15:03:42 #startmeeting KDE SIG Meeting – https://fedoraproject.org/wiki/SIGs/KDE/Meetings/2012-09-25 15:03:42 Meeting started Tue Sep 25 15:03:42 2012 UTC. The chair is Kevin_Kofler. Information about MeetBot at http://wiki.debian.org/MeetBot. 15:03:42 Useful Commands: #action #agreed #halp #info #idea #link #topic. 15:04:07 #meetingname kde-sig 15:04:07 The meeting name has been set to 'kde-sig' 15:04:20 #topic roll call 15:04:32 So, who's present? 15:04:42 * dvratil is 15:04:43 yo 15:04:46 * jgrulich is present 15:04:51 Present. 15:05:04 (Obviously. ;-) ) 15:05:09 present 15:06:08 ltinkl, jreznik_: Are you here? Or busy with other stuff? 15:06:14 #chair dvratil rdieter jgrulich than ltinkl jreznik_ 15:06:15 Current chairs: Kevin_Kofler dvratil jgrulich jreznik_ ltinkl rdieter than 15:06:15 * ltinkl is here 15:06:28 jreznik_ said something about another meeting today. 15:06:48 yeah, some partial overlap I believe 15:07:00 #info dvratil, rdieter, jgrulich, Kevin_Kofler, than, ltinkl present. 15:07:15 #topic agenda 15:07:26 The wiki page is depressingly empty… 15:07:45 * jreznik_ is somehow here... just another conflicting meeting today :( 15:07:47 usual, kde-4.9.1 status update 15:08:03 so sorry for not filling up meeting page :) 15:09:17 #chair jreznik 15:09:17 Current chairs: Kevin_Kofler dvratil jgrulich jreznik jreznik_ ltinkl rdieter than 15:09:21 could discuss new nepomuk-core bug, that's getting quite a few dups 15:09:36 #info jreznik present, but has another conflicting meeting. 15:09:57 .bug 858271 15:10:02 rdieter: Bug 858271 [abrt] nepomuk-core-4.9.1-1.fc17: isEmpty: Process /usr/bin/nepomukservicestub was killed by signal 11 (SIGSEGV) - https://bugzilla.redhat.com/show_bug.cgi?id=858271 15:10:12 We also need to discuss the comps and/or live kickstart cleanups for Beta, now that the Alpha freeze is over. 15:10:20 rdieter: You said you were going to work on those. 15:10:44 yup, on my todo list, near the top 15:10:59 but haven't reviewed anything, so nothing new to report yet 15:11:33 #action rdieter to work on comps and live kickstart cleanups for Beta (nothing to report yet) 15:11:53 But one sub-item we may want to discuss: print-manager by default (s-c-p-kde die die die ;-) ). 15:12:07 (any help with comps/kickstarts would be very welcome) 15:12:26 coud disuss about konsole issue (starting in Document) 15:12:30 #halp any help with comps/kickstarts would be very welcome 15:12:49 than: Already fixed in updates-testing, I tested the fix, it works fine. 15:13:02 * Kevin_Kofler is giving +1 karma. 15:13:09 Kevin_Kofler: reverting in kdelibs? 15:13:14 konsole's been fixed, unless we want to consider reverting the commit(s) that was the source of that issue 15:13:28 than: konsole.desktop: +Path=$HOME 15:13:36 workaround 15:13:53 i prefer reverting the fix in kdelibs 15:14:20 than: I would tend to agree (from earlier discussion today in #fedora-kde) 15:14:21 but it's ok with workaround 15:14:48 but let's save that for later in the meeting, when it's on-topic I suppose 15:14:53 -1 to reverting kdelibs 15:15:28 We can start with that, unless there's more stuff to add to the agenda? 15:16:12 #topic default working directory 15:17:16 So the issue there was that the default working directory for all apps is now ~/Documents and this also affected Konsole. 15:17:50 #link https://bugzilla.redhat.com/show_bug.cgi?id=841471 15:17:51 having the kdelibs workaround to set CWD to Documents made some sense way back when it was originally committed (2005), but these days, much less so. 15:18:16 #info Konsole has been fixed in konsole-4.9.1-2.fc17 to explicitly set the startup CWD to $HOME. 15:18:16 rdieter: i agree with you 15:18:20 it's time for legacy apps to be patched/fixed, and kde workarounds dropped 15:18:23 #link https://admin.fedoraproject.org/updates/FEDORA-2012-14700 15:18:30 IMHO this is the right fix. 15:18:35 (The Konsole fix.) 15:19:02 The XDG user dirs spec says that all documents should by default go to ~/Documents. 