=== racarr|desert is now known as racarr === thumper is now known as thumper-afk [05:35] Good morning [05:41] pitti: morning. exclusion of .la files from dh_install in indicator-datetime http://pastebin.ubuntu.com/6057779/ (while seemingly keeping .a) [05:41] Mirv: that's adding .a files, not keeping them [05:42] Mirv: and as indicator-datetime doesn't have a -dev package or anything like that, it should be verified that it doesn't build *.a files [05:42] (probably it doesn't and the previous removal was just a copy&paste leftover) [05:43] Mirv: is the result of that in some PPA? [05:43] pitti: yes, depending on perspective. it was not keeping both, now it's keeping .a, but... it seems the packages don't really have any .a files so it was apparently not needed [05:43] or a test build? [05:43] pitti: https://launchpad.net/~ubuntu-unity/+archive/daily-build/+sourcepub/3459916/+listing-archive-extra [05:43] pitti: these have always been successfully built and autopilot tested [05:43] if they've gotten into phase that packaging changes need acking [05:44] Mirv: checked that the debs don't have .a, so +1 from me [05:44] pitti: thanks [05:54] Good morning [07:04] Morning! [07:05] morning sil2100 [07:05] sil2100, Good morning [07:13] jibel: how was the yesterday run of the stacks? The 18 UTC tick? [07:14] sil2100, daily-release tests are not going well. I investigated SDK tests yesterday evening and AP just wait forever in the middle of the run after a crash during a test. But I haven't found why it is not proceeding with the rest of the testsuite. [07:15] jibel: ok, so that's reproducible then - so it's like the one time I noticed it [07:16] jibel: you think it can be directly releated to how the tests are written, or maybe something gets stuck (like dbus) on the test machines? [07:16] jibel: I filed a bug about the SDK failing tests today, ie. the apps tests, I believe they're related to yesterday's commits but maybe you're talking about something else? [07:16] or precisely that? [07:16] Mirv: ^ as you see, yesterday it was a terrible day for testing things... [07:17] sil2100: well it's been terrible for some time with the unity tests taking eternally (well, 7200 seconds to be exact) because they hung [07:17] Mirv: not sure if we're talking about the same thing, but yesterday both AP machines were hanging up on the SDK tests, after a failed test they were standing and doing nothing [07:18] If the same happens for other stacks, I would guess it's the machine's fault [07:18] sil2100: yeah right so it's different, there's another problem today but it's upstream problem [07:18] some stacks are going just fine now [07:19] I released indicators in the morning as well [07:20] there is a file conflict in webapps, but that's a different problem [07:20] /var/log/upstart/otto-setup.log: dpkg: error processing /var/cache/apt/archives/unity-webapps-amazon_2.4.16daily13.06.20-0ubuntu1_all.deb (--unpack): [07:20] /var/log/upstart/otto-setup.log: trying to overwrite '/usr/share/unity-webapps/userscripts/unity-webapps-amazon/manifest.json', which is also in package unity-webapps-common 2.4.16+13.10.20130829.2-0ubuntu2 [07:25] sil2100, Mirv SDK is blocked again, same test than yesterday evening [07:25] http://paste.ubuntu.com/6057961/ [07:26] jibel: right, so it's really hanging there. can I stop those jobs so that other stacks can continue? [07:26] pitti, access to ftpmaster is open again, I re-ran update-manager, it picked the right version and it's back to green [07:27] jibel: splendid, that was fast [07:27] Mirv, I'm stopping them, I'd like to archive the run [07:28] jibel: thanks [07:29] Damn... [07:29] Mirv, done, unity8 is running now [07:30] Mirv, jibel: is this happening only for SDK? You think it might be caused by some invalid AP test? [07:31] sil2100: well it happens in webbrowser app test so it should happen elsewhere.. unless of course the recent UI Toolkit change is what makes it act like so [07:32] Mirv: I think I saw the Apps stack yesterday go through