[02:05] === IMAGE 121 building (started: 20150304-02:05) === === chihchun is now known as chihchun_afk [03:05] === IMAGE RTM 250 building (started: 20150304-03:05) === [04:15] === IMAGE RTM 250 DONE (finished: 20150304-04:15) === [04:15] === changelog: http://people.canonical.com/~ogra/touch-image-stats/rtm/250.changes === === chihchun_afk is now known as chihchun === marcusto_ is now known as marcustomlinson [08:59] imgbot, status 121 vivid [08:59] Error: There does not seem to be any build with the number 121 [08:59] hmm [08:59] bah [08:59] W: Failed to fetch http://ftpmaster.internal/ubuntu/dists/vivid/universe/binary-armhf/Packages Hash Sum mismatch [08:59] E: Some index files failed to download. They have been ignored, or old ones used instead. [09:00] * ogra_ restarts the build === greyback_ is now known as greyback [10:00] sil2100: bug #1427710 [10:00] bug 1427710 in maliit-framework (Ubuntu) "A maliit-framework segfault with Qt 5.4.1" [Undecided,New] https://launchpad.net/bugs/1427710 [10:01] Mirv: thanks :) [10:06] bzoltan_: hey, how's the new UITK landing going? [10:06] bzoltan_: I hope it will fix all the UITK autopilot tests on the dashboard, right? ;) [10:07] sil2100: Thanks for asking :) Good progress, the FFe was approved and we passed the first round of tests, kalikiana found one flaky test what he addressed with a single line fix -> https://code.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/objectNameForTextInputHeader/+merge/251705 [10:08] sil2100: I am not sure about the dashboard ...I am validating against the tests made against stock image [10:09] bzoltan_: one-liner fixes are the best [10:09] sil2100: I have heard from iahmad_ about the dashboard ... i am waiting for the information that what exact image introduced the issue [10:10] sil2100: if the fix of those failures are trivial then of course I will push the fix [10:10] bzoltan_: yeah, we're seeing around 20 test failures in UITK right now, would be good to get those down to 0 as it was in the past :) [10:11] sil2100: I see 6 UITK failures on krillin with the #131 [10:11] When running locally? [10:11] sil2100: on krillin [10:12] sil2100: proof http://paste.ubuntu.com/10525293/ [10:16] bzoltan_: I trust your testing ;) It's just that the differences between smoketesting and local testing always bother me, especially that in theory there should be no difference in how tests are run [10:16] grr [10:18] sil2100: you know my opinion about the smoketesting :) It is an unreliable photoshopped test result ... when it comes to UITK landing I trust only my test plan. Since I am using that process I did not cause regression :) [10:19] Usually it was all fine, as recently it was rather reliable, but once again we get seemingly different results [10:19] And this makes me frustrated and hungry === iahmad_ is now known as iahmad [10:20] sil2100: there are two major issues I keep ranting about: 1. some apps do need manual interaction.. like the clock. No interaction means all AP tests fail 2. The krillin with Vivid does enter to reboot loop after certain AP tests. Most typically after the dialer and reminder [10:21] sil2100: since yesterday 6pm I had to falsh with the recovery image 4 times and reflash+setup the whole testing enviroronment each time because of the reboot loops [10:22] This is even more worrying, AP tests should never lead to breaking the device [10:22] I think we might need to poke the respective upstreams about that [10:22] Otherwise we won't be able to setup automation [10:23] sil2100: iahmad: But I am not sure what the probelm is .. the #127-#129 Vivid images on krillin shos the normal 3 AP failures for the UITK [10:23] sil2100: iahmad: http://rtm-dashboard.ci.ubuntu.com/smokeng/vivid/ [10:25] === IMAGE 121 DONE (finished: 20150304-10:25) === [10:25] === changelog: http://people.canonical.com/~ogra/touch-image-stats/121.changes === [10:25] sil2100: I do not know what exactly causes the lockups ... my test plan script does nothing strange ... just reboots, configures and runs the tests [10:28] bzoltan_: in the end what we should really have is like one way and one tool of running autopilot tests suites for both local and automated testing, so that we're 100% sure that in both cases exactly the same steps are performed [10:28] * sil2100 is still waiting for the documentation of that from the CI team [10:29] sil2100: +1 My proposal that the CI tools should be productized, packaged, documented and released for common use was hammered down several times. [10:30] sil2100: I think it is the lack of time and resources [10:43] ogra_: what's up with vivid image taking... 