[00:01] yeah === iveand is now known as rikshaw76 [03:48] hello [04:02] Mirv: hey [04:02] I just CC you for an email with the indicators failures [04:02] I'm off to bed now [04:12] Good morning [04:34] cyphermox: ok, thanks [04:34] morning pitti [04:34] hey Mirv [07:31] Morning! [07:32] hello sil2100 [07:45] pitti, hi [07:46] have you ever tried using python3-pam? it won't even load here ;( [07:46] darkxst: no, I didn't (neither the py2 variant) [07:46] py2 variant seems to work, but ubiquity-dm is py3 [07:47] sil2100: I contacted the SDK team about the unity8 problem now [07:49] sil2100: btw when time and before next tick, please check the merge requests from the status page [07:50] pitti, is there some that is needed to register the module? [07:51] in python3 I get ImportError: No module named 'PAM' [07:51] even though it is installed [07:57] Mirv: ACK [07:57] Mirv: thanks for dealing with that, I still am a bit amazed that this got pass, if they say it's as bad as it is [07:58] But anyway, damn, when you started it was all broken anyway, and now even more work... [07:58] guten morgen [07:59] Mirv: the problem seemed that SDK didn't run any unity8 tests, right? Since I guess if it released, the change didn't break any SDK or Apps tests [08:00] Mirv: sadly it's not possible for us to run unity8 tests alongside SDK ;/ [08:04] sil2100: yeah, the apparmor+dbus uploads from yesterday seem to cause breakage all over [08:04] morning Laney, while testing the background change functionality in the settings app, I changed the background at some point, but I cannot change it back to the default now. So in summary: help, I'm stuck with the snail background! :) [08:04] sil2100: yes, the mergers only run unit tests which is at least part of the problem [08:04] dpm: we disabled that for now, so that'll be why [08:05] in fact it never really worked [08:05] do a dconf reset on the key [08:05] which one is it? [08:05] erm [08:06] maybe /org/gnome/desktop/background/picture-uri [08:07] * sil2100 sighs [08:16] good morning desktopers [08:19] just for info, new dbus in saucy-proposed is creating issues with tests using dbus [08:20] Morning! [08:20] tyhicks is working on it [08:20] "the test cases that are ran on the buildds are broken because dbus-daemon now calls aa_is_enabled() in libapparmor" [08:20] "on the buildds, that is returning an error and dbus-daemon is refusing to start :/" [08:20] " that problem is easy to fix and I have a patch in the works - I need to verify that the other test failures are caused by that same issue" [08:20] [08:20] [08:20] hey sil2100 [08:22] pitti, hey [08:23] hey seb128, bonjuor [08:23] pitti, salut, ça va bien ? [08:23] seb128: FYI, I removed the broken libgtkada from -proposed again [08:23] seb128: oui, et toi ? [08:23] Thanks for the update! Some of our problems might have been also caused by that [08:23] pitti, très bien, merci ! [08:23] pitti, thanks [08:23] seb128: so jenkins is green again, and the FTBFS not an issue any more [08:23] I jump out for a moment now, since yesterday after UDS it was already too late to do anything ;p [08:23] seb128: ah, I'll as jibel about the excuses for cairo [08:23] pitti, weird, http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html still says it is [08:24] pitti, did you just remove it [08:24] or do we need jibel to trigger an update? [08:24] * Laney begins updating gst [08:24] seb128: cf. #u-devel [08:24] when things are updated is still magic to me [08:24] danke [08:25] sil2100, how did vUDS track hosting go? [08:25] Laney, hey, good luck with the update ;-) [08:25] scared [08:41] sil2100, Mirv: bug #1051921 is in the sponsoring queue, does it need to be or is that going to be part of the next precise SRU? [08:41] Launchpad bug 1051921 in Unity 5.0 "lens-bar-keynavigation periodically writes to /tmp/wut.png" [Medium,In progress] https://launchpad.net/bugs/1051921 [08:44] seb128: that's been