has been set to 'kde-sig' 15:04:26 #topic Roll call 15:04:27 * jgrulich is present 15:04:30 * Kevin_Kofler is present, who else? 15:05:03 * tosky here 15:06:09 present 15:07:20 danofsatx-work, dvratil, jreznik, mbriza: Ping? 15:07:26 hi 15:08:08 I'm here, was waiting on you 15:08:44 had to go calm a user for moment 15:08:53 here 15:08:59 * jreznik is around but pretty busy 15:09:02 ping me if needed 15:09:38 hi, semi-present 15:10:13 #chair jgrulich tosky than dvratil danofsatx-work pino|work jreznik mbriza 15:10:13 Current chairs: Kevin_Kofler danofsatx-work dvratil jgrulich jreznik mbriza pino|work than tosky 15:10:39 #info jgrulich, Kevin_Kofler, tosky, than, dvratil, danofsatx-work, pino|work present, jreznik and mbriza semi-present. 15:10:44 #topic Agenda 15:11:32 proposing splitting kde-runtime a little to improve coinstallability with Plasma 5 15:11:34 #chair ltinkl 15:11:34 Current chairs: Kevin_Kofler danofsatx-work dvratil jgrulich jreznik ltinkl mbriza pino|work than tosky 15:11:42 #info ltinkl present (late). 15:12:12 There's officially welcoming danofsatx-work as voting WG member, replacing the inactive Siddharth Sharma, that was proposed before. 15:12:39 oh, welcome :) 15:12:44 sorry, I have to go so I won't be available 15:12:45 welcome aboard 15:13:01 thank you gentlmen, it is my honor to contribute to the best DE on Fedora. 15:13:05 There should be a vote, really. :-) Now we have the agenda though. 15:13:10 danofsatx-work: welcome! 15:13:13 Is there anything else to discuss this week? 15:13:15 welcome! 15:13:33 Kevin_Kofler: can you please update the wiki pages? 15:14:30 #topic Proposal: New WG composition: Dan Mossor (danofsatx-work) replaces the inactive Siddharth Sharma 15:14:41 So, can we formally vote for or against the proposal? 15:14:43 +1 15:14:59 And yes, welcome on board! 15:15:13 +1 from me too 15:15:15 +1 15:15:17 +1 15:15:18 +1 15:15:34 +1 15:15:45 That's +6, i.e. enough. :-) 15:15:54 woot! 15:15:58 * danofsatx-work bows 15:16:06 #agreed The proposal passes (+6, -0, 0 abstentions, 5 absent), welcome on board Dan! 15:16:29 Thank you, again. I am truly honored. 15:17:06 So with the new WG composition, we should reconsider the default browser issue. :-) – Look here: https://fedoraproject.org/wiki/Talk:Fedora_Plasma_Product/Integration – Firefox no longer has a majority. 15:17:20 (17:17 now) 15:18:20 I am for no FF. at all. 15:18:34 Me too. 15:18:44 Firefox be gone! (Ideally from Fedora as a whole!) 15:19:07 I'm still undecided tbh, but do we have to go thru this now? we have more important stuff to do than this imo 15:19:33 however, there are some (many?) features missing from the Rekonq and konquerer options. 15:21:04 What I see is that Rekonq and Konqueror both actually have MORE features than stock Firefox (without third-party, non-Fedora, sometimes even non-Free, extensions). 15:21:10 E.g. builtin ad blocking. 15:21:33 yes, native OS usability is outstanding. 15:21:38 Also kioslave support. 15:21:46 however, I'm also with ltinkl - let's shelve this for F22. 15:22:11 Shelving only makes sense if we ship F21 with what we have always shipped (Konqueror). 15:22:25 F21 won't even be a Plasma Product release, but a KDE Spin one as always. 15:22:51 well, I haven't tried a working live TC yet to know how well everything works. it's on my todo. 15:23:14 there were some hints to upstream mailing lists about mozilla devs not being totally closed about patches, if someone is willing to maintain them 15:23:20 if you look at the technical stuff (https://fedoraproject.org/wiki/Fedora_Plasma_Product/Technical_Specification), there's still a lot of gaps 15:23:33 Installation methods and media 15:23:34 and 15:23:41 Hardware requirements 15:23:43 mainly 15:23:48 we should decide on those 15:24:02 what size/media will we target? 15:24:11 what HW will we support? 15:24:18 32? 