correctly, only with a normal test failure, so maybe indeed something in SDK? Could you file a bug? [07:33] sil2100, only for SDK apparently, but the same test on the apps stack fails with "error: [Errno 98] Address already in use" so we cannot say [07:34] sil2100: I should probably ask if you've checked the stack status page before going to jenkins? ;) [07:36] sil2100: but we all fail updating the stack status anyway, we should fix that at some point so it's actually up-to-date during each tick [07:37] I try to but what I for example don't do every time is try to validate the previous claims in there [07:41] Mirv: I didn't go to jenkins yet, as I don't want to 'interfere' with your tick ;) [07:41] Mirv: I was just asking about the overall status considering yesterdays apocalypse and overall maddness [07:41] With people screaming running around [07:41] ...wait, that was actually just me [07:44] sil2100: haha :) [07:44] sil2100: it's "ok" now with these isolated problems, while other stacks are running and publishing... for now [07:56] Mirv: the problem was that seb wanted the new SDK stack released badly ;) [07:56] Mirv: while we were unable to do that yesterday evening, and even now as well - but maybe it's a good thing, since if there's something broken then bleh [07:58] Mirv: I also hope unity can be released finally - yesterday we had one tick where unity passed the check job, but on one machine only [07:59] I also didn't know if the test run wasn't somehow affected, so I didn't publish - but there are some changes pilled up that would be nice to release [07:59] morning! [07:59] hey Laney, how are you? [08:00] does anyone have a bluetooth keyboard and/or mouse here? === tkamppeter_ is now known as tkamppeter [08:09] hey pitti [08:11] good morning desktopers? [08:11] ! [08:12] Good morning seb128 [08:12] hey jibel Laney, how are you? [08:13] seb128, quite fine and you? [08:13] I'm good thanks [08:15] bonjour seb128 [08:15] pretty good [08:15] pitti, salut, ça va bien ? [08:15] seb128: oui, et toi ? [08:17] pitti, très bien merci [08:21] pitti: I've one BT mouse although I use a logitech usb receiver mouse instead [08:21] if I click Bluetooth 'on' in saucy g-c-c, it just switches back to off for me, though [08:22] Mirv: ah, thanks; the Logitech ones have their own proprietary protocol, so that doesn't help in that case [08:25] does unity receive updates in the lts point releases? [08:32] sil2100, Mirv: hey, what's the status of the sdk stack? [08:33] seb128: tests failing when running apps, bug filed https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/1220056 (and given to SDK team directly) [08:33] Launchpad bug 1220056 in Ubuntu UI Toolkit "Multiple failing apps autopilot tests with the new ubuntu-ui-toolkit build" [Critical,New] [08:33] seb128: probably related to the optionselector changes yesterday [08:33] Mirv, do you know what change is causing that and is somebody working on that? [08:33] sil2100, unity tests are stuck in a loop http://10.97.0.1:8080/job/autopilot-saucy-daily_release/1508/label=autopilot-intel/console [08:33] mlankhorst: not specifically for point releases, but generally yes 12.04 had one in the spring [08:33] jibel: unity as well?! [08:34] Mirv: ^ [08:34] sil2100, yet another issue [08:34] jibel: sil2100: yes that I think I even e-mailed about as well, bug filed [08:34] Mirv: yeah it's going to need one for saucy [08:34] that has been since Friday [08:34] bug #1219636 [08:34] Launchpad bug 1219636 in Unity "unity autopilot tests fails to move mouse, leading to infinite test run" [Critical,New] https://launchpad.net/bugs/1219636 [08:34] Mirv: it needs a backport of pointer barriers, with runtime switching between old api and new saucy api [08:35] Mirv, jibel: let's stop this run maybe [08:35] sil2100: I've stopped it for now in the earlier runs [08:35] so go ahead [08:35] it does abort after two hours automatically, but it's too long [08:36] sil2100: that bug is as a blocker on the stack status for unity [08:37] sil2100, I