[06:25] Riddell: thanks [06:55] Good morning. [07:05] morn [07:31] * apachelogger thinks all software should simply exit with "subprocess returned error exit status 1" [07:31] it increases debugging speed by about 300% [07:56] shadeslayer: bug 1309458 [07:56] bug 1309458 in ubiquity (Ubuntu Trusty) "[kubuntu] ubiquity installer crashes when clicking release notes link with non-US locale" [High,Triaged] https://launchpad.net/bugs/1309458 [07:59] https://docs.python.org/3.4/search.html?q=direxists&check_keywords=yes&area=default [07:59] today is tuesday but it feels a lot like monday looking at my WTF counter [07:59] apachelogger: because peoplem in sane countries were not at work yesterday :D [07:59] what is the holiday on Monday? [08:00] * valorie lives in an insane country [08:00] Easter monday! [08:00] hmmm [08:00] we have a 4 day weekend for easter :) [08:00] it's when you try to get sober enough for work [08:00] so I heard [08:00] yeah, otherwise known as "recovery day" [08:00] :P [08:01] * valorie didn't have enough mimosas to need sobering up [08:01] sounds like a bug [08:02] "The Richer You Are, The Longer You'll Live" [08:03] linkedin sure knows how to send me mails that sound like spam, look like spam and smell like spam [08:04] linkedin sends out something that *isn't* spam? [08:05] I wouldn't know, never seen anything that isn't [08:05] bug 1310620 much wow [08:05] bug 1310620 in gwenview (Ubuntu) "Gwenview crashes while zooming in at high resolution" [Undecided,New] https://launchpad.net/bugs/1310620 [08:06] that image is exciting 255 mib in size [08:07] so while y'all were off recovering, the baloo issue has blown up like a balloon [08:08] that image is of a painting I saw in Madrid! [08:09] I could hardly bear to leave it [08:09] valorie: was that a pun, sounds like a pun [08:09] valorie: what issue? [08:10] some people's system being overloaded [08:11] overloaded? [08:11] of course, the bigger issue is that rather than filing bugs, they are shouting "dictator' and shit like that [08:11] I was unable to get anybody to file a bug [08:11] :( [08:12] mustn't be that big an issue then [08:12] * valorie doesn't want vishesh to get discouraged [08:12] let's not tell him :P [08:13] * apachelogger notes that there's still unhandled exceptions in baloo [08:13] unfortunately, i think he reads kde-devel and the comments on his blog and G+ [08:13] pfff [08:13] the secret is to not read anything other than mails sent to one directly [08:14] lol [08:14] this is my default when life gets too stressful [08:17] it was quite fun to provide some support the first two nights after release [08:18] really most problems were from upgrades with odd PPAs left enabled, etc. [08:18] agateau: bug 1310620 .. very reproducible, it crashes in libjpeg though so I guess that's out of scope [08:18] bug 1310620 in gwenview (Ubuntu) "Gwenview crashes while zooming in at high resolution" [Undecided,Invalid] https://launchpad.net/bugs/1310620 [08:18] http://paste.ubuntu.com/7305319/ [08:19] valorie: that should not happen :O [08:19] all ppas are disabled on upgrades [08:19] unless you mean with old packages from old ppas ^^ === xnox is now known as NoNameYet_xnox [08:20] not sure, there was some "noobie something" ppa that seemed to have pinned a bunch of apps [08:21] \o [08:24] ::workspace-bugs:: [1265897] 14.04 Alpha-1 Plasma Netbook Interface has no Web Browser favorite @ https://bugs.launchpad.net/bugs/1265897 (by Marco Parillo) [08:24] ::workspace-bugs:: [1309494] package kdm 4:4.11.6-0ubuntu1~ubuntu13.10~ppa2 failed to install/upgrade: ErrorMessage: su... @ https://bugs.launchpad.net/bugs/1309494 (by bradleesargent) [08:27] we still don't have 14.10 milestones :@ [08:28] valorie: ppas cannot pin apps unless they provide higher versions than what is in the archive (which is certainly a possibility and xorg-edgers likes to do that a lot ;)) [08:29] that could have been; somebody else had taken over helping that person [08:29] I just saw the paste [08:30] ScottK: bug 1310019 <- can we SRU a new package dependency? [08:30] bug 1310019 in ktouch (Ubuntu) "ktouch shows an empty home screen because org.kde.charts is