64 bit only? 15:24:21 etc 15:24:59 and, while it's probably still a bit too early... Plasma 5? :) 15:25:13 too early 15:25:14 next :D 15:25:18 given that there's still plenty of time before F22 15:25:30 Media: live image, I guess CD won't be possible (I'm having a hard time trying to get a CD-sized Kannolo going even ripping off stuff like SELinux), so I guess we'll stick to 1 G(i?)B. 15:25:45 ye I know it's too early NOW but we must think like several months ahead 15:25:49 Installation: Path of least resistance = stick to liveinst from Anaconda. 15:26:09 agree with liveinst 15:26:10 A couple KDE distros are working on a cross-distro Qt 5 live installer though, let me dig up the link. 15:26:17 change #topic? 15:26:20 * satellit FYI last f21 boot.iso had KDE 0726 15:26:26 size... dunno, 1 or 2 gigs is fine I guess 15:26:34 #topic Plasma Product Discussion 15:26:42 #topic Installation methods and media 15:26:45 sry :) 15:26:48 That could be an alternative for the future (especially if Anaconda developers follow through on their threat of dropping liveinst). 15:27:17 yes 15:27:21 * ltinkl likes that 15:27:34 (but even without Anaconda doing that, it could be an option) 15:28:01 * Kevin_Kofler would love a GTK+-free Plasma Product by default (especially now that we moved the Scientific parts to an add-on and are focusing on Plasma again). 15:28:52 ok, let's vote in the install methods and media? 15:29:08 so that we move the technical docs somewhere 15:29:35 ok, I'm here again, did I miss something? :) 15:30:18 so here's the vote proposal: For the time being, we will be using anaconda's liveinst and targetting 2 GB media 15:30:30 Darn, I can't find the link again. 15:30:32 +1 from me obviously 15:30:35 Why 2 GB? 15:30:40 The target has been 1 so far. 15:30:50 hmm 15:31:02 will we safely fit in there? 15:31:12 (perhaps yes without the sci stuff) 15:31:42 and will there be a way to choose what to install? 15:32:16 Kevin_Kofler: i agree with ltinklm +1 fir 2GB 15:32:20 Ah, there's my link to the installer: 15:32:22 #link https://github.com/calamares/calamares 15:32:48 #info (not ready for production yet, to be investigated in the future) 15:32:51 I think if we don't fit to CD then it doesn't matter if we have 1GB or 2GB 15:33:21 ye, if we leave out the CDs, I don't think it matters if it's 1 or 2 GBs 15:33:45 There might be people with a bunch of 1 gig usb sticks around though. if the size is 1gig, it'll still fit on their old stucks 15:33:50 the "official" Fedora swag USB sticks are 2GB. At least the one I got from Masta at Texas Linux Fest last month was. 15:33:51 better have some spare space left, we can ship extra stuff at least 15:33:51 s/stucks/sticks 15:33:56 Also download sizes matter. 15:34:00 ltinkl: ++1 15:34:32 Installing an image with only the stuff most people will need and then adding the extra stuff is nicer than having a huge image to download with lots of stuff the user doesn't need. 15:34:33 so 1 or 2? :) 15:34:46 Kevin_Kofler: that's true as well 15:34:54 +1 for 2GB 15:35:09 I'd say 1, or if we cannot reach that (but without Firefox we should be able to! :-p ), then maybe something like 1.2, 1.5 or whatever. 15:35:21 Less to download, and leaves room for an overlay on a 2 GiB USB stick. 15:35:28 + 1 for 2GB including firefox 15:35:31 1.44GB 15:35:39 is it important to decide it now? I mean, if it turns that 1.3 GB are enough, are we going to fill the disk until 2 GB are full? 15:35:54 mbriza: :D 15:35:54 I don't want us to go CD → 1 GiB → 2 GiB → 4 GiB → DVD → DVD-DL → BluRay just because we can! 15:36:16 there's plenty of stuff you can put on 2gb keeping that as limit 15:36:25 We need to limit the size inflation as much as we can! (Ideally, stop it entirely, but that doesn't only depend on us, there's creeping bloat all over the distro. :-( ) 15:36:31 1.4GB - http://en.wikipedia.org/wiki/MiniDVD 15:37:03 set that as the target, it leaves ~600MB overlay on a 2GB Live USB 15:37:16 danofsatx-work: I like that (if and only if 1 is really not reachable anymore!), it's a hardware-enforced size limit, so it's less prone to people just bumping it again for no reason. 