stopped it [08:37] Mirv: hm, yesterday it didn't happen actually, at least at the afternoon ticks [08:37] jibel: thanks [08:37] Mirv, do you know what change is causing that and is somebody working on that? [08:38] Mirv, there is no video because recordmydesktop encodes the video when the test finishes, not on the fly [08:39] (was reading your comment on the report) [08:42] sil2100, Mirv: who did you guys ping about the sdk stack issues? [08:42] * seb128 trying to get a reply, we need that stack to land [08:44] sil2100: the commits were yesterday evening [08:44] seb128: bzoltan directly, he said to be looking [08:45] Mirv, thanks [08:49] hi seb128! [08:50] chrisccoulson, hey, how are you? [08:50] seb128, yeah, not too bad thanks. still busy :) [08:50] how are you? [08:50] chrisccoulson, I'm good thanks, really busy as well [08:50] rrreeeeaallly busy ;-) [08:51] heh :) [08:56] sil2100, hey, there? [08:56] seb128: hi, yes [08:56] sil2100, can you help pinging SDK people? [08:56] I'm not having luck getting replies for Mirv [08:56] from [08:57] seb128: will do! But it seems Zoltan is informed [08:57] sil2100, Zoltan asked if you guys pinged the gallery app people about the gallery tests [08:57] sil2100, right and he said the owner of the apps that have failed tests should be pinged [08:57] seb128: yes ;) It's all taking place on ubuntu-touch, but heh... it seems we have some more AP issues [08:58] seb128: and we might need someone from QA on this case as well, as the mediumtests are failing on jenkins while being fine locally (and on the machine) [08:59] QA being jibel? [08:59] seb128: it seemed like the apps weren't failing tests with the previous version of ui-toolkit, so it did not immediately seem like needed to ping the apps people [08:59] or one of the US guys? [08:59] sil2100, apps stack is affected by a similar issue than SDK but on a different test [09:00] Mirv, zoltan suggests that the default person to ping about app tests failing at app people ;-) [09:00] sil2100, http://paste.ubuntu.com/6058195/ [09:00] jibel: that hang you mean? [09:00] sil2100, yes [09:01] seb128: yes, that makes sense in general [09:01] seb128: it's a webbrowser issue, so let me ping oSoMoN maybe as well, but this is getting a bit confusing [09:01] sil2100, Mirv: ok, whatever seems right but please get the sdk stack to land today [09:02] seb128: roger! [09:02] sil2100, thanks ;-) [09:02] my biggest confusion is that sdk ran the apps tests successfully early yesterday [09:02] seb128: sure, one way or another... [09:02] Mirv, could you point to a run that succeeded yesterday? [09:04] jibel: aha.. green light, but skipped. what on earth has been going on with those tests. [09:04] the config has been unchanged itself, now it started running those tests and failing [09:05] I'm... I'm a bit confused right now in overall [09:11] the pieces start to drop in places I think now [09:11] just too many pieces and uncontrolled API changes [09:15] sil2100: what's up with those medium tests, then, any idea? [09:25] Mirv: I have no idea ;/ Last time it was a dep-problem, but now it's a mystery for me, I'm also unable to get much info from the logs there sadly as it's another, different architecture [09:25] ok, SDK team seems to be about to revert the change instead, for now... unless they change their mind and all apps get updated instead. let's see. [09:25] Ok [09:25] that would be the fastest way anyway, revert [09:25] sil2100, Mirv I'll abort apps tests. No need to wait for the timeout as next tick is in 35 minutes [09:25] Mirv: wait, to make things clear... will it also fix gallery-app? So that the mediumtests issue won't have to be fixed? [09:26] jibel: ACK... [09:26] jibel: ok [09:26] Damn, so much broken, broken everywhere [09:26] I see broken stacks [09:26] sil2100: I don't know, I think gallery-app still needs something since it wasn't broken yesterday but on Friday [09:27] Right [09:27] sil2100: we'd have Unity 'green' (yellow) again but it's still with only one autopilot machine. I