2014/05/26/#ubuntu-ci-eng.txt

[02:09] <imgbot> === trainguard: IMAGE 49 building (started: 20140526 02:10) ===
[03:24] <imgbot> === trainguard: IMAGE 49 DONE (finished: 20140526 03:25) ===
[03:24] <imgbot> === changelog: http://people.canonical.com/~ogra/touch-image-stats/49.changes ===
[05:14] <bzoltan> hello, is here anybody who could assign a silo for the line 31?
=== Ursinha-afk is now known as Ursinha
=== Ursinha is now known as Ursinha-afk
[05:41] <bzoltan> Mirv: may i ask for a silo? Line number 31.
[05:47] <bzoltan> zsombi:  do you remember if zbenjamin is off today or not?
[05:48] <zsombi> bzoltan: no, don't remember to have heard that
[05:48] <bzoltan> zsombi: it is too early still :)
[05:49] <zsombi> bzoltan: yeah :)
[05:49] <bzoltan> zsombi:  I am officially off today, but I do some work to keep the show rocking :) a landing would be nice today. I hope Mirv will give us a silo :)
[05:52] <bzoltan> zsombi: after the Silo is ready, I will start an all night AP set ... I think we still have 6-8 hours to push stuff to staging
[05:53] <zsombi> bzoltan: we have, and we have loads of work pending... each MR pending on happroval will take 1.5-2h to autoland... so I think we won't have too many more MRs landing to staging...
[05:54] <bzoltan> zsombi: OK, no problems.
[05:58] <Mirv> bzoltan: yep you have a silo
[05:58] <bzoltan> Mirv: thans
[05:59] <Mirv> bzoltan: actually... blocked by landing-006 hmm
[05:59] <Mirv> well, suru theme can wait
[06:00] <bzoltan> Mirv: OK, I will go away from the desk for few hours, I will start the build before that
[06:00] <Mirv> bzoltan: ok now, landing-016
=== Ursinha-afk is now known as Ursinha
[07:42] <sil2100> ogra_: hey, you in front of your PC?
[07:42] <sil2100> ogra_: I would like a packaging ACK, a quick one: https://ci-train.ubuntu.com/job/landing-011-2-publish/42/artifact/packaging_changes_indicator-network_0.5.1+14.10.20140523-0ubuntu1.diff <- this just adds an installation directory for locale
[07:59] <thomi> cihelp - what room are you guys at?
[07:59] <doanac> thomi: in the ballroom
[07:59] <Ursinha> thomi: the ballroon
[07:59] <Ursinha> room
[08:00] <josepht> thomi: we're in the ballroom :P
[08:04] <thomi> fginther: just wanted to make sure you got my message about a stuck MP?
[08:07] <fginther> thomi, looking
[08:07] <thomi> fginther: thanks
[08:08] <mhr3> sil2100, ping?
[08:08] <sil2100> mhr3: pong!
[08:09] <mhr3> sil2100, there's a problem with migration of click scope, it added dep on ubuntu-sdk-libs, which doesn't do arm64, so it doesn't build on arm64 anymore
[08:09] <mhr3> sil2100, can you do the magic to unblock it?
=== ted is now known as tedg
[08:11] <sil2100> mhr3: ok, so I can't do much regarding click scopes since I have no powers there, I'm still in the age of deb's
[08:12] <kgunn> silo 15 seems weirdly stuck....if someone would help me to kill it?
[08:12] <sil2100> mhr3: maybe Mirv or sergiusens could help
[08:12] <sil2100> kgunn: looking
[08:12] <sil2100> oh
[08:12] <mhr3> sil2100, i meant unity-scope-click
[08:12] <mhr3> sil2100, http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html
[08:12] <sil2100> mhr3: ah, crap, sorry, mis-read
[08:13] <Mirv> mhr3: hmm, that's sad. we can't directly do anything about it. you shouldn't depend on it until ubuntu-sdk-libs is available on arm64, in short.
