[00:00] yea I just like to not wait till the last min [00:03] we are getting noticed thats for sure [00:10] good work tends to do that! [00:13] now, dinner [00:13] ttyl [01:11] palasso_: ping [01:11] palasso_: is this you? https://plus.google.com/113434423178207628891/posts [01:25] valorie: [01:29] dantti: [01:29] Darkwing: [01:29] dantti: sorry wrong person [02:01] himcesjf: hello [03:06] Hey [03:06] Hi ahoneybun [03:12] morning [03:39] himcesjf: hi [03:39] shadeslayer: late night here [03:40] himcesjf: do you have a google plus page? [04:00] dafuq [04:00] W: Failed to fetch https://private-ppa.launchpad.net/kubuntu-ninjas/ppa/ubuntu/dists/saucy/main/binary-amd64/Packages [05:35] Great link about the sysadmins, valorie :D [05:38] ahoneybun, yeah that's me though I don't use it often ;) [05:41] Riddell: ping [05:47] Riddell: Apparently Canonical do think they own the binaries. http://www.canonical.com/intellectual-property-policy claims you need to recompile binaries if you want to redistribute modified versions of Ubuntu. [05:52] ScottK: it also says "The disk, CD, installer and system images, together with Ubuntu packages and binary files, are in many cases copyright of Canonical " [05:52] many cases? like what? [05:52] does that apply to only unity stuff? [05:53] Riddell: I can't figure out how to start an ec2 instance :S [05:53] something is going wrong with key.pem [06:26] good morning [08:46] ScottK: yeah I know, which is nonsense and very objectionable [08:47] shadeslayer: what are you doing? [08:48] shadeslayer: I believe they mean ubiquity/unity/casper/etc. [08:49] Riddell: I tried ec2-run-instances ami-6f951d6e -t m1.small --region ap-northeast-1 --key key [08:49] but that didn't work [08:50] "Client.InvalidKeyPair.NotFound: The key pair 'key' does not exist" uh oh [08:51] shadeslayer: oh ap-northeast-1 [08:51] hm? [08:51] shadeslayer: keys are different for each ec2 regions [08:51] ahhhh [08:51] I've only made them for us east [08:51] ahhh [08:51] I didn't know that [08:52] okay, anyway, don't need one now [08:52] I was trying to figure out UTAH [08:52] UTAH? [08:53] Ubuntu Testing Automated Harness [08:53] I think ... [08:55] guys i ahve asked to krita devs and they said that it's the plasmoid that has a bug [08:55] and it's the widget menu bar [08:55] it doesn't work properly with krita sometimes [08:56] Riddell: http://www.youtube.com/watch?v=-zafY8ZbBoE bug of krita and menu bar [08:59] Riddell: btw I think you forgot to bzr add some files while merging, I fixed those [09:00] for eg. in kiten [09:02] agateau: who maintains menubar? [09:09] :@ [09:09] locale setting is not working [09:09] as expected [09:10] Riddell: I really don't see how we'd correctly get from KDE l10n settings to linux locale setitings [09:10] they are just so different [09:16] apachelogger: ug, why KDE why?! [09:16] KDE is too mighty [09:16] Linux is too dumb [09:20] yofel: oh, btw, someone at some point mentioned that the qt meta git repo tends to break because of their shitty CI [09:21] e.g. submodules hashes not existing [09:21] so I guess atomic packaging may be more sensible [10:00] apachelogger: ScottK shadeslayer et al, the flags are now correctly set [10:00] jussi: thank you [10:00] yw [10:01] flags? [10:03] Riddell: channel flags... new council to be op, former council to not be op [10:04] lthx [10:04] * shadeslayer checks [10:04] jussi: * #kubuntu: You're not a channel operator [10:04] :S [10:05] shadeslayer: /msg chanserv op #kubuntu [10:05] shadeslayer: /msg chanserv op #kubuntu shadeslayer [10:05] so quassel is broken [10:06] !opguide [10:06] no.... [10:06] A guide to help operators learn some of the skills necessary to perform their jobs is available on the wiki: https://wiki.ubuntu.com/IRC/IrcTeam/OperatorGuide [10:06] jussi: "Give op status" doesn't work :P [10:06] I suppose it only works if you're currently op'd [10:06] shadeslayer: to give op status y.. yes [10:07] msg chanserv op #kubuntu-devel [10:07] u noob :P [10:13] apachelogger: maintainers of my former appmenu stuff is Cédric Bellegarde [10:13] apachelogger: gnumdk on irc === KRF_ is now known as KRF [10:32] \o/ [10:33] thanks jussi [10:39] oh a yofel [10:40] yofel: how should we handle recipes for neon kf5? [10:40] I was thinking ... if we have one big