don't know if we could trust nvidia machine enough? [09:27] we've been running with only a single machine before, though, but it's unfortunate intel doesn't succeed there [09:28] the problem is I've no idea how long it'll take for the intel specific problem to get fixed [09:28] Mirv: I had exactly the same problem yesterday, only nvidia was green IIRC [09:28] sil2100: yeah, it's because the intel one gets aborted after 2 hours, as discussed earlier [09:28] Mirv: let's do it like this, let's see if indeed all tests were run, check more or less the failures and publish if it's not terrible [09:29] I just worry about the hang-up issue [09:29] sil2100: shall we ignore the fact that webapps is also failing and technically unity depends on it.. [09:29] That we almost always ignore ;p [09:30] the nvidia is here http://10.97.0.1:8080/job/autopilot-saucy-daily_release/1508/label=qa-nvidia-gtx660/ [09:30] https://bugs.launchpad.net/unity/+bug/1219636 <- I just hope this is not a real regression, but actually there's not much going on on unity! [09:30] Launchpad bug 1219636 in Unity "unity autopilot tests fails to move mouse, leading to infinite test run" [Critical,New] [09:31] sil2100: it does not look worse than on last week when a successful release was made and intel was still running http://10.97.0.1:8080/job/autopilot-saucy-daily_release/1397/label=qa-nvidia-gtx660/#showFailuresLink [09:31] Indeed [09:32] sil2100, on this case it is possible that the test tries to access a launcher that is not visible and enters the infinite loop tryig to reach it [09:32] then we'd need three acks http://10.97.0.1:8080/view/cu2d/view/Head/view/Unity/job/cu2d-unity-head-3.0publish/lastSuccessfulBuild/artifact/packaging_changes_unity_7.1.0+13.10.20130903.1-0ubuntu1.diff + http://10.97.0.1:8080/view/cu2d/view/Head/view/Unity/job/cu2d-unity-head-3.0publish/lastSuccessfulBuild/artifact/packaging_changes_unity-scope-mediascanner_0.1+13.10.20130903.1-0ubuntu1.diff (just typo fix) + http://10.97.0.1:8080/view/cu2d/view/He [09:32] * sil2100 smiles to seb128 [09:32] seb128: ^ [09:33] jibel: that would make sense, and I guess it wouldn't mean it's broken for everyday users [09:34] sil2100, correct, if this theory is right, we'll just have to check what the unwanted additional launchers are. [09:34] Mirv, third link got cut by irc line limit [09:35] seb128: https://jenkins.qa.ubuntu.com/view/cu2d/view/Head/view/Unity/job/cu2d-unity-head-3.0publish/lastSuccessfulBuild/artifact/packaging_changes_unity_7.1.0+13.10.20130903.1-0ubuntu1.diff [09:35] seb128: that's the last one ;) [09:35] seb128: nope the last one was http://10.97.0.1:8080/view/cu2d/view/Head/view/Unity/job/cu2d-unity-head-3.0publish/lastSuccessfulBuild/artifact/packaging_changes_unity-lens-applications_7.1.0+13.10.20130903.1-0ubuntu1.diff [09:35] Wait, no, scratch that [09:35] right, unity was the first [09:36] sil2100, Mirv: +1 [09:36] Mirv: how dare you paste links in different order than it was on jenkins! [09:37] * Mirv published [09:37] \o/ [09:37] sil2100: :) [09:37] Finally! [09:37] SDK probably will get the revert, my main worry is that gallery-app [09:38] what with the gallery-app? [09:38] Mirv: I think we need someone from the QA team that knows his way around those parts, Francis would be the closest bet [09:38] But he starts in around 3 hours [09:38] seb128: this since Friday, and it'd still block SDK https://bugs.launchpad.net/gallery-app/+bug/1218953 [09:38] Launchpad bug 1218953 in gallery-app "Failing autopilot test: gallery_app.tests.test_photos_view.TestPhotosView.test_select_button_cancel(with mouse)" [Critical,New] [09:39] Mirv: we can, of course, merge the fix in manually, but I guess it would be best to fix the issue completely [09:41] hi seb128 [09:42] darkxst, hey [09:42] Mirv, that's not in the sdk failing tests today: http://10.97.0.1:8080/job/autopilot-saucy-daily_release/1506/label=autopilot-intel/testReport/ ? [09:42] Mirv, oh, I guess it's the app stack? [09:43] seb128, can you merge 1219188 so