[08:13] <Mirv> if it was a shortcut for making sure some other deps are installed, then those deps should just be depended to directly
[08:14] <mhr3> Mirv, we had this issue long time ago with something else, the solution was to remove the arm64 binary from the archive
[08:15] <Mirv> mhr3: well it's usually not a solution archive admins will want to make lightly
[08:15] <sil2100> mhr3: let's ask cjwatson if he can help us with that
[08:15] <sil2100> kgunn: yeah, so... it seems it's stuck as mir stopped being buildable on arm64
[08:16] <sil2100> kgunn: due to the lack of libandroid-properties-dev - you would either have to convince the archive admins to remove mir (and deps) from arm64 or do something about the dependency
[08:18] <thomi> cihelp - can someone please allocate me a silo for landing SS row 33?
[08:18] <kgunn> alan_g: ^ i had put mir & unity-mir trusted session branches in the silo to build...but seems we've got  a dep issue on arm64
[08:19] <sil2100> thomi: let me do that
[08:19] <thomi> sil2100: thanks
[08:19] <sil2100> thomi: we usually get pings when 'trainguards' is used ;)
[08:19] <thomi> sil2100: good to know, thanks
[08:20] <kgunn> alan_g: should we fix or ask admins to drop arm64 ? (i'm doubting they'd say "yes" ?)
[08:21] <Mirv> thomi: I'd have added but you have not selected it as being ready
[08:22] <Mirv> thank you :)
[08:22] <thomi> Mirv: ugh, sorry - have now
[08:22] <alan_g> kgunn: I don't follow the problem
[08:23] <kgunn> sil2100: ^ can you elaborate for alan_g?
[08:26] <sil2100> alan_g: so, it seems the latest changes in mir added a dependency on libandroid-properties-dev from what I see which is not available for arm64
[08:26] <alan_g> kgunn: is libandroid-properties-dev something new? What is it for?
[08:26]  * alan_g realises that's probably why he was asked.
[08:26] <sil2100> alan_g: so, we had mir buildable on that platform before, so it's a problem in overall it seems
[08:28] <alan_g> sil2100: I have to track down what libandroid-properties-dev is about. I've not encountered it.
[08:31] <alan_g> sil2100: I'm grabbing kdub to join us. He introduced it...
[08:32] <Mirv> hmm, firefox crashing
[08:33] <kdub_> sil2100, so libandroid-properties-dev is to scan some properties from the android environment... is the problem that we can't build arm64?
[08:33] <ogra_> sil2100, hmm, do we do a meeting in person ?
[08:34] <Mirv> cool I see the very familiar meeting room
[08:35] <Mirv> at least turn the camera even if you're having the meetingin person! :)
=== vrruiz_ is now known as rvr
[08:46] <rsalveti> alan_g_: kdub_: why are you using android properties on mir?
[08:47] <kdub_> rsalveti, we found a place where surfaceflinger set the number of framebuffers via a #define
[08:47] <rsalveti> I could probably split libandroid-properties from libhybris and have that as arch any then, as the code doesn't really depend on android, but would be interesting to know why it's now used by mir
[08:48] <rsalveti> kdub_: right, and you're setting that via a property now?
[08:48] <ogra_> davmor2, popey, are you guys alive ?
[08:48] <sil2100> davmor2, popey: hey, you around?
[08:49] <kdub_> rsalveti,  well, we're increasing the number of framebuffers based on some of the properties
[08:49] <rsalveti> kdub_: right
[08:50] <rsalveti> sil2100: kdub_: so it'd be nice if it could be approved manually for now (proposed transition), and I'll split that package and make arch any
[08:50] <rsalveti> so we can have mir on amd64 working again
=== alan_g_ is now known as alan_g
[08:51] <kdub_> amd64? or arm64?
[08:51] <rsalveti> sorry, arm64
[08:55] <Mirv> sil2100: you could probably find someone over there to approve line 22 / landing-011 packaging changes?
