15:05:16 #startmeeting kde-sig 15:05:16 Meeting started Tue Jul 22 15:05:16 2014 UTC. The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:05:16 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:05:23 #meetingname kdesig 15:05:23 The meeting name has been set to 'kdesig' 15:05:28 #meetingname kde-sig 15:05:28 The meeting name has been set to 'kde-sig' 15:05:40 #topic roll call 15:05:41 * jgrulich is present as usual 15:05:45 hi all, who's present today? 15:05:50 present 15:05:56 Present. 15:06:06 present 15:06:10 * jreznik is semi-here 15:06:53 #info rdieter jgrulich dvratil Kevin_Kofler than jreznik present 15:07:00 #chair jgrulich dvratil Kevin_Kofler than jreznik 15:07:00 Current chairs: Kevin_Kofler dvratil jgrulich jreznik rdieter than 15:08:43 #topic agenda 15:08:50 alright, what to discuss today? 15:09:02 me (present) 15:09:12 #info pino|work present 15:09:14 I can give a status update on GStreamer 1.0. 15:09:16 pino|work: hi 15:09:20 *1.x, I mean. 15:09:51 (Their versioning is ugly, they have libraries with "1.0" in the name and a version number of 1.2.) 15:10:57 that's the ABI version 15:11:19 we can discuss later :) I can do a quick kde-4.13 status update too 15:11:34 dvratil: any kf5/plasma5 news to mention? 15:11:48 yeah, I can do a quick update 15:12:04 #topic gst-1.0 status update 15:12:08 Kevin_Kofler: ? 15:13:28 So the status is that I have updated 2 packages in Rawhide (F22), 5 to go, and the F21 builds to go as well. 15:13:56 hi 15:14:13 I updated QtWebKit (4) with the openSUSE patch to enable GStreamer 1 support, and conditionalized it on F21+/RHEL8+. 15:14:41 (We don't want to build the Qt stack with mixed GStreamer 0.10/1.x on e.g. F20 or RHEL 7, that will cause trouble for sure.) 15:14:48 tosky: hi 15:14:51 #info tosky present 15:15:12 (Unfortunately, the patch is such that it has to be conditionally applied, they didn't go to the trouble of autodetecting the GStreamer version.) 15:15:38 I also updated Phonon-GStreamer to a snapshot from the 1.0-porting-for-merge branch. 15:16:05 sorry, afk ~5 min, feel free to move on to next topic when/if ready before I get back 15:16:19 just fyi: we are going to make an official release of phonon-gstreamer-1 in Randa in couple weeks 15:16:21 The snapshot has the GStreamer 1 porting, and is otherwise equivalent to 4.7.2. 15:16:44 It supports both building against GStreamer 0.10 (where it should hopefully behave the same as stock 4.7.2) and against GStreamer 1. 15:17:04 There too, I added conditionals for enabling GStreamer 1 support only where appropriate (F21+/RHEL8+). 15:17:37 What is missing is the QtGStreamer and KDE-Telepathy stack. (It's really 2 stacks, but they intersect.) 15:18:15 There are 5 packages to update there (and at least for some of them, the specfiles will be GStreamer-1-only, so we'll have to be careful with merging). 15:19:32 I also looked into nucleo's complaint about the weird soname versioning of QtGStreamer: IMHO, we should follow upstream there, the naming is weird, but not broken. 15:20:15 Concretely, the complaint was about the *-1.0.so.0 → *-1.0.so.1.2.0 symlinking (.0 vs. .1.*). 15:20:40 not "intuitive", but still fine 15:20:41 agreed 15:20:45 It turns out that this will work just fine, the soname was bumped through the addition of that "-1.0" part, and it's fine to start at .so.0. 15:22:03 And speaking of weird versioning, here too, we have the use of "-1.0" for 1.2.x versions as in GStreamer itself. This practice is also used in other libraries of the G* stack. It's counterintuitive, but we have to live with that. 15:22:25 version != ABI version 15:22:32 I suppose they use that ".0" so that they can decide when they do 1.2 whether they want to change the ABI (and thus change to "1.2") or not. 15:22:47 pino|work: Well, the thing is, "1" would be enough as the ABI version. 15:22:52 (Similarly for GTK+.) 15:22:57 apparently not enough for them 15:23:30 so if a new version of the 1.x serie breaks the abi, they can bump the abi without issues 15:23:40 They just don't have the balls to fix in advance that the ABI will never change throughout a major version, even though that's the goal (and so far was always the case for those GTK+ etc. libraries, GStreamer 0.x was different). 15:23:58 Right. Then it should be a 2.x version and not a 1.x one. :-) 15:24:13 as long as it's blue 15:24:13 But that's just IMHO and a bit off topic here. ;-) 15:24:59 Oh, and about QtWebKit, if you were wondering about qt5-qtwebkit, that one is already built against GStreamer 1 wherever it is available. 15:25:28 So, to sum up: 15:25:28 , I'd double checked that prior 15:25:51 #info qtwebkit and phonon-gstreamer updated to build against GStreamer 1 on F21+ (and RHEL8+). 