it lands after beta freeze finishes? [09:44] darkxst, I can try to have a look, but I've an endless todolist and those diffs don't seem trivial to review ... you maybe have a better chance trying to grab the day patch pilot [09:44] seb128, ubuntu doesnt use that panel though ;) [09:44] darkxst, that doesn't mean we should upload stuff without review [09:45] nah didnt mean that, just you dont have to worry about regresssions ;) [09:45] I will ping patch pilot [09:45] thanks [09:46] darkxst, https://www.google.com/calendar/embed?src=6k1e5rq45m1bdqq0n1ge3oqaok@group.calendar.google.com [09:46] seb128: apps tests are (for a good reason) ran before sdk is allowed to pass [09:46] Mirv, SHRUG [09:46] Mirv, gallery has like 130 tests failing? [09:47] seb128: no, those are for the whole stack and more related to this latest SDK change that is now being reverted [09:47] ok [09:47] seb128: after the revert we should hopefully be again at this one failing test http://10.97.0.1:8080/job/autopilot-saucy-daily_release/1470/ [09:48] let's see what's the status once that revert is in [09:48] and then gallery-app would hopefully get in [09:48] the fix for that one test === thumper-afk is now known as thumper [10:07] Mirv: we should be safe with the SDK stack and the gallery-app test [10:07] Mirv: as we're not running the failing gallery-app test for the SDK stack, so even if the branch doesn't get merged, it's still ok for SDK to release [10:24] sil2100, Mirv: the ui toolkit revert is in, how do we kick retries? [10:25] sil2100, Mirv: wait [10:26] sil2100, Mirv: they screwed the revert, see -touch [10:29] huh [10:42] seb128: https://bugs.freedesktop.org/show_bug.cgi?id=68210 and https://bugs.freedesktop.org/show_bug.cgi?id=67602 hit us by not debian as we are using mergedlibs and they are not. [10:42] Freedesktop bug 68210 in Writer "Cannot import or read imported tiff images" [Major,New] [10:43] Freedesktop bug 67602 in filters and storage "Unable to open .eps files: "Graphics filter not found"" [Critical,New] [10:44] Sweetshark, hey, ok ... can I help there? [10:44] there is likely more trouble hidden there. A quickfix would be to disable mergedlibs as it is obviously currently broken. The downside of that is it would a/ increase package size b/ slowdown startup (esp. on ARM). [10:45] hmm [10:48] Sweetshark, what is "mergedlibs"? [10:48] seb128: so hmm, I guess we should still do that -- the alternative is to fix that upstream, but I find it hard to judge how much effort it is to fix that properly upstream and if we get that in in time. [10:48] Sweetshark, and why did we enable it and not debian (and when)? [10:50] seb128: mergedlibs puts all the libraries that are used on startup in one big lib instead of a bazillion small ones. Its just one ./configure switch. The advantage is -- as said above: quicker startup because of less dynamic linking work and LTO and a smaller total pkg (also presumably because of LTO). [10:52] OTOH most devs never use mergedlibs as linking half of LO in one lib is severely prolonging your edit-compile-link-test cycles. [10:53] seb128: We had mergedlibs in 4.0 and 3.6 and it worked fine there, but as it looks now I would rather disable that. [10:53] seb128: fwiw SUSE and gentoo also ship with mergedlibs, while redhat and debian do not. [10:53] Sweetshark, did we measure the impact it has on performances when we did it? [10:53] Sweetshark, I guess Suse cares about having it working then? [10:54] seb128: they might just disable it too ... [10:55] seb128: (temporarily, which doesnt help us because we have a hard deadline) [10:58] seb128: when I enabled it I timed it on armhf, but only with the complete distro update: That was LibreOffice 3.5 on precise vs. LibreOffice 3.6 on quantal. IIRC it roughly halfed the first startup from 6sec to 3sec, but that might not have been mergedlibs alone, but pther precise->quantal improvements in addition ... [11:00] seb128: correction -- the change was between LO3.6 on quantal and LO4.0 on raring -- so raring