15:19:29 other apps like xterm/gnome-terminal still have this issue 15:19:40 And in GUI apps, the CWD has normally only one of 2 effects: 1. nothing, then it doesn't matter what we set it to or 2. the default directory for file dialogs. 15:20:03 So IMHO setting the CWD to ~/Documents is exactly what the XDG spec intends us to do. 15:20:12 The ONLY apps where this is causing problems are terminal emulators. 15:20:26 Those should explicitly request the current directory to be $HOME in their .desktop file as Konsole now does. 15:20:31 there are other side effects, but that is the main one 15:20:35 I'm surprised no other desktop has been hit by that. 15:20:57 Kevin_Kofler: does it work for xterm/gnome-minimal with adding PATH=$HOME 15:21:34 One strange thing is, why did the behavior for Konsole change with 4.9? 15:21:36 I recall one bug report whose site was hit hard because they symlinked or redirected the Documents dir to some network share, and now simply logging in held that open/locked indefinitely. 15:22:02 I can't see anything on 4.8 which would set a different CWD for Konsole, yet it starts up in ~. 15:22:08 Kevin_Kofler: krunner (kdeinit actually) was patched recently to respect this too. 15:22:14 But the kdelibs feature to set the CWD to ~/Documents was supposed to be there. 15:22:19 prior it was only apps launched via .desktop files in kickoff, etc... 15:22:21 rdieter: I always use the (classic) menu. 15:22:38 Right here on 4.8.5 on F16, Konsole starts in ~ and has always done that. 15:22:39 so maybe kdeinit fix affected more than just krunner 15:23:11 Changing this in kdeinit means KDE apps will start up in ~/Documents also. 15:23:25 Back in 2005, this was intentionally NOT done in kdeinit so that it'd be applied only to legacy apps, not KDE apps. 15:23:41 KDE apps don't need this workaround. 15:24:24 i'm not sure then. 15:24:27 https://bugs.kde.org/show_bug.cgi?id=183534#c5 15:24:27 Though apart from Konsole, I guess they shouldn't be affected by it, either. 15:25:32 that seems rather old too, doesn't explain any recent behavior change 15:25:35 So that change was to KRun::runCommand. 15:25:37 (that commit) 15:25:42 It's used everywhere, menu etc. 15:25:46 Kevin_Kofler: konsole (4.8.5) still starts in Document via krunnuer 15:25:49 Not just KRunner. 15:26:22 than: Ah OK, so what was actually changed was KRun, which is uses in the menu and NOT in KRunner. 15:26:29 interestingly, my konsole (and yakuake) has always been in $HOME, I never witnessed any use of ~/Documents personally 15:26:35 Very "helpful" naming there. ;-( 15:27:08 sorry, I indeed saw krun, but said and assumed krunner 15:27:28 KRun::runCommand is actually what the menus use. 15:27:42 Not sure what KRunner (Alt+F2) uses. 15:27:58 (Too many ways to run something in kdelibs.) 15:28:22 as i know krunner uses other code 15:30:47 It uses the kinit.cpp code which was changed back in 2008, I guess. 15:31:01 (Not in 2005. The bug report was from 2005, but it took them 3 years to address it.) 15:32:24 anyway, anyone else with an opinion on if we should address this, or what to do about it? 15:32:36 IMHO, keep it as is. 15:32:47 Or get upstream to change it if you think it's broken. 15:33:34 But David Faure seems convinced that it's the right thing to do, he did both the original kinit.cpp change and now also the krun.cpp one. 15:33:34 we can keep it if xterm/gnome-terminal work with adding PATH=HOME in desktop file 15:33:53 Why do we care about running non-KDE terminals at all? 15:34:27 I do have users who still insist on running xterm. 15:34:29 I know neither gnome-terminal upstream nor its Fedora maintainers (they're mostly the same anyway) support running it outside of GNOME in any way. 15:34:48 It still misbehaves with KWin (sizing issues) and I don't think it'll get fixed, ever, or at best by accident. 