handled by others. if it's a potential security issue then cherrypicking would be ok. the 5.0 branch otherwise is stalling on the XIM question. [08:45] (and testing etc as usual for old style SRUs) [08:52] larsu, sil2100, Mirv: I deleted the buggy dbus from saucy-proposed (they are going to reupload it once they fixed the issue), next indicator run should be green [08:52] sil2100, Mirv: let's wait for a publisher run (so the change is in effect) and we can retry the indicators stack [08:58] ok. first on my list is the sdk stack which is about to be published soon, hopefully after the publisher run [09:00] right, it should be fine [09:05] seb128: I'd guess publisher run has now happened, or how often does it? [09:06] Mirv, "less than half an hour", you can check with "rmadison -S dbus", doing that [09:08] Mirv, not done yet it seems [09:08] it should say 1.6.12-0ubuntu2 for proposed as well, or nothing? [09:08] ok, let's wait [09:11] Mirv, it shouldn't have a proposed line [09:13] Mirv, ok, we are good [09:18] so it seems, relaunching indicators [09:31] sil2100: please, the MP:s :) [09:38] well, I can deploy the phone config anyhow [09:41] *grmbl* why are we using this hangouts crap. Still doesnt work [09:41] Sweetshark: because there is no free replacement [09:42] but yes, it's relatively crap from user experience point of view (at least on x86 linux) [09:43] Mirv: mumble is a perfect replacement, if you can live without looking at 6-10 distracted nerds that never look straight at the screen but all over the place to check IRC, etherpad, blueprints and whatnot. [09:44] Mirv: seems to be worse this time I tried to join a few hangouts yesterday to watch and they dropped [09:44] Sweetshark: technically correct answer [09:44] czajkowski: ok, I had no problems but maybe there are networking issues within the infrastructure or something [09:44] possibly. [09:45] xnox, lp:~darkxst/ubiquity/lp1204312 [09:45] Sweetshark: it'd need autouploading of recordings somewhere, though (mumble) [09:45] and live streaming of pure audio without client [09:48] Mirv: ;) [09:51] seb128: now there's some new package in indicators stack, gnome-control-center-datetime, needs new config and run [09:52] Mirv, :-( [09:52] I pinged sil2100 and cyphermox about it yesterday evening [09:52] cyphermox said he would handle it [09:52] seb128: additionally, the cu2d still thinks indicator-network is eternally running on powerpc, whil actually it's a dependency wait on qmenumodel-dev [09:52] which also started yesterday evening at some point, and makes indicator build job to stall eternally. I don't know where it's coming from. [09:53] I've no idea about that :/ [09:58] me neither :( [10:00] seb128: it's somehow related to this latest commit http://bazaar.launchpad.net/~indicator-applet-developers/indicator-network/trunk.13.10/revision/276 which adds the libqmenumodel-dev dependency [10:00] Mirv, that makes sense, it used to build fine on ppc before using qt5 [10:03] Mirv, did you try https://wiki.ubuntu.com/DailyRelease/FAQ#I_know_that_I_want_for_one_time_skipping_one_arch_where_my_package_won.27t_built_for [10:03] Mirv, I would skip it once and then we can delete the ppc binary in the archive [10:04] seb128: not yet, if it works it should be tried now when this tick is in that shape [10:04] Mirv, yes, try that rather than stopping the run next time [10:04] sil2100: can you work on that / follow the builds, I should eat at some point? :) [10:06] sil2100: https://code.launchpad.net/~timo-jyrinki/cupstream2distro-config/indicators_add_extra_gnome-control-center-datetime/+merge/182578 [10:06] deployed already for this run [10:18] Mirv, approved [10:18] Mirv: approved [10:18] Ah, ok [10:18] lol [10:18] sil2100, sorry, I was not sure you were around so I just changed it [10:18] it's almost lunch time after all ;-) [10:19] Mirv: I'm taking over everything now, my system seems back usable now [10:19] (my tick!) [10:19] TRAPPED IN A A GUEST SESSION, SEND HELP [10:19] (the no indicators bug :P) [10:19] lol [10:19] sil2100: thanks, it's been a rough morning :) [10:19] /noon/beginningofafternoon [10:20] * seb128 sends laney a ctrl-alt-del [10:20] Laney: /usr/lib/unity/unity-panel-service & ;) [10:20] or that [10:20] in alt-f2 [10:20] no, you guys suck! [10:20] gnome-session-quit [10:20] :< [10:20] ;-) [10:20] Who would like to quit such a session?! [10:20] we would definitely need an Asian / Australian / NZ guy to the cu2d vanguards [10:21] well I now have a black screen with a mouse pointer [10:21] the guest session was definitely better than that [10:21] I think we'll have to transfer someone from our team to NZ [10:21] Laney, logout is quite buggy in saucy, I need to debug it at some point [10:21] in general or in the guest session? [10:21] it does the same on my normal session, I often end up going to a vt and doing ctrl-alt-del [10:21] hmm [10:21] haven't seen it there [10:22] I usually shut down straight away though [10:22] that works [10:22] it's logout that is buggy [10:22] yeah [10:22] and it's like every 3rd try [10:22] sometimes I do that but I didn't notice it hanging like that [10:22] it tends to do it if you try to log out just after log in [10:22] then it goes downhill [10:22] ah I have a unity-greeter crash [10:22] e.g first nautilus seems to exit [10:22] e.g you get no background [10:23] my shutdown woe is bug #1212142 [10:23] Launchpad bug 1212142 in sysvinit (Ubuntu Saucy) "Slow shutdown/restart on saucy with sendsigs waiting 10 seconds" [Medium,New] https://launchpad.net/bugs/1212142 [10:23] then stuff hang for a minute [10:23] then unity goes away [10:23] then you have an empty screen for another minute [10:23] * Laney reports this crash [10:23] although I'm not sure if it's not been there in various forms for ages, I just did the workaround on 12.04 LTS machine as well to make shutdown nonslow [10:24] we've tinkered a lot in that area this cycle [10:24] upstart / logind [10:24] upstart isn't relevant for the guest session though I guess [10:25] Laney, upstart is what makes you have no indicators :p [10:26] lack of upstart ;-) [10:26] I need to ping robert_ancell about bug #1214504 [10:26] Launchpad bug 1214504 in unity (Ubuntu) "unity-panel-service not running in non-upstart-sessions on Saucy (including guest session)" [High,Triaged] https://launchpad.net/bugs/1214504 [10:28] oh, interesting that it works if you go from lightdm [10:28] didn't know that [10:31] right [10:34] is taking screenshots broken for anybody else? [10:36] http://ubuntuone.com/3616CovHk0RgyA7OvflvF2 [10:36] nope, not me [10:36] tsdgeos: yes but it does fix itself [10:37] if you either A) restart your machine B) do some other restart that seb128 and didrocks told me [10:37] restarting the machine never heps here [10:37] it helps if kill unity and run Plasma though :D [10:37] wfm here, it just sometime gets confused by screen config changes [10:37] e.g if I undock my laptop [10:42] gst done, albeit blocked in proposed until we get libav 9 [10:43] k [10:43] seb128: I don't even dock my machine and it gets confused [10:43] very special [10:43] yeah, no idea about it [10:44] try scrot from the repos [10:45] between compiz eating my cpu today and thunderbird gorging on it I am going to strangle my laptop! [10:46] slomo: ^^^ gst done, fyi [10:47] Laney: sorry, i'm making 1.1.4 tarballs now already ;) you can update it again later today :) [10:47] hahaha [10:47] slomo: much easier to take it from debian packages ;-) [10:47] Mirv, sil2100: skipping indicator-network ppc [10:48] Laney: those i'll do tomorrow :) [10:48] perfect [10:48] larsu, did you look at the accountsservice's vendor stuff/do you know how it works? [10:49] mterry might be able to review that [10:49] he'll be integrating it into the greeter after all [10:49] good point [10:49] let's wait for him [10:50] seb128: I know how it works in theory, but I haven't looked at the patch [10:51] larsu, don't bother, we are going to get mterry to review it, he's on the other side of the settings anyway so it makes sense ;-) [10:51] okay :) [10:51] seb128: k [10:54] k [10:55] sil2100: note that your tick is again being blocked by the indicator build phase, everything waits on that [10:55] Laney: so changes are only the camerabin2 and jpegformat move? [10:56] sil2100: you should abort that, let others finish and then rerun [10:56] Mirv: can't we just skip the indicator? [10:56] Like, with cu2d-skip? [10:56] Mirv: that doesn't work? [10:57] sil2100: try it out, sure [10:57] sil2100: I just quickly noted before my telco [10:58] since I believe nothing significant has yet happend during the last hour progress wise there, because of the stall [10:58] slomo: there's also a bump to the opencv version which is already upstream [10:58] Mirv: cu2d-skip worked \o/ [10:59] slomo: oh and that libtool one that you didn't like [10:59] Mirv: well, it didn't stall for too long, we lost around 10 minutes (since before datetime was still building) [10:59] sil2100: right, great! can you paste me the exact command line you used?-) [10:59] Mirv: ./cu2d-skip indicators indicator-network powerpc ;) [10:59] cu2d-skip indicators indicator-network powerpc? [10:59] (without the ;) in the end :D) [11:00] ;) [11:00] Yes [11:00] right, cool :) [11:00] Laney: libtool one? [11:01] slomo: https://bazaar.launchpad.net/~ubuntu-branches/ubuntu/saucy/gst-plugins-bad1.0/saucy/view/head:/debian/patches/libtool-force-link-lc [11:01] I didn't get to the bottom of where the circular dep was coming from [11:01] Laney: ah, i'm going to add that to the debian package too now [11:01] ok [11:01] are you hitting that problem too? [11:01] sil2100, Mirv: indicators are good to publish [11:01] Laney: don't have time for debugging the real cause of this... i still think it's a toolchain bug [11:02] it's definitely a change in the toolchain but I can't say if it's a bug [11:02] Laney: yes, i386/kfreebsd-i386/ppc... nowhere else [11:02] * Laney nods [11:02] sil2100, Mirv: " Mirv, sil2100: skipping indicator-network ppc" ... when I wrote that I had run the skip command [11:02] doko told me to fix the software but he always says that :P [11:02] Laney: it's probably a feature ;) [11:03] sil2100, +1 for the indicator diffs if you want to publish [11:03] Laney: yes, that seems to be the default mode... you have to prove first that it's a toolchain bug, otherwise it's your bug ;) even if a toolchain change caused it to switch from working to broken [11:03] yeah... [11:04] seb128: ACK [11:04] seb128: ah, I thought not in there but somehow otherwise. ok. [11:04] seb128: published, thanks! [11:04] yw! [11:05] seb128: btw. by any chance, did you pre-NEW and whitelist cordova-ubuntu and cordova-ubuntu-tests already? ;) [11:05] sil2100, no, I didn't [11:06] I think those are packages that robru prepared [11:06] Blocking webapps stack ;p [11:09] sil2100, can we get those off the stack? [11:09] there is a ton of files in there with random licences [11:09] I'm not going to have time to NEW review them before lunch [11:10] Righto [11:15] seb128, Mirv: first thing: https://code.launchpad.net/~sil2100/cupstream2distro-config/disable_cordova/+merge/182602 [11:15] Mirv: http://www.livestream.com/mumblestreaming suggests that the pirate party germany already has some implementation of streaming for mumble. And converting a live stream to a ogg dump should be trivial. [11:19] seb128, Mirv: and this as well, since we want to have a mention of this rebuild in the changelog: https://code.launchpad.net/~sil2100/libunity-webapps/rebuild_libunity/+merge/182604 [11:27] sil2100: ah so is that latter in archive already? [11:27] sil2100: the former approved [11:27] sil2100: force_rebuild should work, yes, if not such a situation [11:28] Mirv: I know, but