was the first release with mergedlibs [11:02] Sweetshark, seems like an optimization worth having... would it be hard to fix the mergedlibs? [11:05] seb128: so fixing the issues I see right now might be fixable in reasonable time (although its painful to debug and work with such a huge lib) -- just finding out what went wrong was hopefully already half the work. My concern is rather other cornercases that are broken and e.g. not yet reported ... [11:05] Sweetshark, are they broken in raring as well? [11:06] seb128: I guess not. I should have decent visibility of "is broken on Ubuntu, but not on Debian" bugs for Raring -- it has been out long enough. [11:07] ok [11:07] Sweetshark, well, your call, if you think the non mergedlibs codepath is better tested upstream and the way forward let's go for that [11:08] Sweetshark, it could be useful to do an email to ubuntu-desktop or devel list summarizing the situation [11:12] seb128: 'your call', thats mean. Lemme try fixing this in this week, and if it doesnt work out go back to nonmerged as a fallback (that would be changing one character in the rules file after all). In this week I can also watch closely for other suspicous symptoms to crop up. [11:12] does that works out with the release schedule? me checks ... [11:13] Sweetshark, https://wiki.ubuntu.com/SaucySalamander/ReleaseSchedule [11:14] it would mean we have the bug still in beta1 -- but thats what betas are for, right? at least we see if there are other horrorstories from this. [11:16] there is no ubuntu desktop beta 1 :-) [11:17] Laney: you can change the names, but I guess people will stick to the customs for a long time ;) [11:42] Mirv, jibel: I'll cancel the SDK tests, they're hung again [11:43] jibel: how can I safely abort this? Can I simply kill the jenkins job? [11:43] jibel: will that do the trick? [11:43] sil2100: I just canceled them too [11:44] sil2100: sorry, I was just about to say "canceled the hung autopilot jobs for you", but then you were here :) [11:44] Mirv: let's rebuild SDK, since the revert landed [11:44] ;) [11:44] sil2100: the red 'x' at http://10.97.0.1:8080/job/autopilot-saucy-daily_release/label=autopilot-intel/ + http://10.97.0.1:8080/job/autopilot-saucy-daily_release/label=qa-nvidia-gtx660/ [11:44] sil2100: yes, fire it up [11:45] Mirv: I wanted to make sure that doesn't break anything, since yesterday I was doing aborts and the AP machines were dying [11:45] Not sure if related, so I asked [11:45] sil2100: ok. that's the way I've done those without AP machines dying. [11:47] sil2100, you can safely cancel the job but sometimes it doesn't completely stop the container and next jost will fail becuase the container is already up. [11:48] jibel: yep, that's what I've noticed [11:49] That's because jenkins sends a SIGTERM to the process tree as user jenkins but process is running as root [11:50] I wrote a script to kill any process created from the jenkins script. I should probably deploy it there === dednick is now known as dednick|lunch [12:19] seb128: I guess the settings stack changes are covered by the touch FFe, right? [12:19] So I can safely publish that probably? [12:21] sil2100, yes [12:37] jibel: eh, apps are stuck now... [12:38] jibel, Mirv: I aborted the apps check, it hung on webbrowser again [12:38] :-( [12:38] sil2100, when did those hangs start? [12:39] sil2100, same pb than previous run [12:39] sil2100: :( [12:39] seb128: they didn't happen last week ;/ All started like yesterday [12:39] sil2100, Mirv: hum, sdk still red... is that the same issue? [12:39] jibel: indeed ;/ [12:39] jibel, what's the issue? do we have a clue what is to blame? [12:40] seb128: eh, it's red because the containers are busy [12:40] jibel: ^ [12:40] busy? [12:40] shouldn't that make them wait? [12:40] jibel: SDK check can't run because of the containers - the thing you said earlier [12:41] seb128: it's a problem that sometimes happens when we abort the AP test machines, the containers don't stop completely and are unavailable [12:41] jibel