8h 20min to build? [10:43] Mirv, you didnt listen in the meeting, eh ? [10:43] :) [10:44] ogra_: sorry, I missed some part of it indeed when doing other stuff :( [10:44] last nights build failed due to a hashsum mistmatch error on the internal archive server ... i re-kicked it again right before the meeting [10:44] busted [10:44] ok, I saw it some slowness the other day too === MacSlow is now known as MacSlow|lunch [12:47] hm, my ubuntu-device-flash seems to have issues flashing my krillin [13:01] * Mirv has pushed three packages to Debian today so far [13:01] since I need them to my PPA, and I need them as unmodified from Debian :) === alan_g is now known as alan_g|lunch [13:06] \o/ [13:38] cihelp hey, any idea why one of our MPs wouldn't be picked up by the -ci job? https://code.launchpad.net/~unity-team/unity8/shellRotation/+merge/248400 [13:38] * ev looks [13:38] I know I can run it manually, thought it'd be useful to see why it's not picked up automagically [13:41] Saviq: how long were you waiting? [13:41] ev, it's almost a day now [13:41] ev, the MP was in WiP for a while [13:41] ev, and now we put it in Needs Review, maybe that's what got the MP scrubber confused? [13:41] (now == yesterday) [13:41] but it's been over a day that it was in Needs Review? [13:42] right [13:50] Saviq: is this Kevin's first branch under CI? [13:52] I suspect that's the problem, unless this has worked for him before [13:52] I've added his LP ID to the whitelist [13:53] so future MPs should work, but let me / the vanguard know if that's not the case === MacSlow|lunch is now known as MacSlow [13:54] ev, huh, we might have just dealt with it manually before, nice find, didn't know there's a LP user whitelist :) === alan_g|lunch is now known as alan_g [13:57] credit goes to fginther for doing a bang up job documenting all of this: https://wiki.canonical.com/UbuntuEngineering/CI/Playbook/UpstreamMerger#Specific_merge_proposals_are_not_being_triggered [13:58] :) [14:00] bzoltan_: hey! [14:00] sil2100: shoot [14:01] bzoltan_: the UE Live is starting in a few minutes, do you think we can have someone from the SDK team doing some updates? ;) [14:10] sil2100: I have an other meeting starting ... [14:10] sil2100: let's see if somebody from my team can make it [14:13] s/three/six/ soon. [14:57] hm, it seems it wasn't able to publish qtubuntu-sensors [14:58] hm [14:59] It didn't even upload? [14:59] Let me check the publish rsync file [15:00] rsalveti: interesting, it looks like the rsync packagelist file is ok [15:00] rsalveti: so it should have published it, as it had this: [15:01] yeah, trying to publish it again after a watch-only build [15:01] ci-train-ppa-service/ubuntu-rtm/landing-003Release14.09Proposed14.09qtubuntu-sensors0.6+15.04.20150226~rtm-0ubuntu1timo-jyrinkiubuntu-rtm [15:01] indeed [15:01] slangasek: ping [15:01] slangasek: could you check on snakefruit if there was a problem with publishing qtubuntu-sensors? [15:06] sil2100, i'll trigger another image soon ... to test pittis big seed changes he did today [15:06] i.e. t make sure we can still build images and the packages still show up as they should [15:06] ogra_: sure thing [15:07] Yeah, +1 on that [15:07] What were those seed changes about? [15:07] systemd as default [15:11] \o/ [15:11] To clarify: systemd as default but not for touch [15:11] err, yes [15:11] rsalveti: ok, now it seemed to work [15:11] i should have been clearer :) [15:12] sil2100: pretty sure that http://people.canonical.com/~ubuntu-archive/cicopy.log is all the logging we have, even for those with direct access to snakefruit [15:12] slangasek: ^- [15:12] oh, well, there's LP logging [15:12] rsalveti: but no idea why it didn't work before, if a package is correctly written down in the rsync file then it should be copied [15:12] cjwatson: oh, ok, bookmarking that (always forget about it) [15:12] Thanks :) [15:13] SOmetimes it's good to have that in case problems appear [15:13] But there's no sign in LP logs that the copy was attempted earlier [15:13] First entry is: [15:13] [2015-03-04 15:05:06,589: INFO/PoolWorker-3] Running (ID 26645187) in status Waiting [15:13] And then that succeeded [15:14] Ideally we'd have, you know, actual logs rather than just the last one [15:14] cjwatson: so before only unity-scopes-shell and unity8 were logged? [15:14] Yeah... [15:14] cjwatson@carob:~$ grep ubuntu-rtm/landing-003 celeryd-production_launchpad_job.log [15:14] [2015-03-04 