15:26:38 #action Kevin_Kofler still has to update 5 packages (QtGStreamer and KDE-Telepathy stacks), and the F21 builds of all 7 packages remain to be made. 15:26:53 I think that's it, or are there any questions? 15:28:07 Let's move on, rdieter? 15:28:12 ok 15:28:18 #topic kf5/plasma5 status 15:28:20 dvratil: ? 15:28:55 KF5 is done completely, Plasma 5 is almost done, now waiting for Copr to start cooperating again and build the last few of them 15:29:17 if it goes as expected, it should be done tonight 15:29:42 dvratil: its becoming a FAQ, we should probably add something to #fedora-kde /topic about it 15:29:54 (not that anyone reads that much, but still :) ) 15:30:00 +1 - maybe link to the Plasma copr? 15:30:38 either that, or a link to something that explains it all, that also includes such a link 15:31:03 there's upstream wiki page that documents availability of KF5/Plasma 5 in various distributions 15:31:05 I'm a bit annoyed by how users are always pestering us for the bleeding-edge *.0.0 KDE Plasma releases; yet, if we ship them, they're complaining about that, too. 15:31:15 We've had that with Fedora 9. 15:31:18 ship it! 15:31:20 where explanation includes why f20/f21 won't include this in stable repos, and its in a copr instead 15:31:28 if I manage to find it, I'll update it and we can link it that page 15:31:42 ok 15:31:46 great 15:32:08 dvratil: https://community.kde.org/Plasma/Packages 15:32:19 thanks jgrulich 15:33:15 jgrulich wins, beat me to it 15:33:16 Can we provide those for F19 also? I guess the KF5 stuff would have to come from a Copr there though, with F21 branched, I assume F19 isn't open for new branch requests anymore, is it? 15:33:34 (We should have requested those right when we did the KF5 reviews. :-( ) 15:33:51 most kf5 doesn't have f20 branches either, iirc 15:33:59 none do 15:33:59 Indeed. 15:34:11 I'd have expected the requests to obviously include all branches, so I didn't check. :-( 15:34:17 as (if I remember it correctly) we agreed that we will keep F20 in Copr 15:34:25 so I was only requesting rawhide 15:34:34 There's no reason we don't want to ship those parallel-installable packages for stable Fedoras. 15:35:05 We should definitely request at least F20. (I think it's too late for F19 unfortunately.) 15:35:09 I guess I can do some kind of mass-request for all kf5 to include F20 branch 15:35:29 and build F19 in Copr 15:35:37 Yeah. 15:35:45 I'm just afraid that F19 might have too old cmake 15:36:01 It's easy to upgrade CMake in Copr. :-) 15:36:06 f19/f20 cmake is the same, fyi 15:36:16 cmake-2.8.12.2-2.fc19, cmake-2.8.12.2-2.f20 15:36:16 ok, then it should be fine 15:36:39 but I would agree if there are priorities, f20 is more important, then f19 if time allows 15:37:14 +1 15:37:20 Newer releases first. 15:37:26 So F22, F21, F20, F19 always in that order. 15:38:39 hmm, there does not seem to be any UI or CLI tool to mass-request branch 15:39:35 true, the SOP is still per pkg in bugzilla, unfortunately 15:40:04 anything else? move on? 15:40:05 Talk to #fedora-admin, maybe? 15:40:22 , worth asking anyway 15:41:02 ok, moving on... 15:41:11 #topic f20/kde-4.13.x status 15:41:28 nothing much to report, other than continued positive feedback 15:43:03 last batch of builds was pushed to -testing on jul 19, so I'd say wait a bit more before pushing stable 15:43:52 maybe late this week, or early next 15:45:03 that's all I have 15:45:07 #topic open discussion 15:45:12 anything else for today ? 15:46:12 Is there any harm to upgrade cmake to 3 ? 15:46:19 on f19 > i mean 15:46:45 Are there things that need CMake 3 to build? 15:47:21 No necessarly but have same cmake on all platforms make sense 15:48:02 I'd figure it *mostly* safe, I'm not aware of any issues with it 15:48:12 Upgrading CMake to a new major version on F19, I'm not sure that's a good idea. Even for F20, you'll have to convince the primary maintainer, who tends to be a bit more conservative than me about upgrading stuff. 15:48:31 (Orion Poplawski is the primary maintainer for CMake.) 15:48:53 Well, just an idea 15:49:18 But we can get it in for F20 (and maybe even F19) if we really need it. 15:49:58 It's always a tradeoff about allowing to build new stuff vs. the risk of breaking the build of existing stuff. 15:50:06 CMake is quite good at backwards compatibility, luckily. 15:50:10 I know 15:50:12 But there can still be surprises. 15:53:41 * rdieter will close meeting soon, if theres nothing else 15:54:12 By the way, if you're wondering why I got only 2 of the 7 updates for GStreamer 1 done, blame the weather. :-) (Too hot to spend time on packaging.) 15:54:33 This week should be more suitable for getting stuff done. 15:56:53 Kevin_Kofler: :) 15:56:56 ok, thanks everything 15:57:03 everyone, for everthing, that is 15:57:07 #endmeeting