branch with all the packaging we could just as well drop the recipes in there [10:40] e.g. qtbase.cfg qtbase/ phonon4qt5.cfg phonon4qt5/ [10:41] also how to do the recipe definition ... [10:41] we could have simple ruby modules as recipes (gives flexibility of writing arbitrary codez within the module) [10:41] or we could have simple shell modules (gives equal amounts of flexibility but since the builder is ruby ... :P) [10:42] or we could have string markup like launchpad (le crap as it requires a parser for no good reason...) [10:44] or we could have some accepted information container format like json (would be between silly markup and raw code modules - having the flexibility but also not having to write code strictly speaking) [10:44] pick your poison please :P [10:48] putting the recipes with the packaging is fine with me. [10:48] As for the recipes, feel free to use ruby if you already wrote the rest in ruby. That can be turned into something shell-callable easy enough [10:48] generally it will work like this: builder creates/updates pbuilder -> executes builder, calling builder itself inside pbuilder -> pbuilded builder munches on recipes (pending decision on how that hsould happen) -> updates bzr/git cache -> merges in build directory -> bumps version -> pbuilder-satisfies deps -> builds src -> uploads [repeats for each package] [10:48] launchpad markup only makes sense with the parser from bzr dailydeb, and I don't quite get what you want to do with json [10:49] so... the envrionment the sources are built in is clean in general but not recreated for each source for performance reason [10:50] yofel: json being the more reasonable approach to the markup business [10:50] how do plan to update the bzr/git cache from inside pbuilder? (bind mount the cache dir?) [10:50] yes [10:51] builder.rb build/ cache/git/ cache/bzr/ [10:51] @args = "--basetgz '#{BASE}/#{POCKET}.tar.xz' --compressprog xz --distribution #{POCKET} --extrapackages '#{EXTRAPKGS}' --bindmounts #{BASE}" [10:51] * yofel puts a note somewhere to learn ruby [10:52] eh, builder should be easy to read :P [10:52] mostly it's just a shell script using classes [10:52] http://paste.kde.org/757388/ [10:52] I might want to do something else than reading every now and then :P [10:53] ah, looks easy enough indeed [10:53] 208 to 214 would be approximately what a package recipe would loook like [10:54] (less ugly, but you'd create your copy jobs to copy repos or arbitrary paths around) [10:59] looks reasonable enough to me [11:03] apachelogger: what might be needed for KF5 would be a way to make a source depend on the same-day version of the previous tier uploads [11:05] yeah, variable substitution like in kde-l10n-common I'd say [11:05] right, something like that [11:06] just another job ;) [11:08] apachelogger: is the kf5 builder stuff in a branch somewhere? [11:08] nope [11:09] just save the paste as builder somewhere and run it :P [11:09] won't produce results though as stuff is commented out [11:09] lol, use gist then ^^ [11:09] whatever that is :P [11:10] though you could just put it into the project-neon repo [11:10] first it needs to do something :P [11:10] never used gist.github.com? [11:10] no [11:10] I tend to not use github [11:11] (fortunately I need to for work reasons) [11:11] ah. imagine pastebin with git repo [11:11] what'd be the point of that? [11:11] I'd just put it in a repo [11:11] dunno, but people still use it ^^ [11:11] silly people [11:12] or rather, browser people :P [11:12] supposedly diffing is the only useful thing [11:33] apachelogger: so medibuntu decss on vlc server? [11:34] yes [11:34] Riddell: so there seem to be serious issues with QtScript on Ubuntu. Amarok crashes left and right, and the BT is always deep in QtScript JIT code [11:35] on Gentoo this doesn't seem to happen at all [11:35] here is one bug report: https://bugs.kde.org/show_bug.cgi?id=261839 [11:35] KDE bug 261839 in Playback "Crash in [QTWTF::HashTable sometimes I can't even get it to start, when Resume Playback on Start is enabled === debfx_ is now known as debfx [11:36] I spy different backtraces [11:37] QtScript is some scary foo [11:37] and outdated Qt in the prominent ones [11:37] yes it is [11:37] apachelogger: check the last comment by me, it's a recent BT [11:37] with Raring [11:37] it's also different from what the previous backtraces head [11:38] *had [11:38] apachelogger: there are several different crashes, but all in QtScript [11:38] so? [11:38] well, something is wrong with QtScript [11:38] or with the QtScript bindings [11:40] *shrug* [11:42] shrugging is not nice, folks, really, this needs to be fixed [11:42] what needs to be fixed? [11:43] the Kubuntu QtScript package, as this is a very Kubuntu specific bug, can't be reproduced on other distros [11:43] there are 3 distinct crashes in that report..... [11:43] and the majority of them on an old Qt [11:43] apachelogger: don't look at the old stuff, only look at thre recent reports [11:44] markeyŝ backtraces are all recent [11:44] and incomplete [11:46] apachelogger: the new ones are not incomplete [11:46] again, look at the most recent backtraces [11:46] I am reasonable certain that amarok has more than one thread [11:47] what the heck are you talking about? [11:47] the relevant parts are there, stop being a dick [11:47] yeah [11:47] so here is the thing [11:48] if you want someone else to look at stuff [11:48] then you should provide all the information [11:48] right now I see a backtrace with one thread [11:48] markey: ^ [11:48] and I happen to know that the qtscript jit shit had at least one crash in the past that was in fact thread related [11:48] so showing me one thread makes me go: "that backtrace aint complete" because it is not complete... [11:49] here you go: http://paste.ubuntu.com/5729104/ [11:50] markey: dpkg-query -l libqt4-script [11:51] here's the output: http://pastie.org/8000126 [12:00] hiho, kde 4.10.4 is scheduled for this week > [12:00] ? [12:00] markey: no failed assert? [12:01] soee: was this a statement or a question? [12:02] apachelogger: no assert, nope [12:02] Mamarok, the ? sign is the answer for your question :0 [12:02] :) [12:03] soee: http://kyofel.dyndns.org/kubuntu/build_status_4.10.4_saucy.html [12:03] kdepim fails because of boost weirdness [12:04] shadeslayer, so saucy first than backports ? [12:04] yes [12:04] files changed around quite a bit [12:04] which is why I haven't uploaded for raring [12:04] shadeslayer, ok thank you for the info [12:04] fix saucy, then we won't have to fix raring, quantal and precise [12:04] ;o [12:04] s/fix/upload [12:04] anyway [12:05] -./usr/share/icons/hicolor/80x80/apps/khangman-harmattan.png [12:05] whut [12:05] -.- [12:09] yofel: apparently ubuntu has had ddebs for some time [12:09] http://ddebs.ubuntu.com/ [12:09] and soyuz will have ddebs soonish [12:23] markey: which scripts and applets do you have? [12:23] apachelogger: do you know how to debug ksplashqml ? [12:23] debug? [12:24] amarok [12:24] apachelogger: now to understand if it look well and if the code is good i am using the kcm module to test the theme [12:24] isn't there another way ? [12:25] apachelogger: "LyricWiki", "Cool-Streams", "LibriVox", "Free Music Charts" [12:25] Riddell: want to take over 4.10.4 for saucy? I'm done for the day :) [12:25] apachelogger: it's easy to reproduce when you enable "Resume Playback on Start" [12:26] markey: trying that [12:26] and have it start playing stuff right away [12:26] so far no luck [12:26] shadeslayer: mm can do, much more to do? [12:26] not really [12:26] I think this "Free Music Charts" script triggers crashing a lot [12:27] markey: thing is... there was upstream breakage in a similar call chain which should not happen in 4.8.4 [12:27] just 3-4 packages + oxygen-icons and kajongg I think [12:27] Riddell: http://kyofel.dyndns.org/kubuntu/build_status_4.10.4_saucy.html [12:27] markey: *** Error in `amarok': double free or corruption (!prev): 0x0000000000a55170 *** [12:27] on exit [12:27] hm [12:27] I don't get that on exit [12:27] markey: which applets do you have in teh context view? [12:27] Riddell: it's not as bad as it looks on that page [12:28] apachelogger: that's from pgst [12:28] apachelogger: Current Track, Analyzer, Wikipedia, Lyrics [12:28] well [12:28] apachelogger: it only happens with pgst 1.0 [12:28] and hurray, you get that too [12:28] oh right, gstreamer [12:28] yep [12:28] pGST [12:28] haven't been able to debug that [12:28] whatwhatwhat [12:28] markey: you are on pgst1? [12:28] ah [12:28] apachelogger: using default