14:15:06,813: INFO/PoolWorker-2] Running (ID 26644652) in status Waiting [15:15] [2015-03-04 14:15:12,145: INFO/PoolWorker-1] Running (ID 26644656) in status Waiting [15:15] [2015-03-04 15:05:06,589: INFO/PoolWorker-3] Running (ID 26645187) in status Waiting [15:16] huh, strangeness, need to check maybe there's some small bug in the copy2distro script or something [15:18] sil2100: Or it's possible it failed with a network blip or something, complained on stdout, but then decided it had processed the rest of the job ... [15:18] s/stdout/stderr/ [15:18] Proper logging at some point :-/ [15:27] Mirv: thanks for the update [15:27] cjwatson: +1 [15:27] cjwatson: I can try adding some logging from the copy2distro side anyway [15:28] sil2100: np [15:30] === IMAGE 122 building (started: 20150304-15:30) === [15:56] sil2100: could you tell me what do I do wrong here https://launchpadlibrarian.net/199340335/buildlog_ubuntu-vivid-amd64.ubuntu-ui-toolkit-gles_1.2.1431%2B15.04.20150304.1-0ubuntu1_BUILDING.txt.gz [15:56] Looking [15:58] bzoltan_: tst_performance exited with 666 <- jokingly saying it seems teh devil possessed your package! [15:58] Digging deeper === chihchun is now known as chihchun_afk [16:03] * bzoltan_ goes to fetch a pint of holly water [16:05] bzoltan_: it seems that the warnings you get during unit tests seem to be causing the FTBFS [16:07] bzoltan_: didn't you already build this UITK successfully in the PPA? [16:13] sil2100: The main package is built and I am running the test plan like second time on it. That is what makes it wierd ... [16:19] bzoltan_: weird, looks like some missing dependency [16:20] bzoltan_: maybe something changed with the new Qt 5.4? And maybe you need to install something specific for QtQuick.Layouts to be available? [16:21] Maybe it's now missing for the -gles versions build-deps in its debian/control [16:21] Although I'm sure we'd notice it during Mirv's initial rebuilds then [16:22] bzoltan_: yeah, it seems you're missing the qml-module-qtquick-layouts dependency [16:22] bzoltan_: you'll have to add that to the -gles packaging [16:23] bzoltan_: since I see it's being installed by the normal UITK during build, while it's not in the -gles versions - meaning the dep is missing, and it's what's providing that lib [16:36] popey, robru, davmor2, ogra_, rvr: are there any updates from you guys re. landing things? Since if not we might want to skip today's evening meeting [16:37] sil2100: not from me [16:37] nothing from me except the seed changes [16:37] sil2100: No new silos today [16:38] sil2100: ohh, that explains [16:42] popey, robru, davmor2, ogra_, rvr: ok, let's skip it then :) [16:42] great [16:44] * ogra_ twiddles thumbs waiting for the importer [16:51] trainguards: Hi! I need to change the branch that the train watches for landings from lp:unity-system-compositor to lp:unity-system-compositor/ubuntu. Is this the right place to ask? [16:52] alf_: hey! What do you mean by 'the branch that the train watches'? The train watches the branch that is the MR destination which you want to release [16:52] alf_: which means basically CI Train doesn't watch just one branch, you can already release a package from branch lp:foo/a, then from lp:foo/b etc. [16:53] alf_: all that CI Train needs is to have previous branch history in the branch so that he can find the last released package's version tag in it === dpm_ is now known as dpm [17:00] sil2100: location on rtm passed is the only news from me another battery drain hopefully fixed :) [17:00] === IMAGE 122 DONE (finished: 20150304-17:00) === [17:00] === changelog: http://people.canonical.com/~ogra/touch-image-stats/122.changes === [17:00] * ogra_ looks at the build log [17:00] imgbot, status 122 vivid [17:00] sil2100: Thanks, I was misunderstanding how the train works, sorry for the noise. [17:00] Status: succeeded, Started: 2015-03-04 15:27:09 UTC, Finished: 2015-03-04 16:25:15 UTC [17:00] Build URL: https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/vivid/ubuntu-touch/+build/21786 [17:00] Changelog: http://people.canonical.com/~ogra/touch-image-stats/122.changes [17:01] alf_: no worries :) What you might want to do though is poke CI if you want the merge-CI to work [17:01] alf_: since I think CI setups the merger-bot per-branch [17:01] imagbot, stunt [17:01] imgbot, stunt [17:01] * imgbot rolls on its back and purrs [17:42] sil2100: regarding failed qtubuntu-sensors publication, nothing is wrong with copy2distro, the packagelist is just invalid: https://ci-train.ubuntu.com/view/2.