15:16:44 #startmeeting kde-sig -- https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2013-09-17 15:16:44 Meeting started Tue Sep 17 15:16:44 2013 UTC. The chair is rdieter. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:16:44 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:16:51 #meetingname kde-sig 15:16:51 The meeting name has been set to 'kde-sig' 15:17:04 #topic roll call 15:17:14 hi, whos present today? 15:17:21 Present. 15:17:25 present 15:17:25 present 15:17:29 * oddshocks lurking 15:18:25 than_ : ping 15:19:04 #info Kevin_Kofler rdieter mbriza jgrulich oddshocks present 15:19:13 #chair Kevin_Kofler mbriza jgrulich 15:19:13 Current chairs: Kevin_Kofler jgrulich mbriza rdieter 15:19:56 let's just run through the agenda per wiki (see /topic), and then see if we have time at end for anything else 15:20:00 #topic kde-4.11.1 status 15:20:32 let me echo what I sent to kde@fpo list... 15:20:35 As far as kde-4.11.1 goes, status is that it's been in kde-testing for awhile now, getting some good testing. I just submitted it in several pieces to bodhi for f19/f20 yesterday. It seems to be too big to submit as a single piece to bodhi anymore. 15:20:50 several things I wanted to work on : 15:20:52 * enable automatic passwordless-kwallet (for f20+ at least) 15:20:54 * backport https://git.reviewboard.kde.org/r/109609/ 15:21:02 present 15:21:09 #info than_ present 15:21:11 #chair than 15:21:11 Current chairs: Kevin_Kofler jgrulich mbriza rdieter than 15:21:12 #chair than 15:21:12 Current chairs: Kevin_Kofler jgrulich mbriza rdieter than 15:21:47 that's all I have wrt 4.11.1, any other comments/concerns? 15:22:16 F18 builds? 15:23:11 I suppose I could do unofficial f18 builds, unless you're suggesting more than that? 15:23:44 was meaning to do el6 ones sooner or later, could just as well do f18 at the same time 15:24:05 I'm always the one who suggests doing official updates for Fn-1, but I guess I've just opened a big can of worms… 15:25:26 rdieter: you mean 4.11 for el6 ? 15:26:23 Yes, in kde-redhat, where he has 4.10 already built. 15:26:45 than: yes 15:26:46 great ! 15:26:54 It's not eligible for EPEL because it replaces the stock RHEL KDE. 15:27:14 i know 15:29:13 anything else? (otherwise I'll move on to f20 features/changes items... sddm, plasma-nm) 15:31:02 If you do unofficial builds of 4.11 for F18 soon, I can probably test them. 15:31:19 #topic sddm - https://fedoraproject.org/wiki/Changes/SDDMinsteadOfKDM 15:31:20 ok 15:31:24 I'd really like to test the kdepim performance improvements, I hope I won't be disappointed! 15:31:50 mbriza: mind giving a brief update/status-report 15:32:03 wrt sddm progress 15:32:31 so, briefly, there's still some mess regarding pam, resulting in me basically rewriting the whole authenticator class 15:33:22 Are we sure the PAM issues are even SDDM's fault? 15:33:33 not really but other DMs work better 15:33:37 There seems to be some PAM/systemd breakage lurking in F20+. 15:33:49 at least on my machine 15:34:22 Kevin_Kofler: could be 15:35:36 but yeah, having an online session after logging out is definitely wrong 15:35:36 pam_systemd does most of the work for session regristration, no? 15:36:10 well, the current problem is you can't log in again after logging out, i'm not sure if this is directly related to pam_systemd or any other pam module 15:36:16 * rdieter forgets, spoiled by it "just working" for so long 15:36:22 mainly because when you restart sddm, it works fine again (for one log in) 15:36:51 .bug 1005233 15:36:52 so i think there's some state information lingering in sddm, causing it to fail once tried again 15:36:54 rdieter: Bug 1005233 lightdm greeter does not allow existing users to choose a different DE when they logout or "switch users" - https://bugzilla.redhat.com/show_bug.cgi?id=1005233 15:37:06 fyi ^^ I think this may be a symptom of the same or similar problem 15:37:16 thanks, i will take a look at it 15:37:45 planning on reassigning that to systemd , and proposing it a blocker 15:38:11 ah, yeah, i already saw that, this is definitely wrong 15:38:41 mbriza: the problem you describe on not being able to login again... does that happen on f19 too, or just f20+ ? 15:39:06 well, currently it doesn't happen anywhere because it's "fixed" by calling pam_end() twice on the same handle 15:39:19 heh, ok 15:39:23 which is wrong of course 15:39:33 pam_end_harder 15:39:35 and afaik it causes automatic relogging for autologin 15:40:15 may be a pam problem then, eh? 15:40:25 i doubt that 15:40:54 other DMs just work with the same flow of pam calls as we have 15:41:16 per the bug I referenced, all DMs fail the same way for me. 15:41:21 currently i'm experimenting in ending the session after the X process has exited, the session process has exited and such... still no success though 15:42:12 hmmm, unfortunately i don't have my f20 machine here right now, it'd be worth trying again to be sure 15:42:18 only significant pam difference f19 vs f20, is bug 969174 15:42:20 .bug 969174 15:42:24 rdieter: Bug 969174 pam_sepermit is causing xguest to ask for password on screenlock. - https://bugzilla.redhat.com/show_bug.cgi?id=969174 15:42:59 anyway, didn't mean to sidetrack things here... 15:43:03 anything else wrt sddm ? 