has some fix for that but it wasn't deployed yet [12:41] hum, k [12:41] jibel: ^ [12:41] this week is quite a fail one so far [12:41] seb128: it seems all has really REALLY bad luck now [12:41] seb128, it started somewhere between Sep 2, 2013 1:01:30 PM and Sep 2, 2013 10:14:51 PM [12:42] jibel, do you think it's likely a distro change? we are in beta1 freeze, we didn't get that many stuff landing... [12:44] jibel: can you unblock the containers? What should I do in such a case when you're not around? Should I simply stop the containers or something more? [12:45] seb128, yeah, I've no idea what is causing this hang yet. There is no evidence on the test machine, autopilot is running, and found nothing special in the logs [12:45] jibel, :-( [12:45] seb128, strace shows that AP is looping on a select, which actually means nothing [12:45] jibel: maybe we should have the autopilot development guys check it out? [12:46] sil2100, yup [12:48] sil2100, the containers are unblocked, you can also run: sudo lxc-stop -n $(lxc-ls|tail -1) directly on the machines, that will stop the run and create all the artifacts [12:48] jibel: thanks! Ok, so just stopping is enough, yay [12:48] Re running SDK [12:56] seb128: you got mail [12:57] seb128, sil2100 http://paste.ubuntu.com/6058785/ this is the list of packages that differ between the last successful run of apps and the first run that completely failed [12:57] Sweetshark, k [12:58] jibel, so could be an ui toolkit issue... [12:58] sil2100: https://code.launchpad.net/~schwann/gallery-app/gallery-atest-toolbar-opened/+merge/183195 got approve from jenkins! [12:59] top-approving it so it can land [13:00] sil2100, Mirv: sdk still red :/ [13:01] seb128: it's actually success, but with a warning about removed tests [13:01] Mirv, yeah, I was wondering since intel/nvidia are green [13:01] seb128: so sil2100 can probably just publish it after double checking that it's correct that some tests were now disabled [13:01] \o/ [13:06] ;/ [13:06] \; [13:06] Checking [13:13] Ok, I can't find anything strange [13:14] It's all a bit suspicious, but I think it's not suspicious enough not to release those stacks === psivaa is now known as psivaa-lunch [13:16] sil2100, do it! ;-) [13:16] sil2100, if there are issues we can blame it on jibel and the unstable QA config ;-) [13:16] * seb128 hides from jibel [13:17] ;D [13:17] Ok, working on getting everything out [13:17] \o/ [13:18] seb128, you'd rather blame the app team that removes tests to make them pass ;) [13:18] speaking of removing test [13:19] jibel, chrisccoulson: the firefox tests seems like they never managed to go green, we keep overruling them, should we just drop them? [13:19] seb128, IIRC gnome-session failed to start so tests do not even run, but someone more knowledgeable than me must have a look [13:19] Published, yay [13:20] sil2100, \o/ [13:21] seb128, there has been a lot of work to have them green in raring, it'd be a pity to remove them [13:21] jibel, well, reality is that they seem to be always red and creating issues, so they are creating work rather than being useful ... though I agree than having them working would be nice [13:21] chrisccoulson, ^ [13:22] seb128, I can setup a testing environment and give access to any resource we have to help fixing them [13:22] seb128, unfortunately, the tests leave no clue in the logs of why gnome-session doesn't start, and i can't reproduce it here [13:23] stgraber managed to reproduce it while we were at debconf [13:23] I think we were using the prepare-testbed stuff [13:24] yeah, was easily reproducable when using the kvm based autopkgtest stuff as described on developer.u.c [13:24] was also reproducable on a standard saucy machine as long as it's not using an nvidia driver === dednick|lunch is now known as dednick [13:40] fginther: hello! === psivaa-lunch is now known as psivaa [13:41] fginther: about the removal of those hooks... I have also been wondering about removing the phablet ppa and the sdk ppa - but I wanted to