=== jono is now known as Guest40438 [07:13] veebers: hey, still around? === alan_g is now known as alan_g|afk [10:24] howto apply ubuntu settings to compiz compiled from source? [10:24] I have compiled and installed compiz but when I login the only thing I get to see is the wallpaper [10:24] so I change to a VT and start compiz using: env DISPLAY=:0 compiz –replace composite opengl move resize decor compiztoolbox mousepoll wall expo animation switcher unityshell [11:32] Esokrates: you need to start compiz like this: [11:32] compiz --replace ccp & [11:33] the 'ccp' is important, it tells compiz to load the settings management plugin [11:33] which pulls in all the other plugins [11:33] and applies the stored settings [11:33] i remember to have tried this too [11:33] have got my mail? [11:34] umm, i install unity from source using dpkg-buildpackage and dpkg -i a running compiz --replace works fine for me [11:35] but yeah, that's a horrible workflow :-) [11:35] compiz --replace just kills compiz and does not start anything [11:36] Esokrates: compiz --replace *ccp* [11:36] yeah, i believe you its not working for you, sry [11:36] i think he has alried tried that too [11:36] alreadz [11:37] if you're doing it from a vt, you need to do COMPIZ_CONFIG_PROFILE=ubuntu DISPLAY=:0 compiz --replace ccp [11:37] (that will make it pick the gsettings profile and the saved ubuntu settings) [11:37] smspillaz, oh i did not know that one, i will try that === francisco is now known as Guest13450 [11:56] smspillaz, does not work [11:57] smspillaz, i get error like "A handler is already registered for /org/freedesktop ..." [12:01] smspillaz, everything runs fine, but i cannot adjust settings [12:09] luv, unity has nothing to do with that problem, as it already ran with default compiz before I compiled a patched branch from source [12:09] luv (a patched branch of compiz) [12:23] Esokrates: deactivate debus [12:23] *dbus [12:23] in ccsm [12:23] that will at least make it shut up [12:23] Esokrates: also can you post the full output somewhere? === _salem is now known as salem_ [12:47] Mirv: sil2100: published last part of daily release documentation! If you have any question that are still pending, do not hesitate to ask me, I'll add them to the FAQ :) [12:49] ok :) [12:55] Hi :) Could somebody please reapprove: https://code.launchpad.net/~mc-return/compiz/compiz.merge-fix1020822-gears-plugin-does-not-build-anymore/+merge/146285 [12:55] and for anyone wondering about where didier's documentation is: https://wiki.ubuntu.com/DailyRelease [12:55] * MCR1 forgot to adjust debian/compiz-plugins.install [12:56] MCR1: done, although click-missed Rejected first... sorry about that [12:56] Mirv: Thanks a lot. That was fast :) [12:57] no problem, I trust Brandon and the packaging fix seemed fine :) [12:57] Mirv: It is fine, don't worry ;) [13:05] smspillaz, http://pastebin.com/y47K1Ukk [13:05] smspillaz, http://pastebin.com/gHgWQLk9 [13:06] bregma: Hi :) There are a few serious Grid plugin problems with excellent bug reports left unfixed, which imho should really be fixed for 13.04. Would it be okay to add those to: https://blueprints.launchpad.net/ubuntu/+spec/desktop-r-unity-polish , so we won't forget them ? [13:07] MCR1, https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/1115341 [13:07] Launchpad bug 1115341 in compiz (Ubuntu) "Grid resize: Placing maximized windows on the top edge using shortcuts does not work" [Undecided,Confirmed] [13:07] MCR1, https://bugs.launchpad.net/compiz/+bug/1115344 [13:07] Launchpad bug 1115344 in compiz (Ubuntu) "Grid + Maximizing (button) + snapping by mouse to Grid position changes cursor position relative to window " [Undecided,New] [13:07] Oh, Esokrates - I did not know you are here :) Excellent reports, btw. [13:08] MCR1, https://bugs.launchpad.net/ubuntu/+source/compiz/+bug/1115350 [13:08] Launchpad bug 1115350 in compiz (Ubuntu) "Grid: Dragging screen centric windows (e.g. "Sim unlock required") causes cursor being changed relative to the window" [Undecided,New] [13:08] MCR1, https://bugs.launchpad.net/compiz/+bug/1082001 [13:08] Esokrates: That is how bug reports *should* be written -> the shortest & exact way to reproduce the bug -> should help a lot in fixing those... [13:08] Launchpad bug 1082001 in compiz (Ubuntu) "Grid/Expo: Corner-resized Grid windows confuse Expo and follow