[08:59] <popey> sil2100: ogra_ public holiday today
[08:59] <sil2100> popey: uh oh! Ok :)
[08:59] <popey> many UK and US people will be out I believe
[08:59] <popey> unless they're in Malta
[08:59] <popey> happy to dogfood this morning though ☻
[09:00] <popey> if thats what you were after
[09:01] <sil2100> popey: I know you're holidaying, but could you maybe publish gallery-app as click for us :) ?
[09:01] <popey> if someone pushes it to the store, sure
=== cjohnston changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp - sprinting this week, slow to respond | CI Train support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: -
[09:02]  * Mirv can push
[09:04] <sil2100> Oh, it wasn't pushed?
[09:05] <popey> i see no new gallery in the app store
[09:08] <popey> Mirv: just ping me when/if you push it
[09:08] <Mirv> I'm slightly uncertain if when upstream says they've tested it and we release gallery-app via CI Train, is it that they also tested the click package?
[09:08] <popey> thats why I always run the AP tests for click packages which I see in the store
[09:09] <sil2100> rsalveti: so, that dependency will be resolvable soon, right?
[09:12] <sil2100> Mirv, popey: so, the new version of gallery-app has only AP changes
[09:13] <sil2100> Mirv, popey: that was basically why we thought releasing it is not necessary, as I had the impression that phablet-click-test-setup fetches latest bzr
[09:13] <Mirv> sil2100: ok, so I can probably upload it then. it also seems it's functional on device (pkcon install-local)
[09:14] <Mirv> sil2100: oh right, it was that
=== jhodapp_ is now known as jhodapp
[09:15] <rsalveti> sil2100: yup
[09:16] <Mirv> popey: https://myapps.developer.ubuntu.com/dev/click-apps/507/
[09:20] <popey> sil2100: i spoke to someone at Malta last week who said gallery has more than just AP changes
[09:20] <popey> one of the QA guys
[09:20] <popey> anyway, testing now
[09:20] <sil2100> popey: well, I just know what I landed - we first landed something that had more than AP changes
[09:21] <sil2100> popey: but that got released to the store already
[09:21] <popey> yes
[09:21] <sil2100> popey: the next landing had only one merge which had only an AP change
[09:21]  * popey grabs 985 from the store
[09:21] <sil2100> popey: which is the thing we want to release
[09:22] <sil2100> https://code.launchpad.net/~artmello/gallery-app/gallery-app-autopilot_issue/+merge/220633 <- and this was the only change we pushed through train, so I guess nothing else should be in trunk waiting for release
[09:25]  * popey starts tests
[09:41] <popey> sil2100: ogra_ Mirv approved gallery app into the store
[09:41] <popey> Ran 37 tests in 867.284s
[09:41] <popey> OK
[09:41] <sil2100> popey: thank you!
[09:41]  * ogra_ hugs popey 
[09:41] <popey> np
[09:41] <sil2100> That is what we were expecting :)
[09:43] <robru> sergiusens, you in malta?
[09:48] <ogra_> robru, he is ... (room C2)
[09:48] <robru> ogra_, where is that?
[09:48] <thomi> fginther: I replied to your email, but I'm wondering whether I can read from the DB that the ci dashboard uses directly - who would i talk to about that ideally?
[09:48] <ogra_> end of the long corridor ... last room on the right
[09:49] <robru> ogra_, thanks
[09:49] <ogra_> (if you come from the coffe ara)
[09:49] <Mirv> robru: he was last week at least
[09:49] <Mirv> but yeah he had a long trip
[09:51] <fginther> josepht, is there an API that exposes the CI Dashboard image test data?
[09:52] <plars> rsalveti: hi, we're looking at using the emulator to run tests, and there's an active MP to lay the groundwork for that out there, but we're hitting some crashes in the emulator, especially when trying to run larger testsuites like webbrowser. Have you seen this?