as I mentioned, if I do a force_rebuild, there will be no changelog entry why we did the rebuild, no? [11:28] sil2100: there's a changelog entry "forced rebuild" [11:28] Mirv: there is? Oh [11:28] but yes, no packagekit mention [11:29] Mirv: I guess I'll invalidate the merge request and do the force_rebuild [11:29] Since no use to loose time [11:30] Mirv: thanks for approving! [11:30] Redeploying and re-running in a moment [11:54] Damn... [11:54] What's with those LP network problems?! [11:55] httplib2.ServerNotFoundError: Unable to find the server at api.launchpad.net [11:55] Will have to re-run some stacks [12:02] Mirv: sadly... https://code.launchpad.net/~sil2100/cupstream2distro-config/platform_extra_click_pkgs/+merge/182616 [12:02] Mirv: new extra packages coming from click :o [12:05] approved [12:06] Thank you! [12:06] Redeploying [12:13] Mirv: sometimes I wish we had a better way of dealing with that 'extra packages' thing, I think I need to sit down and think of a better way [12:15] sil2100: yes, it's definitely already in next cycle's 'to-be-thinked-about' plans, with occasional discussions here. didrocks has the best thoughts on it as usual. [12:15] those seems like they should put the stack on UNSTABLE/manual approving rather [12:15] Mirv: https://code.launchpad.net/~sil2100/cupstream2distro-config/webapps_extra_packagekit/+merge/182619 [12:16] e.g "things went ok, but you have an extra package pulled in, do you want to publish" [12:16] seb128, Mirv: right, since right now it's sometimes blocking a stack for longer, or at least putting additional 'paperwork' work on us [12:17] the system could be more clever for sure [12:17] like gnome-control-center-datetime is a binary produced by one of the packages of the stack [12:17] those should be ok without being manually listed [12:17] seb128: we can also manually override it in the jenkins job temporarily when there's something important (without the need of modifying cu2d-config and redeploying), but it's still just working-around a problem [12:19] Mirv: pretty please when you have a free moment ;) ^ [12:21] sil2100: there's no such thing.. :) [12:22] approved [12:22] the final Qt 5.1.1 just got released an hour ago [12:22] happily it's identical to the snapshot I already packaged [12:26] :D === dednick is now known as dednick|lunch [12:38] Mirv: do you remember if we set a stack to manualpublish, does it require a redeployment? ;/ [12:38] Mirv: i.e. do I have to redeploy now all the stacks just because I want all of them to go into manual mode? ;p [12:39] you shouldn't listen to asac :p [12:39] that manual mode stuff is crazyness [12:40] its the opposite ;) [12:40] the path to sanity [12:40] stop landing stuff [12:40] i am not stopping landing stuff [12:40] sure, until people get on us because stuff are not landing [12:40] i am just stopping automatically landing stuff so we can coordinate the landings [12:40] we will manually land stuff [12:40] one by one [12:40] stack by stack [12:41] can you provide us the resources to do the manual work as well? [12:41] first those that need something in soon [12:41] those manual review are not going to automagically happen [12:41] sil2100: no, I don't remember [12:41] who is going to do them [12:41] especially during vUDS hours [12:41] seb128: well, we currenlty do it automatically [12:41] so we just be an expanded bot [12:41] asac, keep in mind that nobody is going to do manual review for the next 18 hours at least [12:42] yeah i hope we can go to automatic after we see an image [12:42] asac, it's going to be vUDS then eod [12:42] But but but [12:42] asac, I bet you that by tomorrow morning people are going to complain about "why are stuff not landing" [12:42] seb128: i sent a mail [12:42] to ue-leads/management [12:42] yeah, still [12:42] send them to me [12:42] I read it [12:42] if they complain [12:42] k, I can do that [12:43] Well, seb128 is right, since