missing" [Undecided,New] https://launchpad.net/bugs/1310019 [08:31] or would a recommends be preferred? [08:42] so chromium is kinda borked in trusty :( [08:49] mek was saying that canonical made some changes that he thought were rather unsafe [08:49] "untested" I think is the word he used [08:50] jussi: what problem have you run into? [08:50] apachelogger: 254MB for a jpeg... wow [08:50] I run chromium, and so far, no probs [08:50] * agateau downloads [08:51] valorie: where do I start... flash not working, delete and backspace not working, intermittently not able to enter text. [08:51] time to switch to Kubuntu default browser... firefox :) [08:51] flash stopped working for a day, then started again [09:17] youtube review "I think dolphin is actually the only file manager gaining features." [09:18] * apachelogger giggled a bit [09:19] good morning [09:43] apachelogger: dafuq @ the ubiquity crash\ [09:43] how does changing lang affect os.direxists [09:43] it doesn't but according to official docs that function doesn't exist :P [09:43] still syncing image [09:43] at 18kbps -.- [09:44] shadeslayer: https://docs.python.org/3.4/search.html?q=direxists&check_keywords=yes&area=default [09:44] https://docs.python.org/3.4/library/os.html [09:44] apparently one is supposed to use os.path [09:48] oh [09:48] 14.10 doesn't even have a name yet [09:48] this is all rather inconvenient [09:48] we should have taken the week off [09:50] shadeslayer: https://www.kubuntuforums.net/showthread.php?65214-Blueshell-Firefox-KDE-Package-Error firefox-kde-support apparently not built on trusty [09:51] mhm [09:54] apachelogger: requested builds [09:56] apachelogger: btw can you make out why gwenview depends on baloo [09:56] image tagging? [09:57] because I don't see it being installed during the build, and we don't depend explicitly on it [09:57] or rating maybe [09:57] yofel: sure, but we don't explicitly depend on it, so how does it end up being a dep? [09:58] hm [09:59] hm indeed [10:08] shadeslayer: libbaloo*.symbols has a dependency on baloo defined [10:08] so anything that links against one of those will depend on baloo [10:12] I see [10:13] apachelogger: yofel so vHanda wants the baloo dep gone, any ideas how to proceed? since even if I remove the depends from the symbols file, I'd have to rebuild all packages that build depend on baloo-dev [10:14] I want the users to be able to remove baloo without dragging out the rest of KDE [10:14] shadeslayer: why should the dep be gone? [10:14] could explicitly set it to recommends [10:14] because it is a runtime dependency, which is *optional*. [10:15] shadeslayer: or add it to the seed [10:15] mhm [10:16] but it'll need everything recompiled to change that depends [10:16] yep [10:21] * shadeslayer is annoyed at python [10:21] stupid thing [10:22] can't even copy a file [10:22] the baloo packaging must have an excitingly useful changelog if it doesn't even document implicit shlibdep abuse [10:23] Riddell: shadeslayer : So what kind of ETA are we looking at? [10:23] 6 months [10:23] Riddell: apachelogger http://paste.kde.org/pbga1nvsr [10:24] oh [10:24] . [10:24] right [10:24] apachelogger: shadeslayer : You guys cannot change a simple depends for another 6 months? [10:24] sorry, let me rephrase that again [10:24] and incorrect depends [10:24] *an incorrect depends [10:25] apachelogger: it's only half a dozen packages, and we have SRU exception for KDE SC so at worst when 4.13.1 goes in [10:25] no because it's not a simple dependency but a circlejerk [10:26] who comes up with shit like that anyway? [10:27] it's implemented the same way as kde-runtime on libkdecore5 [10:27] explains who comes up with shit like that then [10:27] and the next SRU should fix it, right [10:27] go debian! [10:27] where to? [10:28] insanity? ^^ [10:28] ah good point! [10:30] apachelogger: Riddell https://code.launchpad.net/~rohangarg/ubiquity/ubiquity/+merge/216691 [10:30] apt-cache rdepends baloo |tail -n +3 |grep -v baloo |wc -l [10:30] 10 [10:31] Riddell: that's not half a dozen :P [10:31] but we have the favor of the gods of bullshit fixery and only actual SC packages are affected [10:32] even so, that