consult it with you first [13:42] fginther: since I guess even if this might provide some override behavior, I guess we don't want to be testing/building things when using non-daily-build/non-distro versions [13:42] fginther: so, you think we can remove those two as well? [13:46] sil2100, are you just referring to removal of the hooks, or the ppas as well? [13:48] fginther: for now maybe just the hooks, since PPAs can still be used for local testing [13:49] But I guess for mediumtests and daily-release, those PPAs might only cause trouble [13:49] As we might be testing and building against something that's not going to be released [13:49] sil2100, yes, I think that's reasonable [13:50] fginther: ok, I'll prepare a branch for your review in a moment [13:50] sil2100, cool, thanks [13:50] fginther: btw.... ;) Are you busy busy? [13:50] sil2100, I'm less busy this week [13:50] sil2100, do you have a major project for me :-) [13:51] fginther: since we have that branch in gallery-app that just fails building, it's failing on mediumtests - the author says he cannot reproduce it on the machines (phones) and locally [13:51] fginther: sadly, no major project ;p Just the standard hacking [13:51] fginther: https://code.launchpad.net/~schwann/gallery-app/gallery-atest-toolbar-opened/+merge/183195 [13:51] fginther: at first it was a dep problem, we fixed it, but then it's hm, strange [13:52] fginther: we don't know how the mediumtests are working and where to get specific info, but for instance CI approved it and then it failed to merge [13:52] I re-approved, but I guess it won't help much [13:52] hmm [13:52] There's not much change in the code, so this is a good question mark [13:53] om26er, can you help take a look at this gallery-app test? ^^ [13:53] fginther: let's move to -touch, and poke gusch and om26er there [13:53] ack [13:53] fginther, sil2100 sure [13:53] Thanks guys [13:54] fginther: hey :) [13:54] asac, morning [13:54] hi ... -> touch :) === m_conley_away is now known as m_conley [14:27] sil2100: did cordova get sorted out after all? [14:33] cyphermox: the issue that Robert pointed out in the e-mail seems fixed, but it's not out yet because of webapps problems [14:36] cyphermox: let me find the bug [14:36] ack [14:42] cyphermox: actually, the bug was not really accurate, so I made a new one, since I guess the previous issue got fixed: https://bugs.launchpad.net/ubuntu/+source/webapps-applications/+bug/1220277 [14:42] Launchpad bug 1220277 in webapps-applications (Ubuntu) "unity-webapps-amazon overwriting file from unity-webapps-common" [Undecided,New] [14:55] grr [14:56] come on SetTime, you know you want to work [14:59] Laney: no it doesn't really it wants to drive you around the bend! [15:04] fginther: what do you think? https://code.launchpad.net/~sil2100/cupstream2distro-config/hook_cleanups/+merge/183686 [15:10] czajkowski: Might have been me that was defective. :P [15:10] * Laney tests [15:11] Yes, indeed it was. ¬_¬ [15:12] One part gave milliseconds while the other wanted microseconds [15:12] Turns out they don't like it very much when you mix those up [15:19] sil2100, I'll review in a moment [15:29] hey everyone, it's meeting time [15:30] qengho, mlankhorst, Sweetshark, Laney, tkamppeter, larsu: hey, it's meeting time [15:31] seb128: as usual, I'm still in the indicator call :) I have my notes ready though! [15:31] I hope everyone is fine! [15:31] larsu, no problem, there is time before your turn [15:31] Yo. [15:31] qengho, hey [15:31] * Trying to tighten the loop of chromium releases. [15:31] - I'm taking over the webapps patches. Previous maintainer is too busy. [15:31] - I am trying to automate the webapps testing. This is still a mystery to me. [15:31] - On new release, I'm finishing any release I start, instead of restarting if a new upstream release appears. Today's or tomorow's release will be 29.0.1547.62 (28 Sept) instead of 29.0.1547.65 (yesterday). [15:31] * Trying to make chromium the best choice for security-conscious users. [15:31]