me and Mirv are the ones doing most stack management, cyphermox as well [12:43] thats what i said anyway in the mail :) [12:43] its not only you who is on vUDS [12:43] everyone is [12:43] seb128: that's why I asked asac to push that info to management and leads [12:43] we will still juggle the most essential stuff [12:43] If it happens [12:43] the goal is to go back to automatic [12:43] :) [12:43] Since I don't want to have blood on my hands! [12:44] i just dont believe its better to let all stuff land [12:44] let's see [12:44] it's going to be interesting [12:44] uncontrolled [12:44] mir didn't land for almost a week [12:44] yeah its a first experiment [12:44] they are not going to be happier [12:44] yeah. those i woul dlike to take a mnual look at [12:44] i dont understand their case [12:44] and why it didnt land [12:44] for now i assume our gates have catched stuff :) [12:44] because they keep changing abi [12:44] catched it upstream, which is a good thing [12:44] so unity-system-compositor and other stuff needs to be rebuilt with the new abi [12:44] seb128: I guess it should land now, as the jobs are green ;p [12:45] seb128: yeah. so thats a valid reason to no land from the looks [12:45] which requires us to publish several of the other stack to unblock mir [12:45] Indeed [12:45] not sure that i am particularly unhappy to nhot have those abi breakages in the image right now :) [12:45] mir is not in the image [12:45] seb128: ok so its about unfortunate split of stacks? [12:45] yes [12:45] e.g. not all the right pieces can live in their own area? [12:45] we need the unity stack to go through publishing [12:46] because it has some components that needs rebuilding with the new mir [12:46] yeah...i think we should change that and introduce like a "one-time-landing-stack" [12:46] e.g unity-system-compositor [12:46] where such big things can be staged properly in one shot [12:46] yeah understood [12:46] the problem is that stack as grouped as they are because component are inter dependant [12:46] you can't land a new unity without retesting indicators [12:46] because unity changes could break indicators [12:46] yeah i get the idea [12:47] It's a big web of inter-dependent components sadly [12:47] well, anyway [12:47] right... it always as like that, so lets check how we can improve our concept [12:47] let's see how things goes [12:47] to better capture those cross-stack transitions [12:47] asac, keep in mind also that beta1 is next week and the archive is going to be frozen tomorrow evening [12:48] Mirv: I'm getting tired of those 'extra packages installed' bugs ;/ [12:48] archive = main [12:48] but that includes the indicators [12:48] and unity [12:48] and mir [12:48] so we should try to get those juggled in manually [12:48] in a working form [12:48] i dont think automatic landing would give us that :) [12:48] well, as said manual is fine [12:48] we just need resources to do the manual work [12:48] seb128, Mirv: asking just in case I missed something in the morning - is it safe to publish the SDK stack? [12:49] right. but the point is that i dont see how manual creats more work [12:49] sil2100, it was published earlier? [12:49] it just crewates the work visibly up front [12:49] rather than in a hidden firedrill afterwards :_ [12:49] seb128: yes, in the morning it seems [12:49] well, it implies somebody to look, review and press the button [12:49] By Mirv's manual re-publish [12:49] so somebody around to do that [12:49] sil2100, is there any change since? [12:49] thats fine. i am sure we can do the button push pretty cheaply and we get more control [12:49] during this critical time of development [12:49] which should give us a net win :) [12:50] let's see [12:50] seb128: we rely on brave folksa everywhere. we will work on creating more redundancy on all stages of the CI engine [12:50] so we can go off on holiday :) [12:50] or sleep even [12:50] \o/ [12:50]