will require substantial adjustment [10:32] apachelogger: reverse-depends -r trusty baloo only gives me 6 [10:32] s/6/7/ [10:32] shadeslayer meant: "apachelogger: reverse-depends -r trusty baloo only gives me 7" [10:32] I am seeing kdepimsies libs [10:32] which means all of kdepim will need adjusting [10:32] Riddell, shadeslayer, vHanda: I do believe this is wildly out of scope for stable updates [10:33] wouldn't that mean just rebuilding kdepim? [10:33] shadeslayer: no, it means figuring out which thing that uses the pimsies library actually uses baloo and then add baloo as a recommends for that package [10:34] and if you miss something then there'll be a bug report "random application you never heared off does not work!!eleven" [10:34] ^ all of this is why implicit link dependency injectin is utter bullshit [10:35] uhm, wouldn't the current rdepend list be the one to add recommends to? [10:36] as that's the list of packages that use baloo in some way [10:36] ^ that's not wrong and it's not right [10:36] (ok, unless we're talking about dbus, but that's not even correctly mapped even now) [10:37] a library that uses libbaloocore probably doesn't necessarily need baloo [10:37] ah, true [10:38] an application using that library also might or might not need baloo [10:38] which is why I think this is out of scope [10:38] it's random guessing without testing at best [10:38] it doesn't [10:39] vHanda: what doesn't? [10:39] the 'baloo_file' and other processes are completely indepndent. [10:39] the libraries do not require them. [10:39] vHanda: the baloo package also contains the akonadi stuff [10:39] in fact it contains everything that is not strictly libbaloo*.so* [10:40] Can you guys split it up so that the baloo_file executable can easily be removed? [10:40] not for 14.04 [10:40] or maybe [10:40] ScottK: what's the sru team stance on introdcuing new dependencies? [10:41] new deps are fine, new binaries... not so much [10:42] yofel: the wiki page does not say anything about that :P [10:42] actually, the only mention I see is WRT upstream releases "Note that some noise introduced by autoreconf is okay, but making structural changes to the build system (such as introducing new library dependencies) is generally not. " [10:42] it's all much shitty [10:43] shadeslayer: please put down a card for 14.10 to remove the implicit linker bullshit from baloo and figure out dependencies proper [10:43] doing [10:43] shadeslayer: and an SRU card for 14.04 to figure out how to get baloo_file as an optional dependency [10:43] well, it makes sense in general. It's just that the KDE folks don't really consider dep and source layout stability to be a worthy point of point releases -.- [10:43] * apachelogger remembers to cleanup the boards and sighs [10:44] Is there any way to get it done faster? [10:44] 6 months is too long. [10:44] that depends on what course of action the sru team would like us to take [10:45] either way the next ISOs are rolling in 4-6 months (14.04.1 and 14.10) [10:46] vHanda: best possible scenario we can push updates for 14.04 with 4.13.1 (which would be ~1.5 months with QA and everything) [10:47] due to the way the dependency is created to begin with a faster update is not an option unfortunately [10:49] Riddell: apachelogger can you review https://code.launchpad.net/~rohangarg/ubiquity/ubiquity/+merge/216691 [10:50] I do not have ubiquity review powers [10:50] apachelogger: but do you think it looks alright? [10:52] shadeslayer: that also does not matter because I don't speak python well enough :P [10:53] heh [10:53] it certainly doesn't use nonexistant functions anymore :P [10:53] well, it works in my vbox [10:54] https://play.google.com/store/apps/details?id=org.kde.kdeconnect_tp&reviewId=Z3A6QU9xcFRPRlNfR29TdldsMk5QRTMzTjVmdm55c0F0VGZYbWU4VXZWQzNaUGVmbWc1aW43aEVIQUhTUXh4T2xCTVh5Xy1IeEFib2dNRV9IU1FmMS14eXc [10:54] @_@ [10:56] apachelogger: splitting a package would require QA? [10:56] vHanda: oh, splitting might be faster [10:56] still needs a week of testing though [10:56] that's mandatory for all updates to stable releases [11:07] shadeslayer: what's the difference between os.direxists and os.path.exists ? [11:08] Riddell: there is no direxists on Python 3 >.