[09:53] <cjwatson> sil2100,alan_g: so I could remove mir/arm64, but this looks entirely trivial to fix in mir, it just needs:
[09:53] <cjwatson> -               libandroid-properties-dev,
[09:53] <cjwatson> +               libandroid-properties-dev [i386 amd64 armhf],
[09:53] <alan_g> kdub_: ^
[09:53] <cjwatson> it used to be [armhf], but that was removed in lp:mir/devel r1458.1.14, presumably in order to be able to use it on x86
[09:54] <rsalveti> plars: one is bug 1319582
[09:54] <ubot5> bug 1319582 in android (Ubuntu) "emulator: 'Failed to start RenderThread' after opening/closing applications" [High,Confirmed] https://launchpad.net/bugs/1319582
[09:54] <rsalveti> plars: but I'd like to see the issues
[09:55] <rsalveti> so we can try to isolate them this week, and try to get them fixed
[09:55] <cjwatson> the code that's using that header is already guarded by MIR_BUILD_PLATFORM_ANDROID, which is already off on arm64
[09:55] <cjwatson> kdub_: ^-
[09:56] <plars> rsalveti: cool, I'll look to see if that's the same thing I'm getting... that error is what I can expect on stdout/stderr from where I run the emulator?
[09:56] <kdub_> cjwatson, iirc, think we build both platforms on the different architectures
[09:58] <thomi> trainguards - I wonder if someone could please reconfigure silo 19? I've had to add a new MP that pulls in a second source package
[09:58] <kdub_> if adding the architecture designations in debian/control fixes the problem, i'm okay with that too
[09:58] <sil2100> thomi: sure, on it!
[09:58] <thomi> sil2100: thanks
[09:59] <rsalveti> plars: yes
[09:59] <sil2100> thomi: done! yw
[09:59] <thomi> sil2100: awesome
[09:59] <cjwatson> kdub_: No, I checked, we don't build android on arm64.  Indeed I'm pretty sure I introduced the change to not build android on arm64.
[10:00] <cjwatson> kdub_: See debian/rules inside "ifeq ($(DEB_HOST_ARCH),arm64)".
[10:00] <kdub_> cjwatson, ah, right. i keep reading arm64 as amd64
[10:00] <cjwatson> Yeah, common problem :-)
[10:00] <cjwatson> kdub_: We do indeed build both platforms on i386/amd64
[10:01] <cjwatson> (and armhf of course)
[10:01] <kdub_> right, so that debian/control change makes more sense
[10:01] <kdub_> although, building libproperties for arm64 also makes sense
[10:03] <cjwatson> yeah, I don't really mind which, though the mir change would be quicker I expect
[10:03] <cjwatson> and is correct anyway right now given the platform build setup
=== didrocks changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: cihelp - sprinting this week, slow to respond | CI Train support - US: robru, rsalveti - EU: sil2100, Mirv | CITrain support no answer: use mup bot after 30 minutes, but choose right timezone | Known issues: -
[10:12] <fginther> thomi, there isn't an available to read from the dashboard db
[10:12] <thomi> fginther: because it's not stored?
[10:13] <thomi> fginther: or because it's hidden behind a firewall?
[10:13] <fginther> thomi, there isn't a externally excessible API
[10:13] <thomi> fginther: but if the data is stored in postgres, and you guys are really nice to me, perhaps I could connect to that read-only?
[10:14] <thomi> wouldn't need to do it very often either
[10:14] <thomi> hell, even regular (perhaps weekly) sql data dumps would work :)
[10:19] <plars> rsalveti: I'm not seeing a message like that, just a segfault and the emulator dies: http://paste.ubuntu.com/7520332/ It could be that it's just hidden somewhere though. Maybe I could retry once there's a fix for this one
[10:21] <rsalveti> plars: seg fault is interesting, have the core dump?
[10:21] <rsalveti> plars: mostly happening with the webbroser-app?