the user from workspace to workspace" [Undecided,Confirmed] [13:09] Esokrates: Yep, I wrote the last one ;) [13:10] MCR1, then you also know this one: https://bugs.launchpad.net/compiz/+bug/1115351 [13:10] Launchpad bug 1115351 in Compiz "Workspace switcher: Right clicking launcher items makes quicklist always appearing by changing the workspace (switcher) " [Undecided,Confirmed] [13:10] Esokrates: Yes, because I confirmed it today ;) [13:12] smspillaz, both outputs: i started compiz (on a VT) using the commands found in the pastebin and killed compiz using Strg+C === MacSlow is now known as MacSlow|lunch [13:21] didrocks: \o/ Reading it all up now since it's all finished ;) === dandrader is now known as dandrader|afk === dandrader|afk is now known as dandrader [14:10] fginther, have you seen http://10.97.0.1:8080/job/ps-unity-autopilot-release-testing/label=autopilot-ati/75/console ? looks like a bad log gathering script [14:18] mterry, looking [14:22] popey: do you experience bug #937118 still? [14:23] bug 937118 in linux (Ubuntu) "8086:0084 Wireless stops passing packets" [High,Expired] https://launchpad.net/bugs/937118 [14:24] mterry, looks like machine filed to install: "UTAHProvisioningException: Machine install timed out." [14:25] fginther: there were crashes on the indicator tests though (and the nvidia trick to avoid the xorg crash didn't work) [14:25] grrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrrr [14:25] fginther: btw, do you know if veebers is still looking at those? I thought he would fix/relaunch what's needed during the night when this happened [14:25] fginther, ah yeah. Right you are on UTAH. OK, I guess ignore that failure [14:25] mterry: can you please open a bug against UTAH? [14:26] didrocks, who owns indicators again? [14:26] mterry: with the date of the day, they want to track failures [14:26] didrocks, k [14:26] mterry: cyphermox, can you check with him? [14:26] I think we need uploading the crash xorg files and ping bryceh [14:26] and relaunch testing the stack [14:26] didrocks, xorg file doesn't help on nvidia really [14:27] mterry: there is one on intel though? [14:27] for the indicator stack [14:27] ah [14:27] didrocks, ps-indicators-autopilot-release-testing/label=autopilot-nvidia had the nouveau driver installed [14:27] cyphermox, you know how to grab the xorg crash file? It's in the build artifacts [14:27] fginther: urgh, we test on nvidia blob though? [14:28] didrocks, it is supposed to be nvidia. I wonder if something tragic occurring during provisioning [14:29] didrocks, and AIUI veebers is still looking at these, but he's on holiday now [14:29] fginther: ah, making sense :) [14:30] didrocks, I'll see what I can dig up on the nouveau instead of nvidia failure [14:32] didrocks, https://bugs.launchpad.net/utah/+bug/1116307 [14:32] Launchpad bug 1116307 in UTAH "Machine install timed out" [Undecided,New] [14:33] thanks mterry [14:34] cyphermox: are you looking at the indicator stack issue? [14:35] mterry: it's in the build artifacts, you mean, already grabbed? [14:37] cyphermox, yeah. But download it and put it somewhere bryce can see, and poke him about it [14:37] cyphermox, and make sure to get the intel one, since the nvidia one is usually not useful [14:37] mterry: cyphermox: I'll let you relaunch the 2 stacks, only for checking? [14:37] didrocks: sorry, was on the phone (still am) upgraded to raring, not seen the issue since [14:37] part 5 explains exactly what to run btw :) [14:38] popey: you're quite lucky, happening a lot here since a few weeks [14:42] didrocks: btw I do get that wifi bug too [14:43] you can disable wifi with the switch to reset the driver, no? [14:43] cyphermox: yep [14:43] not ideal though… [14:43] well no [14:44] #ubuntu-kernel pointed me to https://patchwork.kernel.org/patch/2007911/ [14:44] seems 3.7 is safe, not 3.8 [14:44] but it's likely an easy enough issue, it seems to be specific to like, RX queue or something [14:44] i get a different message [14:45] didrocks: see if you get a message like: iwlwifi 0000:03:00.0: fail to flush all tx fifo queues [14:46] seems to be the one thing that shows up everytime wifi stops working === MacSlow|lunch is now known as MacSlow [14:46] hum, not really [14:48] didrocks: tbh my laptop spends most of its time in a docking station so on wired network these days ⍨ [14:49] popey: yeah, hard to see wifi issues I guess then ;) [14:49] popey: I'm not the only one, it's already a start… [14:54] heh, intel was a compiz crash, nvidia was a xorg crash [14:54] cyphermox, please note that the nvidia test was using the nouveau driver [14:55] meh [14:56] do not hesitate if you need help to relaunch testing the stack [14:56] both would be using a free driver. that would mean both crashes are important [14:56] (even if the FAQ published today is giving the command) [14:56] didrocks: so in case of failure you just relaunch? [14:56] cyphermox: well, log bugs with the crash (retraced locally if possible) [14:56] and yeah, relaunch but don't rebuild [14:56] just use the "check with whole ppa" case [14:57] tbh I'm kind of running in all kinds of directions at once right now, but let's try to get this done now [14:57] mterry: can you give cyphermox a hand? I'm still catching up [14:57] cyphermox: your turn you will be next in that case :p [14:57] mterry: wait, I'll look at everything first, see if I can do it on my own [14:57] cyphermox, ok [14:58] cyphermox: please do paste the command first ;) [14:59] mterry: oh, did you get the pastebin of deploying a new component from yesterday? === francisco is now known as Guest8850 [15:06] didrocks: mterry: so what's the syntax for the cred file again? [15:06] cyphermox: you didn't receive my email with them? [15:06] ah it's in email ok [15:06] yep :) [15:10] didrocks, I got the pastebin, thanks [15:10] great :) [15:12] didrocks: so cu2d-run -R indicator-stack -r raring --check-with-whole-ppa would be the right command, yes? [15:12] err, wait, not sure about the stack name [15:13] cyphermox: the stack name is just "indicator", and you are changing the head (we will have raring when s… is opened) [15:13] cyphermox: so either: cu2d-run -R indicator -r head --check-with-whole-ppa [15:13] dah [15:13] or head is the default [15:13] yep yep [15:13] so you can remove the -r head :) [15:13] cyphermox: btw indicator*s* [15:13] would it make sense to alias head to the current dev release (or the other way around) [15:13] :) [15:14] so that it would be more obvious and you can then just use the release name as expected? [15:14] cyphermox: yep, please open a bug against cupstream2distro [15:14] sure, I'll even propose a patch [15:14] cyphermox: but normally for the dev release, you don't provide anything [15:14] great! [15:14] just cu2d-run -R indicators --check-with-whole-ppa [15:14] for instance [15:14] yeah [15:14] easiest :) [15:14] I'm just thinking that there might be other people one day to work on this [15:15] or releasing to multiple releases at once or whatever [15:15] yep [15:15] anyway, just waiting to get the confirmation that I got the necessary perms and I'll run that [15:15] didrocks: hmmm, btw! Since I noticed thomi's unity fixes for AP got into lp:unity - but I don't see the optimization changes re-introduced in lp:autopilot [15:16] sil2100: maybe reintroduce it? [15:16] didrocks: you think I should re-introduce them and just have thomi review and approve? [15:16] yeah, please do ;) [15:16] we have enough failures on the current daily release [15:16] thanks for looking sil2100 [15:16] sil2100: btw, did you try locally? [15:16] like, this doesn't indeed regress? [15:16] (if you get both) [15:17] didrocks: I'll try before pushing it further ;) [15:17] thanks! [15:17] np! Let's see now... [15:20] didrocks: done [15:20] cyphermox: so, as you can see the -check job is the only one running right now (with -head master job) [15:21] let's wait for the results :) [15:43] mterry: I'm merging indicator-bluetooth, are you going to do the bootstrapping too? [15:44] cyphermox, I hadn't thought about it :) I can sure, let me know once you merge [15:44] oh, I can do it too ;P [15:48] mterry: I think it's good now [15:48] didrocks: so we add to autolanding and everything? [15:50] cyphermox: yep, once the boostrap is done, I'll let you add the indicator-bluetooth to the stack [15:50] bootstrap* [15:54] cyphermox, do you need an approval on the bootstrap merge, or did you just do that directly? === chiluk is now known as chiluk_away === chiluk_away is now known as chiluk [16:09] mterry: well, we should be doing MPs if that's what you mean [16:12] it's really straightforward though so it's simple enough to approve quickly ;) [16:15] cyphermox: compiz crashed on