> [11:08] tsk [11:08] yeah ... [11:09] * apachelogger doesn't want to look at muon crashers -.- [11:09] 'Morning all [11:12] https://errors.ubuntu.com/problem/945085a4e972ad84a6076ed5abf7c6cba253b72c [11:13] driver manager ... wut [11:13] I have no clue [11:13] many crashes there [11:13] oh [11:13] actually [11:13] fails assert [11:14] Q_ASSERT(m_transaction == nullptr); [11:14] I think I know why [11:14] I don't think I like it though [11:15] loading loading loading xD [11:15] so slow [11:16] apachelogger: https://trello.com/c/4Jw6u7Fx [11:16] what's that? [11:16] kdeconnnect 0.6 is out apparently [11:17] or someone claimed so anyway [11:17] don't see a tar [11:17] maybe not out, but general question ^^ [11:17] anyway [11:17] not important right now [11:17] shadeslayer: I don't see how that backtrace happens [11:18] vHanda: http://paste.kde.org/p0oszucsz [11:18] you'd have to click apply twice in a row [11:18] but it gets deactivated after first clicky [11:18] apachelogger: try Apply and then OK ? [11:18] no dice either [11:18] hm [11:19] oh [11:19] actually the states are bugged there [11:19] it calls disableui after the transactions have been setup [11:20] I still don't see how you'd get two calls though [11:20] since that is in the mainloop you probably could not cause two clicks anyway as the loop is blocked [11:21] * shadeslayer ponders when 14.10 opens [11:21] when it gets a name [11:21] mark is probably on easer vacation, come back next week [11:21] *easter [11:21] should have given it a name before going on vacation -.- [11:21] ^^ [11:22] apachelogger: I thought we fixed https://errors.ubuntu.com/problem/205be0085e711d8928ed136719b457bca28ca00a [11:22] this has many incarnations [11:22] as you can see from the bt any old exception could cause it and there's no way of telling which :P [11:23] I see [11:23] so there's still some unhandled exceptions [11:23] yeah [11:23] probably only one though [11:23] given the reportyness [11:23] 25 crashes after release is not very substantial [11:23] * shadeslayer wonders if apport automatically sends reports after release [11:24] apachelogger: true that [11:24] most of them are from .95 [11:24] shadeslayer: it should [11:24] AFAIK that was the plan anyway [11:24] i.e. instead of getting to report a bug you can send the crash data [11:24] whoopsie should I believe [11:25] which is the middleground between silly report dialog and no crash data after release [11:25] yofel: whoopsie needs apport to first aggregate the crash [11:25] if it doesn't then whoopsie has nothing to send [11:25] ah, true that [11:26] shadeslayer: driver manager hates me :'< [11:27] :) [11:27] I now used the modifier key [11:27] still no crashy [11:28] * apachelogger turns off polkit [11:28] * apachelogger wonders how qaptworker got translations all of a sudden [11:29] Riddell: thx for merging, are you also going to release? [11:29] shadeslayer: well that's my next question [11:29] oh but no xnox to ask it to [11:30] shadeslayer: does this only affect !en_US? [11:30] Riddell: I think it also affects en_US [11:31] right, I'll update bug 1309458 [11:31] bug 1309458 in ubiquity (Ubuntu Trusty) "[kubuntu] ubiquity installer crashes when clicking release notes link with non-US locale" [High,Fix committed] https://launchpad.net/bugs/1309458 [11:31] much silliness from my side [11:33] oh [11:33] OH [11:33] OHHHHHHHHHH [11:33] Oo [11:33] much silly [11:33] shadeslayer: I think I refactored the ui disabling away [11:34] O_______O [11:35] lol [11:35] mh [11:35] no [11:35] shadeslayer: am I missing something or did driver manager never lock the actual selection widgets? [11:36] i.e. I can switch driver, click apply and continue to switch driver and click apply again [11:36] and I don't see disabling code in earlier revisions [11:36] you are potentially correct [11:36] all your fault, I knew it :P [11:36] do we want a u-series kickoff meeting? [11:37] Riddell: yes, I'll send a mail about that soonishy [11:37] * yofel notes he's on vacation this week [11:37] Riddell: feel free to organize meeting though [11:37] ah and yes [11:37] Riddell: earliest next week [11:37] * apachelogger only has mail stuff about general work epics for the future [11:38] shadeslayer: the widgets are not part of a super widget? :O [11:39] * apachelogger forgot all he knew already ^^ [11:41] apparently not ^^ [11:41] shadeslayer: ok, I think I have a minimally invasive fix [11:42] and I really think the kcm locked at some point [11:42] perhaps that was before you introduced the widget class or something [11:42] maybe [11:42] I recall it locking , but don't remember for sure [11:42] brb lunch [11:43] * apachelogger branches trusty [11:46] mhh [11:46] there's more issues with that [11:58] apachelogger: If it fixes a bug, sure. [11:59] apachelogger: If depends is the technically correct relationship, then do that. [11:59] ok thanks [11:59] vHanda: ETA 1-2 weeks [12:00] shadeslayer: course of action for baloo SRU ... split indexer executable into own package, make baloo recommend the new package [12:00] apachelogger: I don't think we should split it in an SRU. [12:00] That's a bit different. [12:01] ScottK: how so? [12:02] 1. It's a packaging choice. There's no bug that's being fixed. 2. That means every single KDE user has to dist-upgrade instead of upgrade to end up with exactly the same thing they have now. Seems like a lot of user annoyance for not much. [12:02] Additionally, baloo seems reasonably well behaved to me, so I'm not sure what we gain. [12:03] ScottK: the bug is that the indexers are not a required piece [12:04] ScottK: the other course of action is removing the bogus dependency injection in some 7 packages [12:04] File indexing can be turned off. [12:04] Pim indexing can't be, so you're stuck with that regardless. [12:05] vHanda: ^ pim indexing different? [12:05] I don't see the difference between turning file indexing off and having it do nothing and removing it. [12:05] vHanda: ^ please make your argument [12:05] ScottK: I think vHanda is on lunchbreak though, best discuss this with him [12:05] Oddly, I'm making his argument on kde-devel. [12:05] Sure. [12:06] * apachelogger as usual understands both arguments but doesn't care either way :P === apachelogger is now known as apachelogger_ === apachelogger_ is now known as apachelogger [12:06] I've no issue with splitting things up in the development release, but unless the situation changes in 4.13.something, I don't see the point in changing what we have for the current release. [12:08] BBIAB [12:41] bah, I can't get libreoffice to crash when testing bug 1290514 [12:41] bug 1290514 in qt4-x11 (Ubuntu Saucy) "LibreOffice KDE4 backend crashes in Qt4 recursive paint events" [Undecided,Fix committed] https://launchpad.net/bugs/1290514 [12:47] Back. [12:53] apachelogger: I saw your bug for the driver manager, but I didn't see the package. Did you upload it yet? [12:54] Uploading kubuntu-driver-manager_14.04ubuntu8_source.changes: done. [12:54] Successfully uploaded packages. [12:54] ScottK: just now [12:54] OK. I'll have a look then. [12:54] related autoreport https://errors.ubuntu.com/problem/945085a4e972ad84a6076ed5abf7c6cba253b72c [12:54] There is is. [12:54] it is [12:55] shadeslayer: I am pushing a trusty branch, please make sure to setup dpkg-mergechangelogs or you'll have a hard time merging branches ;) [12:57] apachelogger: Accepted. [12:57] thx [12:58] shadeslayer: ^ will need your testing plz [12:58] I think I covered all cases, but you never know ^^ [13:03] * ScottK wonders if apol is the relevant Aleix in https://wiki.debian.org/SummerOfCode2014/StudentApplications/FlorisAndreiStoicaMarcu [13:03] (GSoC project for Muon in Debian) [13:04] I guess [13:15] ScottK: we were debating if the enable / disable button should be added. Another option is that we simply tell users to uninstall it. [13:16] currently they cannot do that as that would remove dolphin + pim + other packages [13:16] vHanda: I read the discussion on kde-devel ML. === rdieter_laptop is now known as rdieter [13:16] okay [13:16] I think disable makes more sense. [13:16] ohm [13:17] apachelogger: plz be wraping git entries at 80 chars [13:17] In any case, for Kubuntu 14.04, which has 4.13.0, we are constrained by