%20Publish/job/ubuntu-rtm-landing-003-2-publish/67/artifact/packagelist_rsync_landing-003-14.09/*view*/ distro version is missing, which means that's not getting set properly by the build job. I'll [17:42] prep a branch that fails the publish job in this case rather than silently succeeding with this broken packagelist. [17:55] trainguards: you can delete silo vivid 007; we created a new one with alex-abreu to hold more fixes; sorry for the duplication [17:55] dbarth_: no worries, thanks [17:55] robru, thx [17:56] alex-abreu: you're welcome [17:59] sil2100: https://code.launchpad.net/~robru/cupstream2distro/dont-publish-invalid-packagelist/+merge/251789 if you're still around === alan_g is now known as alan_g|EOD [18:06] robru: looking :) [18:07] sil2100: robru: How can I convince the silo to take the trunk of the MR and use the just modified debian/control file for building the -gles package? [18:08] https://ci-train.ubuntu.com/job/ubuntu-landing-003-1-build/156/console [18:08] it complains about a missing dep what I just added [18:09] bzoltan_: I'm not sure what you mean. if there's packagng changes in the non-gles you need to sync those into the -gles branch. [18:10] robru: I did that [18:10] sil2100: tthanks [18:11] bzoltan_: well that job you linked pretty clearly shows that it's merging in your branch which has modifications to debian/control, so the only thing I can guess is that your debian/control is not correct. [18:11] robru: the consol logs says that it pulls the qml-module-qtquick-layouts but he build does not pull it https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-003/+build/7031245/+files/buildlog_ubuntu-vivid-amd64.ubuntu-ui-toolkit-gles_1.2.1431%2B15.04.20150304.1-0ubuntu1_BUILDING.txt.gz [18:11] robru: approved, thanks :) [18:12] sil2100: thanks! [18:12] robru: the failure clearly says that -> tst_performance: QWARN : tst_Performance::benchmark_GridOfComponents(list with Captions, preset: caption) file:///build/buildd/ubuntu-ui-toolkit-gles-1.2.1431+15.04.20150304.1/tests/unit/tst_performance/ListOfCaptions.qml:19:1: module "QtQuick.Layouts" is not installed [18:13] robru: and that QtQuick.Layouts is provided by the qml-module-qtquick-layouts ... what is indeed listed in the build-deps [18:14] robru: and the consol logs shows that it is pulled, but the build is missing it [18:14] bzoltan_: let me take a quick look [18:15] bzoltan_: https://ci-train.ubuntu.com/job/cyphermox-test/810/console this is the debian/control file as it was built & uploaded to launchpad. [18:15] bzoltan_: if that's correct, then it's a bug in launchpad. [18:16] robru: it has the qml-module-qtquick-layouts, so it should be a correct one [18:17] bzoltan_: so somehow launchpad ppa got something different than what we uploaded. I've never heard of that before ;-) [18:17] bzoltan_, robru: so... the problem looks like the version [18:17] robru: I have hard it once from Mirv [18:17] bzoltan_, robru: because of the hacks that we use for the -gles packages, you rebuild the package with the same version - and this is not possible [18:18] bzoltan_, robru: you cannot re-upload a package with the same version number to the PPA [18:18] sil2100: how to fix that? the -gles should follow the main package. Should I bump the main version and rebuild it? [18:18] bzoltan_: I suppose so. [18:18] bzoltan_: change the ubuntu version :) [18:18] bzoltan_: 1.2.1431+15.04.20150304.1-0ubuntu1 to 1.2.1431+15.04.20150304.1-0ubuntu2 [18:18] (in -gles) [18:19] It should still work as we only take the tarball, which only has the upstream version (1.2.1431+15.04.20150304.1) [18:20] sil2100: yeah makes sense, but what happens if non-gles is 0ubuntu1 and gles is 0ubuntu2? isn't there some code somewhere that freaks out if those don't match exactly? [18:20] robru: yeah, but we can use the ignore flag for that (if it's still there) [18:21] This is the easiest way in this case [18:21] Otherwise we'd had to rebuild both packages, wasting time and resources [18:21] sil2100: which ignore flag are you thinking? IGNORE_MISSING_TWINS? [18:21] s/had/have [18:21] robru: yeah [18:21] Although... [18:21] It might not be needed, I didn't do any version checks [18:21] And not sure we have any other checks anywhere [18:22] (even beyond the train) [18:22] sil2100: ignore missing twins is only if the whole package is missing, doesn't change any version handling. but also I don't think the train forces those versions, I mean like somewhere in the packaging, if a dependency says like ==binary:version [18:23] robru: sil2100: i give it a try with bumping to ubuntu2 ... let's see, if it does not work then i will rebuild the main [18:23] robru: there shouldn't be any place like that, as the two aren't bound together in any way - we only request publishing o both since otherwise people would forget about them and leave the emulator outdated [18:23] s/o/of [18:24] sil2100: hm ok, yeah i don't see == anywhere in the packaging anyway. [18:34] bzoltan_: oh, what was wrong with 4.1-0ubuntu2? [18:39] robru: it did not find the orig.tar.gz ... i take the sure and brutal way... rebuild the main and it will get new version what the -gles will follow [18:40] bzoltan_: hm, that's weird. ok... [18:45] bzoltan_: huh? Why didn't it? [18:46] * sil2100 sighs [18:46] Absurd [18:47] huh? [18:48] bzoltan_: but I don't see you changing the version [18:48] bzoltan_: did you bump it to ubuntu2 ? [18:49] There's no sign in Launchpad logs of an upload of ubuntu2 [18:50] Oh, right, not an upload, it was a citrain build job failure [18:51] Maybe it's failing to handle the + or something? [18:51] Yeah, but it should have found the .orig file, as it's in the pool [18:51] http://ppa.launchpad.net/ci-train-ppa-service/landing-003/ubuntu/pool/main/u/ubuntu-ui-toolkit/ [18:51] Not sure [18:51] No use to debug it now though as bzoltan_ rebuilt the original packages [18:52] sil2100: the MP is weird. he changes it to 4.2-0ubuntu1 then 4.1-0ubuntu2 then back to 4.2-0ubuntu1. I have no idea what's going on [18:53] I think bzoltan_ is doing some confusing magic there ;) [18:53] sil2100: oh I guess he uploaded the 4.2 by mistake and then it was too late to go back [18:53] He's far past his EOD so I guess that's why he didn't want to play more with it [18:53] Yeah, maybe === blr_ is now known as blr [19:42] robru: want to land a change in media-hub that is not functional, just changing the video formats used when running the unit tests in order to remove the dependencies of a few complicated packages [19:42] as we want to move media-hub to main [19:42] robru: do we have someone from QA to help me understanding if this would require sign-off? [19:43] rsalveti: ToyKeeper is my go-to qa person, but i think it's a bit early for her right now. otherwise jibel [19:43] yeah, jibel doesn't seems to be around [19:43] davmor2: still around? :-) [19:44] rsalveti: I am [19:44] davmor2: this is the MR I want to land in vivid: https://code.launchpad.net/~rsalveti/media-hub/migrating-tests-open-formats/+merge/251669 [19:44] davmor2: just changing the video/audio file format to ogg, so we don't need to depend on fluendo/libav [19:44] davmor2: do we need QA to sign that off? [19:45] no functional changes [19:47] rsalveti: if it is landing through a silo then yes it would need signing off. But I guess it would only take a few minutes to test, record a video is it format x rather than y can it be played back on phone and pc right? My only concern would be that MS and Apple pc's wouldn't be able to play it any more right? [19:48] rsalveti: what would you use to record if it isn't fluendo/libav is it ogg directly or something [19:48] davmor2: doesn't change anything on the functional level, as we're bringing the list of additional plugins as part of our seeds [19:49] davmor2: this only changes the unit tests [19:49] those videos/formats were not used outside the unit tests [19:49] basically the problem is that we were testing with mp3 and avi files, and for that to work we need to build-dep on libav and fluendo-mp3 [19:50] but nothing should change for the normal ubuntu touch user :-) [19:50] will make it to require qa then and add some comments that all we need really is smoke testing it [19:51] as nothing changed on the functional level [19:51] rsalveti: so I see no issues landing that we'll soon know if it breaks anything [19:51] right [20:05] === IMAGE RTM 251 building (started: 20150304-20:05) === [20:05] that's me, after our landings we got earlier today [20:05] regarding GPS and gstreamer [20:55] brb, lunch [20:56] trainguards can you remove L46 from the stylesheet, the silo have been cleared already [20:56] and reconfigure silo 13 [20:57] (vivid silo 13) [21:20] === IMAGE RTM 251 DONE (finished: 20150304-21:20) === [21:20] === changelog: http://people.canonical.com/~ogra/touch-image-stats/rtm/251.changes === [21:20] alex-abreu: one sec [21:22] alex-abreu: good to go [21:31] robru, thx [21:31] alex-abreu: you're welcome [23:31] bregma: are we not qa'ing desktop stuff? guess not [23:32] bregma: or is that on the phone? [23:59] robru, Thanks!