15:35:05 reverting the change in kdelibs causes side effect that other apps like openoffice starts in $Home instead in Document 15:35:09 updating the .desktop files won't hurt anything, I'll patch xterm 15:35:26 * rdieter hopes libreoffice supports xdg_dirs these days 15:36:07 You're very optimistic. ;-) 15:38:59 So, do we have any agreement? Doesn't look like it… Maybe defer to next week? 15:39:53 There are 2 more topics to discuss, can we defer and move on? 15:39:59 let defer to next week 15:40:08 next topic please 15:40:18 #info deferred to next week 15:40:49 #topic #858271 nepomuk-core crash 15:40:56 #link https://bugzilla.redhat.com/show_bug.cgi?id=858271 15:41:01 .bug 858271 15:41:03 Kevin_Kofler: Bug 858271 [abrt] nepomuk-core-4.9.1-1.fc17: isEmpty: Process /usr/bin/nepomukservicestub was killed by signal 11 (SIGSEGV) - https://bugzilla.redhat.com/show_bug.cgi?id=858271 15:41:29 Yet another Nepomuk crash. :-( 15:42:12 is it regession in 4.9.1? 15:42:23 basically, anyone able/willing to triage this, and report upstream? 15:42:36 than: seems so 15:42:48 There's no evidence that it has happened before, at least. 15:44:31 dvratil_: could you please triage this and report upstream? 15:44:46 ok 15:44:54 dvratil_: great, thanks! 15:45:34 So this is either a KInotify bug or Nepomuk abusing KInotify somehow. 15:46:08 There's a QList which is checked for emptiness, but which seems to be entirely invalid (NULL pointer or whatever), even isEmpty crashes. 15:46:32 Valgrind may or may not help, depending on what exactly the bug is. 15:46:37 It needs to be reported upstream in any case. 15:46:47 (unless it already is) 15:46:56 I'll check it tomorrow, report upstream and try to fix it if possible 15:47:14 #action dvratil_ will report the bug upstream and try to fix it 15:47:34 OK, thanks! I think we can move on. 15:47:43 #topic print-manager by default 15:48:19 i've used print-manager on/off over the past few weeks, only tested as systray applet for print jobs. 15:48:22 So my proposal: now that print-manager is in F18, replace system-config-printer-kde with print-manager in F18 and all future releases, effective immediately. 15:48:36 imo, it's still a bit less user-friendly and quirky, but gets the job done 15:48:50 Is it more or less quirky than s-c-p-kde? ;-) 15:49:01 it can also be used to configure printers. I only tested this once when 0.1 version was available, and it failed. 15:49:18 more quirky than s-c-p-kde 15:49:35 err, sorry, there's 2 pieces it can replace, s-c-p-kde and kde-printer-applet 15:49:45 it's more quirky than the latter 15:50:05 +1 for print-manager 15:50:07 It's hard to have more bugs than the main s-c-p-kde app. 15:50:10 ;-) 15:50:12 and it failed as a replacement for the former, last I tested. needs more testing obviously 15:50:35 so, ok, +1 to try it out. 15:50:36 It's true that the s-c-p-based kde-printer-applet worked surprisingly well. 15:51:48 So I think we agreed on print-manager, or does anybody object? 15:52:28 #agreed default to print-manager in F18+ 15:52:34 #topic open discussion 15:52:38 Anything else? 15:53:12 f18 kde-4.9.1 update needs karma 15:54:10 or timeout will hit today I think, then I can queue stable tomorrow 15:54:31 f17 kde-4.9.1 update went stable a coupld days ago already 15:55:13 it's why I haven't queue'd konsole-4.9.1-2.fc18 for updates-testing yet 15:55:32 I think the timeout will be in 4 hours. 15:55:39 k 15:55:42 Uh, 6 hours. 15:55:51 can probably skip karma'ing then 15:56:36 It's 15:56 UTC now, the timeout should be at 21:41:17 UTC, unless they already changed it to the 7 day timeout. 15:57:53 fyi, submitted https://admin.fedoraproject.org/updates/xterm-278-3.fc18 15:57:54 This Bodhi bureaucracy really sucks, it's not like the update is magically going to become more stable in less than 6 hours. 15:59:38 #info rdieter submitted https://admin.fedoraproject.org/updates/xterm-278-3.fc18 for the initial working directory issue. 16:01:24 Looks like we're out of time… 16:01:44 #endmeeting