15:37:29 danofsatx-work: good idea, like it 15:37:54 (That was also a big reason I disliked doing away with CD size, I *knew* we'd have this discussion again so soon. :-( ) 15:39:52 so go for 1.4GB, MiniDVD format? 15:40:04 (seems like a good compromise too) 15:40:10 yep, makes sense 15:40:31 Can we really not do 1 anymore? 15:40:48 we'll see in the end I guess 15:40:53 And if yes, why? Just because you want Firefox? Then that's yet another killer reason to NOT ship it! 15:41:20 * ltinkl didn't say he wanted Firefox :) 15:41:25 Firefox drags in a whole extra HTML engine (and a huge one at that), when we already have 2 (!) on the KDE spin. 15:41:35 OK, just because *someone* wants Firefox. :-) 15:41:38 Sorry. 15:41:59 Code duplication that WILL bloat our spin. 15:42:17 That, and integration is why Firefox is bad on a KDE/Plasma image. 15:42:36 Then there are all the reasons why it's just bad to begin with, but we don't need to rehash those now. :-) 15:42:51 next time I see the word "Firefox" in this channel, I'm leaving ;) 15:43:59 Let's move on to the next topic, or we won't get done… 15:44:01 let's move on, I'll put the "1.4GB, MiniDVD in the tech specs" as a tentative target 15:44:03 ye 15:44:10 ltinkl: OK 15:44:14 #topic Hardware requirements 15:44:37 Uhm, that wasn't really on the agenda. ;-) 15:44:57 CPU requirements are defined by Fedora as a whole. 15:45:07 sry, I just moved one bullet down in the tech specs 15:45:11 32-bit i686 = 686 with cmov 15:45:18 ltinkl: before going on with this, wasn't there an agenda proposal from dvratil? 15:45:18 the live respin is gettin up there, 924 megs 15:45:22 x86_64 = base x86_64 (which includes SSE2) 15:45:34 tosky: Right. 15:45:35 tosky: sry, must have missed that, I came a bit too late 15:45:40 The only thing we can tweak is memory. 15:45:45 dvratil: all yours then 15:45:48 And there we should go with upstream recommendations. 15:46:10 #topic Proposal: splitting kde-runtime a little to improve coinstallability with Plasma 5 15:46:14 dvratil: Your turn. 15:46:29 Soooo, in order to have maximum co-installability between KDE 4 and Plasma 5, we need to split two things from kde-runtime: that's khelpcenter (simple app - I have no idea why it even is in kde-runtime), and -docs 15:46:53 here's updated spec file, so if someone can review it after (or during) the mtg: http://paste.fedoraproject.org/121721/46411140/ 15:47:06 dvratil: because applications ship documentation, and it was deemed good to have something for users to read them 15:47:23 (or something like that, if i remember correctly baaaack then) 15:47:24 I'd like to get this to Fedora asap, as that's the last thing that blocks the Copr repo with Plasma 5 to install all packages 15:47:43 pino|work, could be kde-baseapps or something, but that's not relevant atm 15:47:45 dvratil: make sense 15:47:55 khelpcenter is provided standalone now in Plasma 5 15:48:06 so it will go as a smooth update to khelpcenter 4.13.3 15:48:10 I guess we can split it out, make kde-runtime Requires: it, then the Plasma 5 repo can provide a build with a higher EVR. 15:48:11 dvratil: how does it this improve the coinstallability? Just for khelpcenter? 15:48:23 and kio-extras-docs Obsoletes kde-runtime-docs 15:48:48 (assuming khelpcenter from Plasma 5 can read the docs for KDE 4 (and 3 for that matter) apps, otherwise we need another solution (renaming the executable and patching stuff to find it, ugh…)) 15:48:59 I would expect so 15:49:53 it should be able to do it, yes 15:50:09 did the actual base format of the docs themselves change from 3 to 4 to Plasma 5? 