15:43:17 in particular, anything you need help with or testing particular features? 15:43:56 theme - have a draft, nobody has still seen the source but it looks kinda like this http://ma.rtinbriza.cz/s/esdedeem.png and i have no idea how to finish it and who to discuss it with 15:44:20 in particular, i'd love if somebody tested itv with some directory service 15:45:18 if you have any setup somewhere, otherwise i'll just ping the freeipa guys in rh 15:46:13 mbriza: we use active directory @ work, so I could test that 15:46:34 but that's all generally handled via pam, so I wouldn't expect any problems 15:47:06 * rdieter likes the basic theming, good start. 15:47:40 Yeah, it's an OK design. 15:47:55 #info mbriza working on improving pam, has a draft non-userlist theme 15:48:27 Are you going to also maintain a version with the userlist or only that one? 15:48:54 so far i'd just stick to this one 15:49:03 sddm has other themes that support userlist, I dont think we need to provide another one 15:49:39 ok, moving on? 15:50:47 #topic plasma-nm - https://fedoraproject.org/wiki/Changes/Plasma-nm 15:51:11 jgrulich: your baby, can you give a brief update on plasma-nm status? 15:51:53 well, on F19 it works okay, but it seems we have some problems on F20, but don't know whether the problem is NetworkManager 15:53:30 let's make sure to file bugs to track these items 15:53:54 (if not already, of course) 15:54:30 the only existing problem now is that you can't see IPv4 address in connection details 15:57:00 in particular, can't see ipv4 addr for wired connections (I can see wireless ipv4 addr ok) 15:57:32 rdieter: hmm, that's weird 15:57:47 * Kevin_Kofler cannot find jriddell's script to migrate to plasma-nm. 15:57:57 It's not in the Kubuntu package. 15:58:17 because there is no difference between getting IPv4 addr from wired and wireless connections 15:58:38 * rdieter double checks 15:59:18 confirmed, it shows fine for wiki 15:59:52 that's another thing, Aaron Seigo and Sebastian Kugler want for us to rename plasma-nm to have the same name like the old one for better upgrade experience 16:00:45 so when we rename it, we won't need any migrate script 16:01:10 I think that makes sense. 16:01:25 hrm... that means when/if renaming happens, we can no longer deploy it to < f20, unless you mean to migrate them too eventually? 16:01:28 The name is also hardcoded in the initialization code in kde-workspace if I'm not mistaken. 16:01:55 So you also avoid having to install an init script for new users if you use the well-known name. 16:02:42 rdieter: we can have both, the one under kde-plasma-nm and the old one as it is now, we would only rename the plugin name 16:03:35 jgrulich: but wouldn't they conflict then? 16:04:01 rdieter: yes, that would be the only change in packaging 16:04:46 icky 16:05:21 Conflicts: is evil… 16:05:21 I suppose we could patch for < f20 to not conflict, if we *really* wanted to 16:06:16 Yeah, I think it also makes sense to keep the name we already ship in F19-. 16:06:42 But then of course we'd probably want a migration script to switch back to the old name on upgrades to F20. 16:07:18 (In fact IMHO we need that script anyway because we already shipped the new (soon to be old) name.) 16:07:24 I don't think it's possible to just simply change the plugin name by some patch due to translations 16:07:40 There's no way around a migration script. 16:08:57 I could live with updating only for f20+ when/if it is renamed 16:10:01 it's only a tech-preview for < f20 anyway, we could do unofficial builds of there's demand for it 16:10:27 anything else? nm-related? I guess we're running short on time 16:11:01 fyi, hero haraldh whipped up a systemd build to test, may help fix some/all of our issues, http://koji.fedoraproject.org/koji/buildinfo?buildID=465157 16:14:50 #topic KDE test day 16:15:02 last item on agenda was this, who added it? 16:15:06 me 16:15:13 I guess its about picking a test day? 16:15:27 just a reminder that we should organize some KDE test day 16:16:16 last few releases we've had test days, and have got a lot of good feedback, polish, bugs from them 16:16:39 and it will be useful for plasma-nm 16:16:54 anyone interested in leading test day organization? 16:17:14 I can help 16:17:24 but the question is, when? 16:17:27 Sadly, more than Test Days, we need Fix Days. ;-) 16:18:24 Finding people to FIND bugs is the easy part, the hard part is finding people to actually FIX them. :-( 16:18:29 https://fedoraproject.org/wiki/QA/Fedora_20_test_days 16:18:38 Kevin_Kofler: every day is a fix day. 16:18:41 ha ha 16:18:51 :-) 16:18:58 for plasma-nm it's obvious who will fix them :) 16:19:39 imo, lets take the test day topic onlist, ask for feedback about dates, and go from there? 16:19:52 +1 16:20:08 ok 16:21:11 I think its a given that we want a test day 16:21:27 its just a matter of documenting test cases, and stuff we want tested... and picking a date 16:22:32 and only picking the date we can wait for (not too long though), the others do asap 16:24:58 lets wrap up, went way longer than intended. :) 16:25:02 thanks everybody 16:25:04 #endmeeting