from Raring [12:28] shadeslayer: the double free you mean? [12:29] apachelogger: yes [12:29] kk [12:32] markey: got it [12:33] shadeslayer: saucy has nothing to do with raring, backports aren't SRU-able. You need to re-run the script once .3 is in -updates [12:33] apachelogger: got a crash? [12:33] yes [12:33] ok [12:33] http://wstaw.org/m/2013/06/03/plasma-desktopyY2118.png [12:33] also that [12:34] Oo [12:34] what's that? [12:34] lol [12:34] shadeslayer: and what about ddebs? Those have existed since the age of time [12:34] yofel: oh, I wasn't aware of them :S [12:34] markey: my context view [12:34] I thought they were still not implemented [12:35] shadeslayer: https://wiki.ubuntu.com/DebuggingProgramCrash [12:35] isn't that dbgsym? [12:35] wait [12:35] I'm confused now [12:35] shadeslayer: that's the same damn thing [12:35] whoopsie [12:35] lol [12:35] packagename is -dbgsym, file extension is .ddeb [12:35] for some reason I thought ddebs meant delta debs :/ [12:36] ^^ [12:36] they aren't debs if they are delta -.- [12:37] markey: I think ben is on to something [12:37] 2.7 does not crash [12:37] apachelogger: no, actually I looked into that. it's not related to that code [12:38] it still does not crash with 2.7 [12:38] I doubt that [12:38] also consider: on Gentoo it never crashes with QtScript [12:38] I find that odd [12:39] that means nothing for qtscript [12:39] it's highly dependent on platform/compiler/flags [12:39] true [12:39] the crash I was talkign about earlier with the same backtrace was such a thing.. it would crash only on amd64 because of platform specifics [12:40] apachelogger: how about trying to disable the JIT? [12:40] maybe that could fix it [12:41] no [12:41] that would work around it [12:41] anyway [12:42] it does not crash with 2.7.0 [12:42] so I suspect a change in master screwed with it somehow === sem_ is now known as Peace-wifi [12:51] Hiyas all [12:55] someone using 13.04 on laptop ? [12:55] soee: me [12:55] yofel, can you boot system and leave it for night and next day check how kded4 behaves ? [12:56] soee: the system has been running for 2 days now and nothing unusual so far (neither the week it had been running before that) [12:57] for me if i do such thing or im using laptop for few hours kded4 starts to eat whole free memory and i have to reboot to be able to use it [12:58] after whole night when i checked in the morning it was using > 2GB ram [12:58] so then system stops to response or do it very slooow [12:58] hm... you can unload kded modules over dbus somehow I believe. Maybe that could help with debugging. [12:58] the only option is to kill kded4 when im starting my work on lap [13:06] yofel: unloadModule [13:07] shadeslayer: explain that to soee, I forgot the exact invocation [13:07] qdbus org.kde.kded /kded org.kde.kded.unloadModule moduleName [13:08] qdbus org.kde.kded /kded org.kde.kded.loadedModules will give you all the loaded modules [13:08] apachelogger: ok here's the deal. I'll set my git to 2.7.0 tag and try with it. we'll see [13:08] easy enough [13:09] shadeslayer: thanks! [13:09] np [13:09] soee: for reference: the modules I have in use: http://paste.kde.org/757436 [13:09] soee: saw mention of kded mem usage, may want to look @ https://bugs.kde.org/show_bug.cgi?id=271934 (which is apparently some odd interaction between kauth and cron jobs) [13:09] KDE bug 271934 in general "kded4 process grows on memory usage (possible leak)" [Normal,Reopened] [13:11] rdieter, yes this might be my case [13:11] they mentioned it happens on old HP devices [13:12] and im using old dv5 [13:26] apachelogger: ok, I'm running 2.7.0 now, it crashes as well :( [13:31] darn QtScript [13:41] why is it not crashing for me then Oo [13:41] markey: same bt? [13:41] yep [13:41] well [13:42] let me get you a BT from that one [13:42] it's coming from this Free Music Charts script [13:43] it seems to crash a little less often. maybe only 1 in 5 startups [13:43] that could be coincidence though [13:43] apachelogger: http://pastie.org/8000533 [13:46] markey: does it crash always for you? [13:46] because even with master it only crashes after a while or perhaps not at all (or I just didn't wait long enough) [13:46] apachelogger: no, as I said. maybe 1 in 5 starts crashes [13:46] right [13:46] sadly [13:46] race condition :P [13:46] Heisenbug... [13:46] or reentrancy [13:46] weeh [13:46] yes... [13:47] or still architecture specific memory crap [13:47] #12 Phonon::Gstreamer::Pipeline::cb_error (bus=, gstMessage=0x7f2159a92ca0, data=0x25c6a90) at ../../gstreamer/pipeline.cpp:486 [13:47] it's curious that pgst would get the error callback still [13:48] so I guess I'll put my money on memory crap for the time being [13:48] bug in QtScript you mean? [13:49] could be that something messes up the heap and gstreamer goes "fu" and then jit goes "fu, I am dead" [13:49] markey: for example [13:49] I mean it would be a bug in qtscript anyway [13:49] this is a nightmare :( [13:49] yes [13:49] markey: talk to #qt people maybe [13:51] markey: tell them we have http://bazaar.launchpad.net/~kubuntu-packagers/kubuntu-packaging/qt/view/head:/debian/patches/kubuntu_38_revert_fix_jit_crash.diff [13:51] albeit that patch should have fixed this trace [14:18] apachelogger (and markey): i dont think that revert/patch is needed or desirable in 4.8.4 anymore (the original 4.8.3 qtscript problem was fixed in a small followup, see https://bugreports.qt-project.org/browse/QTBUG-27322 and https://codereview.qt-project.org/#change,37896 which should be included in 4.8.4) [14:18] unless that reversion is something different, in which case, ignore me and carry on [14:18] yeah, no one ever figured out if it was in 4.8.4 [14:19] by reverting that, you may be seeing the original QTBUG-23871 [14:22] yofel: ^ [14:24] is this related to the Amarok crashes? [14:24] hmmm [14:25] it might be, I think [14:27] shadeslayer Riddell: my upgrade using muon is stuck on "Waiting for configuration file" for about an hour now...can I just kill it? [14:27] apachelogger: hm? [14:27] :S [14:28] mck182: attach gdb and check the backtrace ? [14:28] mck182: wait for JontheEchidna to tell you how to debug it ^^ [14:28] yeah ^^ [14:28] it should be telling you about a difference in an /etc/conf file on the frontend [14:29] JontheEchidna: nope, no such dialog anywhere on my desktop [14:29] :( [14:29] unless it's really cleverly hiding [14:29] that's pretty weird, since the frontend did catch that it's supposed to be prompting. I'll have to play around with that I suppose. [14:30] JontheEchidna: what do I do? [14:30] yofel: see what rdieter wrote above, regarding QTBUG-23871 [14:30] load file [14:30] populate(self) [14:30] JontheEchidna: could it be interesting to look at the process tree? [14:30] ruby is really creepy at times [14:30] * mck182 looks [14:30] JontheEchidna: to see if apt is stuck at some command line front end [14:31] It's waiting on Muon to present a "do you want to replace this conf file that you've changed with the updated one from the package" dialog, but Muon's not doing it, even though it knows it should [14:31] well, there's qaptworker with qthread/dpkg as a child [14:32] I've not found a good way to test conf file replacement, due to it requiring a package update to change a conf file [14:32] so it's possible that there are bugs like this I guess.... [14:32] JontheEchidna: I like how you talk about Muon as if it was a stuborn child :) [14:32] btw. probably the most stupid plasma bug ever - https://dl.dropboxusercontent.com/u/6761102/recorditnow_tmp.mp4 - two screens, trying to unmount my disk, no clicking involved [14:32] haha, it kind of is in some ways :P [14:33] JontheEchidna: so...what should I do with it? [14:34] mck182: you'll have to kill muon and qaptworker, run "sudo dpkg --configure -a" and run another upgrade [14:34] ohhh [14:34] oh well [14:34] mck182: awesome! [14:34] yofel: /opt/project-neon/share/project-neon/pkg-project-neon.mk [14:34] agateau: it is! [14:34] yofel: fiddeling the buildsystem will be fun [14:35] started that 3 weeks ago, gave up after 300 replaces of /opt/project-neon for /opt/project-neon-kf5 :S [14:35] apachelogger: looking at qt [14:35] agateau: until you want to actualy unmount the drive [14:35] mck182: so this is with the updater frontend? [14:35] yofel: groovy [14:35] JontheEchidna: I guess...? [14:35] I am checking out for today anyway [14:36] mck182: this is what "sudo umount" is for :) [14:36] apachelogger: that's a symlink to ../0/project-neon.mk, you should be looking at that and default-settings.mk [14:36]