15:50:13 I think kdelibs4support could be needed, though 15:50:25 danofsatx-work: it's not plasma-related, and yes, the DTD was updated 15:50:33 arguably we could split the kde-runtime package a bit more (-kcm, -kio, kglobalaccel, kde-cli-tools) - but the question is whether that makes sense atm 15:50:39 ok, but is it backwards compatible? 15:50:41 dvratil: kde-runtime has more docs than just the kio-extras ones. 15:50:55 Kevin_Kofler, khelpcenter installs it's own docs 15:50:56 danofsatx-work: backward compatibility is provided in kdelibs4support 15:51:01 ok 15:51:10 * danofsatx-work goes back to his corner 15:51:31 And does it still work for KDE 3 docs? IIRC, the KDE 4 khelpcenter can view those just fine. 15:51:52 dvratil: wouldn't kdelib4support clash with kdelibs-4 then? :) 15:52:05 ltinkl, kdelibs4support is coinstallable with KDE 4, it's a framework 15:52:28 dvratil: ok, we will have another clash then, the translations 15:52:36 Kevin_Kofler: I didn't try that in a long time; I think we could have broken some documentation few years ago, so maybe docbook 4.1.x documentation could not work even now 15:52:39 ltinkl, how come? 15:52:42 they get installed into the same dir, most of the times with the same name too 15:52:55 hmm 15:53:04 only on the applications level 15:53:06 not in Frameworks 15:53:11 some tools got renamed (like kreadconfig->kreadconfig5) 15:53:12 (the names there are different) 15:53:13 ye, apps 15:53:15 I can view the Quanta Plus docs in khelpcenter, but there are strange error messages on the console. 15:53:23 But the stuff shows up. 15:53:34 dvratil: ye the frameworks are fine, but not the apps 15:53:34 ant that could be solved by splitting the KDE 4 translation packages 15:53:51 dvratil: I don't see how 15:53:53 and the Plasma 5 package would obsolete that specific KDE 4 translaction subpackage 15:53:57 (The errors are about meinproc failing with exitCode 2.) 15:54:05 dvratil: it can't 15:54:08 (meinproc4 actually) 15:54:09 ltinkl, why? 15:54:11 dvratil: let me explain 15:54:20 ltinkl, don't tell me all translations in KDE 4 are in one huuuuuuge file 15:54:21 all the translations are in one RPM 15:54:36 now suppose you have plasma 5 and digikam 15:54:39 from KDE 4 15:54:52 You can do what I did to the KDE 3 kde-i18n. 15:55:01 you can't obsolete KDE 4 translations RPM with the one from Plasma 5 15:55:02 for i in (everything that's in KDE 4) 15:55:05 rm -rf $i 15:55:12 end 15:55:14 ltinkl, you can obsolete kde-18n-digikam-Czech 15:55:18 ah, darn 15:55:24 you would have to obsolete all langauges manually 15:55:25 there's no such thing... 15:55:39 ltinkl, that's why I said we would split kde-i18n-$language packages 15:55:40 all the translations for one language are in one RPM 15:55:41 to per-app 15:55:54 that means updating _all_ the KDE 4 apps 15:55:55 have per-application subpackages (where it makes sense) 15:56:02 Kevin_Kofler: meinproc4 fails maybe because of some entity not defined or so; if you enable more debug, run khelpcenter from the command line, and you see more information, please send them to me 15:56:02 why? 15:56:08 The kde-i18n packages are still like that. (These days, they should probably use a whitelist instead, but I never got around to it.) 15:56:48 Distro with Plasma 4 as default: delete all the conflicting translations from Plasma 5, and vice-versa. 15:57:06 That's how we did it for kde-i18n vs. kde-l10n back in the days (and still do for kde-i18n, as I said). 15:57:18 ye... 15:57:27 sound a bit harsh 15:57:30 *sounds 15:57:43 I actually used scripts to get a list of all the translations that conflicted, and then scripted the rm -rf into %install. 