policy in what changes we can make since it's already released to end users. [13:17] vHanda: BTW, I quite like baloo with pim in 4.13. [13:18] ScottK: but this would just be splitting it up. [13:18] Is there some way you could ask for an exception? [13:18] I could, but I'm not sure it's worth it. [13:19] shadeslayer: tell debcommit [13:19] The problem is that we'd then have to install a new package, which requires extra user interaction on the upgrade. [13:19] So we'd be inconveniencing/possibly confusing 100% of our users for the few that might want to uninstall the file indexer. [13:19] for me it would since it would allow the compaliners to easily remove baloo [13:19] and be done with it [13:19] Generically, I think a disable button/checkbox is better anyway. [13:20] I understand the frustration. [13:20] I think it's more natural to think in terms of turning things on/off than to uninstall packages to do so for most people. [13:23] Plus I don't think there's a way to remove baloo via any GUI methods by default since we only ship with discover [13:23] opinions on bug 1310828 welcome, nothing major [13:23] bug 1310828 in kdepim (Ubuntu) "The package "storageservicemanager" should be a dependency of Kmail" [Low,Confirmed] https://launchpad.net/bugs/1310828 [13:27] apachelogger: Where do I find this greyed out U/I? [13:27] ScottK: kmail settings -> configure -> composer -> attachments tab -> bottom left button [13:28] Thanks. [13:28] it allows you to automatically push attachments of considerable size to dropbox (for example) rather than attach directly, however the manager apparently can manage cloud stuff in general [13:28] e.g. create new folders and stuff [13:29] apachelogger: I vote leave pim as is and add a thing in kubuntu-docs explaining you have to install the packages if you want this. [13:29] valorie: ^^^ [13:30] * shadeslayer thinks the storage manager thing is quite confusing [13:30] yes [13:31] that's what I was whining about [13:31] shadeslayer: feel free to file improvements suggestions [13:38] ::workspace-bugs:: [1311105] KDE locks up at startup if window decorations changed from default. @ https://bugs.launchpad.net/bugs/1311105 (by Kyle) [13:44] Quintasan: u got the sddm packaging stuffs? [13:44] apachelogger: I believe I do [13:47] apachelogger: Are you OK with adding it to the docs, but not changing the package? [13:47] apachelogger: Yup, everything I did is on my HDD [14:30] ScottK: I doubt the docs would see translation, so I think we can just add it as known bug on the release page and move on [14:30] at least that way it gets exposure [14:30] Quintasan: please throw it up somewhere [14:31] Quintasan: on a related note, it probably would be nice if you could try to update your packaging for inclusion in 14.10 once that opens development [14:39] apachelogger: I'm not convinced it's a bug. I think few enough people would use it that I don't know that it should be there by default. Also, interaction with non-free web services tends to need a lot of maintenance when they change stuff, so I'm not sure we should take that on. [14:39] ::workspace-bugs:: [1311105] KDE locks up at startup if window decorations changed from default. @ https://bugs.launchpad.net/bugs/1311105 (by Kyle) [14:41] apachelogger: Sure thing. [14:49] ScottK: let's just not document it at all then ;) [14:50] adding it to kubuntu-docs wouldn't really add value IMHO [15:00] apachelogger: https://code.launchpad.net/~quintasan/kubuntu-packaging/sddm [15:00] It's only copyright PITA atm [15:01] though [15:01] https://github.com/sddm/sddm/tree/master/debian [15:01] blargh [15:01] shadeslayer: Y U NO TELL ANYONE [15:02] hm? [15:02] I am unsure how good that packaging is [15:02] I think it's ~decent [15:02] Better than what I currently have [15:02] Why it's still not in Debian [15:06] well, fwiw I think that debian folder needs to be removed [15:07] Quintasan: apachelogger can you see kdeconnect screenshots in muon now? [15:07] or for that matter, any screenshots? [15:07] shadeslayer: nope [15:08] do I have to update muon? [15:08] nope [15:08] Quintasan: can you see any screenshots? [15:08] Digkam give me a big X [15:08]