ati in the last run it seems [16:15] cyphermox: mterry: can you retrace this for bregma? Seems we are starting to get regular compiz crashes now [16:16] as we don't have the dbgsym in the ppa ( :( ), I guess we need manual retracing on i386 [16:17] cyphermox, https://code.launchpad.net/~mterry/indicator-bluetooth/bootstrap/+merge/146664 [16:17] * mterry gets out his i386 virtualbox [16:19] mterry: I guess you need to apt-get source from the ppa and rebuild with nostrip or with pkg-create-dbgsym installed [16:19] didrocks, yup. it's a pain :) [16:20] mterry: yeah, apparently one day, we'll have the dbgsym in the ppa… [16:20] didrocks, that would be swell [16:20] basically, if we activate them right now, we won't have them in the distro [16:20] which is worse :) [16:20] mterry: if your setup is ready, maybe you can try to retrace the intel one as well, with a little luck, it's the same issue? :) [16:21] I'll see if either of them help [16:21] thanks :) [16:21] didrocks, that seems like a weird choice to be forced upon us (distro or PPA) [16:21] bregma: if you can put that one on the priority (top priority) as it seems those crashers are quite frequent and block daily release [16:21] didrocks, I wonder how it's implemented [16:22] mterry: "a bad dirty hack" told infinity [16:23] didrocks, I don't see the intel crash file? I do see it for the last ati build === dandrader_ is now known as dandrader|afk [16:26] mterry: in the previous run: http://jenkins.qa.ubuntu.com/job/ps-indicators-autopilot-release-testing/104/label=autopilot-intel/artifact/results/artifacts/_usr_bin_compiz.1000.crash [16:26] ah [16:27] didrocks: yeah I noticed the compiz ati crash [16:27] cyphermox: mterry is retracing them, I hope we can get those fixed so that we can daily release again [16:27] yeah [16:35] didrocks: didn't run all the tests yes, but only some of them - and it seems to work fine with thomi's changes [16:36] sil2100: great, please merge then :) [16:36] thanks ;) [16:41] hey mhr3, dee leaks! [16:41] :-D [16:41] kenvandine, in python? yes [16:41] sort of :) [16:41] in c, no :P [16:41] so i wrote a new friends-service in vala [16:41] which is the master for the model [16:42] memory usage is low! [16:42] and awesome [16:42] get_row leaks in vala [16:42] actually, scratch that, it can be double-freed when using from vala [16:42] the service never changes the model, it just loads it from the resource manager and shares it [16:43] that's all my dee leak knowledge [16:43] then the python based dispatcher populates it with data when it refreshes [16:43] if we leave the dispatcher running all the time [16:43] the service barely grows [16:43] but, if we make the dispatcher exit after all threads finish [16:43] everytime it gets started again to do a refresh [16:44] the service grows a bunch [16:44] hmmm [16:44] mhr3, and the larger the model, the bigger the growth each time [16:44] if the model has like 8000 rows in it, each time the dispatcher connects to it the service RSS grows by 40M or so [16:45] also, my qml based client causes the same thing [16:45] i'm not sure that's really a leak [16:45] it might be just fragmented memory [16:45] hmm [16:46] serializing 8000rows will eat up quite a bit of memory [16:46] yeahm but does that happen for each slave that connects? [16:46] yes [16:46] and not freed? [16:47] it does get freed, but if the allocator puts something on top of those 40mb, it wont return it to the os [16:47] oh [16:47] so you'll see ever increasing rss [16:48] but maybe there is a real leak, only valgrind/massif data will convince me :) [16:49] with G_SLICE=always-malloc pls ;) [16:49] ok [16:49] thx [16:59] mterry: did you succeed this retracing or compiz is slowly building in the vm? :-) [17:01] didrocks, I had to rebuild because my normal nostrip tricks didn't work on compiz. Rebuilding with pkg-create-dbgsym [17:02] mterry: oh really? weird, I was using DEB_BUILD_OPTIONS=debug,nostrip which is working with dh7+ [17:02] but yeah, I fall in love with pkg-create-dbgsym :) [17:06] didrocks, I've been cargo-culting this line: DEB_BUILD_OPTIONS=debug,nostrip,noopt CFLAGS="-g -O0" CFLAGS_APPEND="-O0" debuild -i -I [17:06] didrocks, for some reason I've encountered packages that need each of that at one point [17:07] I'm surprised it didn't work on compiz, I might have done something stupid [17:11] mterry: right noopt, CFLAGS are normally overriden though? I'm surprise this is needed sometimes [17:12] mterry: anyway, I'll check with infinity as those dbgsym exists somewhere if there is a dirty way to get them when needed without copying to the distro :) [17:12] didrocks, this could have been from years ago. ::shrug:: Like I said, I've been cargo-culting that line for a long time [17:12] a dirty way for an ugly hack ;) [17:12] heh ;) [17:14] Hrm [17:14] "Can't read pathname for load map: Input/output error." [17:14] didrocks, have you ever seen gdb do that? ^ [17:15] * mterry googles [17:18] mterry: weird, and if you specify the debug path manually? [17:25] didrocks, it sees the dbg symbols [17:25] didrocks, but I don't get a stacktrace. So I assumed that message was important. Google seems to think it's less important... [17:25] mterry: googline didn't really help here either, maybe check with doko? [17:26] yeah, it seems "minor" from what I saw, but if you didn't get a stacktrace… ;) [17:27] mterry: maybe you want to ask pitti for the correct dbgsym, assuming it's the creation being the cause? === Guest89634 is now known as jono === jono is now known as Guest96957 [17:35] fginther: is https://code.launchpad.net/~autopilot/autopilot/reintroduce-thomis-make-faster/+merge/146654 stuck? I don't see the autopilot autolanding job running [17:35] sil2100: FYI ^ [17:35] mhr3, http://ubuntuone.com/0AQvy3yDmcQCLz3Ejxn6pE [17:35] didrocks, looking === MacSlow is now known as MacSlow|dinner [17:36] mhr3, i guess it is the serializing === dandrader|afk is now known as dandrader [17:39] Just hope it won't be a problem that we're re-merging the same diff for the second time [17:40] Since I remember once compiz doing strange things during a merge once just because of things like this [17:43] didrocks, mterry, can you clarify a place where compiz is crashing in the AP tests so we can try to reproduce manually? [17:45] bregma, around 15:33:34 in this log: http://10.97.0.1:8080/job/ps-indicators-autopilot-release-testing/label=autopilot-ati/105/artifact/results/artifacts/ap_test_debug_log.txt [17:45] bregma, which doesn't seem like anything happened by that log [17:46] bregma, I got the timestamp from the .crash file [17:48] bregma, I'm not sure if that timestamp is after apport finished making the .crash file. So the crash itself might have happened a bit earlier [17:48] there's an awful lot of noise in that log since every single test is failing at startup due to the broken AP change [17:48] bregma, (the .crash file has a "Date" file inside) [17:48] bregma, yeah [17:49] bregma, let me see for the intel crash the day before [17:50] the .crash file shows a smashed stack (untraceable), which usually means either a bad function pointer, and ABI mismatch, or a data walker [17:50] didrocks, https://code.launchpad.net/~autopilot/autopilot/reintroduce-thomis-make-faster/+merge/146654 running now [17:50] fginther: thanks, are you warned by mmrazik's script about stuck jobs? [17:50] bregma, http://10.97.0.1:8080/job/ps-indicators-autopilot-release-testing/label=autopilot-intel/104/artifact/results/artifacts/ap_test_debug_log.txt [17:50] bregma, 02:42:50 is the Date on that one [17:51] didrocks, yes, this one had not triggered it yet [17:51] ok :) [17:52] bregma, all tests before that time look normal, except perhaps the very first one: unity.tests.test_panel.PanelCrossMonitorsTests.test_hovering_indicators_on_multiple_monitors [17:52] bregma, it may have taken that long for apport to process the crash file... [17:56] how could we confirm its not an ABI break? Could we just update the abi number and assert thats not the problem? [17:56] mterry: you'll fix the issue with your a11y branch? [17:56] cyphermox, oh... let me see what it is [17:57] cyphermox, that was jenkins being dumb. I'll set to approved [17:57] oh, right [17:58] alright :) [17:58] err, for bootstrap shouldn't it have been revision 42?? [18:00] cyphermox, eh, the other commits didn't have anything for jenkins. I guess 42 would have been fine too [18:00] or more accurate [18:01] cyphermox, in short yes, but it doesn't happen to matter [18:34] indeed, it doesn't really matter for this one [18:35] * cyphermox dials down the pedanticnessness === MacSlow|dinner is now known as MacSlow === dandrader is now known as dandrader|afk === salem_ is now known as _salem