15:57:43 you will end up with workspace localized, but most apps in english 15:57:51 tosky: When I get around to it, sure. 15:57:55 oki 15:58:06 dvratil: nope 15:58:40 dvratil: The secret is to delete the translations of the version you don't ship. :-) 15:58:47 dvratil: you just emove stuff that exists in KDE 5 and KDE 4, ie that has a newer (ported) version 15:59:09 I see 15:59:13 (We still do blacklist kdewebdev translations from kde-l10n, by the way.) 15:59:24 (because kdewebdev in KDE 4 is a bad joke) 15:59:33 (no Quanta, nor anything else that does web development) 16:01:17 I'm not saying it's a nice and clean solution, but it's an effective one with the constraints that we have (one big l10n package with per-language subpackages). 16:01:58 I really really don't want to do split l10n packaging. 16:02:18 There are already over 50 subpackages, adding a per-app splitting dimension will make them go in the thousands or worse! 16:03:18 I think it's well over 60 :) 16:03:51 That's mathematically over 50. ;-) 16:04:24 Anyway, doing per-app splits will make them too many! 16:04:58 ye, that would be hell 16:05:17 oh and 16:05:27 the frameworks ship translations bundled :( 16:05:34 Though it could make us beat TeXLive's package count. ^^ 16:05:35 to add some extra complication layer 16:06:07 Didn't they also regress some of the frameworks to using Qt's silly Not Invented Here translation system? 16:06:19 but they won't clash with existing stuff so it doesn't matter 16:06:25 there is no split translation packages, every package will bring the translations 16:06:27 * Kevin_Kofler hates that Qt-only translation system. 16:06:29 like other libraries 16:06:49 even if they are gettext based, this has nothing to do with ts vs po 16:06:51 tosky: But the point is that it makes it harder to blacklist the translations if they conflict with KDE 4 ones. 16:06:52 https://bugreports.qt-project.org/browse/QTBUG-40444 16:07:05 Those .qm ones at least won't conflict. ^^ 16:07:10 (but they suck in other ways) 16:07:11 I understand that, I have no other solutions 16:07:12 the frameworks (the Qt-only ones) have this problem 16:07:30 * ltinkl notes arguing with ossi is like banging head against wall 16:07:40 * Kevin_Kofler knows that feeling. 16:08:02 so ye, frameworks use PO files but they get transformed into QM 16:08:08 and there the fun begins :) 16:08:31 Ewww… 16:08:48 Why don't you just stick to gettext? Everyone has gettext installed, it's part of glibc, darnit! 16:09:07 not all frameworks, luckily 16:09:16 applications are going to use ki18n 16:09:21 ye just the Qt-only ones 16:09:41 Qt-only KDE frameworks shouldn't exist to begin with. 16:09:43 because Qt-only code can't load mo files coming from gettext 16:10:39 anyway, offtopic here, another one on agenda? 16:10:43 * ltinkl has to run in a minute 16:11:27 #topic Flock 2014 16:11:45 #info Next week, there will be Flock 2014 in Prague (Aug 6 to 9, 2014). 16:12:03 I suppose several of us will be there. :-) 16:12:19 Unfortunately, there isn't anything KDE-related on the schedule. 16:12:32 (at least not directly) 16:14:38 Anything interesting KDE SIG / Plasma WG members shouldn't miss? I guess cwickert's Spins discussion/workshop at the very end will be interesting. 16:16:17 Looks like there isn't much more to say on this topic, so… 16:16:19 #topic Open discussion 16:16:27 … anything else? Or can I close the meeting? 16:16:51 Kevin_Kofler: pls add danofsatx-work to the membership, ty 16:16:58 OK, will do. 16:18:06 I gotta run, thanks for the welcome folks. 16:18:27 FYI, I won't be available for the meeting next week. 16:18:45 I'll leave you others to figure out whether you'll be enough to hold a meeting, and if yes, who will lead it. 16:18:46 not sure if I will be, either - it's finals week for me 16:19:48 If you know you won't attend, please say so ASAP so we don't end up with 2 or 3 people wasting their time waiting for everyone else. If you know you WILL be available, that's also good to know in advance, too. Thanks for your cooperation. :-) 16:19:59 That said, thanks all for coming! 16:20:09 #endmeeting