[10:22] <rsalveti> can try reproducing that as well here
[10:22] <plars> rsalveti: I haven't tried others except webbrowser and friends_app
[10:22] <plars> rsalveti: it happens reliably for me with webbrowser though, and pretty quickly too I think
[10:22] <rsalveti> great, I'll check
[10:23] <plars> rsalveti: I think it had gone through 8 or 9 tests before it crashed for me this last time
=== Ursinha is now known as Ursinha-afk
=== Ursinha-afk is now known as Ursinha
[11:53] <plars> rsalveti: another question we had - is there a way to control the adb serial used by the emulator?
[12:08] <Mirv> sil2100: robru: I start doing test builds in the Qt 5.3 test silo for some of our components too. so if you see any new prepare-silo landing conflicting with "landing-005", just ignore conflicts.
[12:09] <sil2100> Mirv: ok, sure
[12:09] <robru> Mirv, thanks
[13:18] <plars> rsalveti: there's another failure mode I'm seeing too, where it seems to drop out pretty fast back to a shell, but the console isn't echoing keyboard input when you type (have to reset the console) http://paste.ubuntu.com/7521356/
[13:18] <plars> any way to get more debug data from it other than the .crash file?
[13:19] <rsalveti> plars: yeah, in this case the crash file would be more useful
[13:24] <robru> tedg, you around? had a question about your pay-service landing in silo 17
[13:42] <robru> dbarth, ping?
[13:45] <dbarth> robru: pong
[13:45] <robru> dbarth, hey are you in malta? how's it going?
[13:45] <dbarth> hiya
[13:45] <dbarth> back in paris this week
[13:46] <dbarth> was going through mp and bugs, but i have a few silos to land i think
[13:47] <robru> dbarth, yes, I was just noticing you have a few silos that look like they are moving a little bit slowly. I was just wondering if you needed any help testing those
[13:51] <dbarth> robru: I appreciate any help ;)
[13:51] <dbarth> robru: the big OA silo i haven't touched yet
[13:51] <dbarth> line 16
[13:52] <dbarth> the greasemonkey one, also; i have some weird behavior
[13:53] <robru> hmmm
[13:53] <robru> dbarth, well if you can write me an email with some testing instructions I would be happy to report my results
[14:11] <plars> rsalveti: I have a ~200MB .crash file, but I don't think I have debug symbols to make it very useful
[14:12] <rsalveti> plars: can you upload it somewhere and open a bug against the android package?
[14:25] <robru> dbarth, replied to your email
[14:34] <dbarth> ok
[14:34] <dbarth> robru: then i guess you can free up this one until justinmcp gets back to us
[14:34] <plars> rsalveti: https://bugs.launchpad.net/ubuntu/+source/android/+bug/1323320 (attachment took a bit)
[14:34] <ubot5> Ubuntu bug 1323320 in android (Ubuntu) "[EMULATOR] XIO: fatal IO error 4 (Interrupted system call) on X server ":99"" [Undecided,New]
[14:35] <robru> dbarth, thanks!
[14:35] <rsalveti> plars: thanks
[14:45] <ev> https://rt.admin.canonical.com/Ticket/Display.html?id=68826
[15:48] <sil2100> ogra_: let's skip this meeting now - there's not enough time till dinner and besides, not much changed from the morning
[15:48] <sil2100> ogra_: + all US guys from the landing meeting are here with us on the sprint
[15:50] <ogra_> sil2100, slacker :P
[15:50] <sil2100> ogra_: not slacker! I'm too busy for meetings ;)
[15:51] <ogra_> (i'm sitting two tables left btw )
[15:51] <ogra_> haha
[15:57] <mhr3> sil2100, silo? pretty pls?
[15:58] <sil2100> mhr3: doing! Sorry, got really code-occupied :)
=== Ursinha is now known as Ursinha-afk
[18:43] <bzoltan> rsalveti: could you please assign a silo to the line 38?