[02:05] <imgbot> === IMAGE 58 building (started: 20141218-02:05) ===
[03:30] <imgbot> === IMAGE 58 DONE (finished: 20141218-03:30) ===
[03:30] <imgbot> === changelog: http://people.canonical.com/~ogra/touch-image-stats/58.changes ===
[06:13] <robru> alesage: is there something wrong with the deb in the PPA? https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-016
[06:14] <robru> alesage: citrain only has source packages in the server, and they're not exposed to the web (although I could reach in and tar one up if you wanted it). all the actual built debs are just in lp.
[06:15] <bzoltan_> robru: Mirv: is there any news about the Vivid landings?
[06:15] <robru> bzoltan_: what about vivid landings? normal as far as I know.
[06:16] <bzoltan_> robru:  normal? Yesterday there was a massive block because of the autopkgtest ...
[06:16] <Mirv> robru: vivid landings are stuck because autopkgtests fail
[06:16] <bzoltan_> robru:  the UITK is at least blocked
[06:16] <robru> bzoltan_: ah I wasn't aware, sorry. haven't heard anything
[06:16] <Mirv> robru: because of bug #1399597 that has no fix that wouldn't regress something else
[06:16] <ubot5> bug 1399597 in oxide-qt (Ubuntu) "The last oxide upload breaks autopkgtests with a package installation failure" [Critical,Confirmed] https://launchpad.net/bugs/1399597
[06:17] <Mirv> robru: so currently we can either a) break SDK, b) break many packages from migrating to vivid-release, c) bring oxide-qt-codecs-extra to desktop seed, none of which is wanted. we're currently again at b) after the last failed attempt
[06:17] <bzoltan_> Mirv:  why the c) is bad?
[06:18] <robru> Mirv: yikes, I've been so neck-deep in train code I haven't even noticed what's been happening in the archive
[06:18] <Mirv> bzoltan_: because it'd bring h.264 support to the cd images, which is a patent problem
[06:18] <Mirv> (or so I understood it)
[06:18] <robru> Mirv: how is the sdk broken by this? surely the least-bad option is to break just sdk temporarily until they can fix it?
[06:19] <Mirv> robru: click chroots fail to create, and we've just some Chinese SDK usage thing starting if I understood correctly. now it's working, but if we revert to the original situation then it will be broken.
[06:20] <Mirv> the technically correct thing would be to revert to the original situation, but no-one wants to break SDK either..
[06:20] <Mirv> so, currently unity-scope-click autopkgtest fails at http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html which prevents at least UITK, but also things like Unity8 from migrating if they're published
[06:20] <robru> Mirv: technically correct is best correct ;-)
[06:22] <Mirv> and it seems apt problem really, since it tries to configure a package that dpkg already removed
[06:39] <Mirv> bzoltan_: I think chrisccoulson had an idea on how to fix the problem, but I'm not sure if he's having the time to do that. something about moving files around and using alternatives system.
[06:43] <bzoltan_> Mirv:  waiting for the magic wand when actually nobody has time time to fix the problem i snot really a good solution
=== chihchun_afk is now known as chihchun
[07:23] <bzoltan_> robru: breaking the SDK would mean no app development for Vivid target and no qmake support .. so technically it would roll back the SDK by 3 months
[07:24] <robru> bzoltan_: oh you wouldn't have to roll back 3 months. you'd just be blocked temporarily until you found a workaround.
[07:25] <bzoltan_> Mirv:  so what pulls the extras to the desktop?
[07:25] <bzoltan_> robru: We are looking for the workaround since Septermber ...
[07:28] <Mirv> bzoltan_: if the order is switched which picks the autopkgtests, the extra package is listed first so it gets installed on desktop too instead
[07:29] <Mirv> so this is what chris rejected but was temporarily in archives https://code.launchpad.net/~timo-jyrinki/oxide/prefer_codecs_extras_lp1399597/+merge/243946
[07:30] <Mirv> then this https://code.launchpad.net/~timo-jyrinki/oxide/arch_specific_replaces_lp1400275/+merge/244307 was the next attempt to keep chris happy but not break sdk or autopkgtests.. 50% succeeded, or not really, the arch specific replaces did not help apt to not break autopkgtests
[07:30] <Mirv> so we're essentially in the same place as after https://code.launchpad.net/~bzoltan/oxide/packaging.vivid-improved-conflicts/+merge/243515
[07:32] <Mirv> apparently Chris himself has not had time for working on this however, unfortunately
[07:32] <Mirv> he should be pinged today whether he could work on his newest idea today
[07:33] <Mirv> I'll free up that one silo that couldn't be landed anyway
[07:42] <bzoltan_> Mirv: it sounds like a super top priority issue ...
=== chihchun is now known as chihchun_afk
[07:49] <Mirv> I made the reopened bug Critical, and yes I agree it should be top priority
[08:45] <bzoltan_> Mirv:  what was the package what directly pulls the extra and confused the apt?
[08:46] <Mirv> bzoltan_: ubuntu-sdk-libs
[08:47] <Mirv> no idea if it would help if that dependency was optional too but the other way around. I don't know if the seeds support that.
[08:47] <bzoltan_> Mirv: why do not we remove that oxideqt-codecs-extra from the seeds then?
[08:47] <Mirv> bzoltan_: because that was the requirement to get it on the phone images. ogra_ would know more, and might also have a solution for us...
[08:48] <Mirv> like, could the dependency be pulled via other route, or would http://bazaar.launchpad.net/~ubuntu-core-dev/ubuntu-seeds/ubuntu-touch.vivid/view/head:/sdk-libs support "* oxideqt-codecs-extra | oxideqt-codecs" (in case that would help with the autopkgtests)
[08:48] <bzoltan_> Mirv:  OK.. so that is a fundamental _BUG_ we want a package directly burned on the phone image, but we do not want the same  package on the desktop image ... that is the mother of the whole conflict.
[08:49] <Mirv> oh but he's on vacation now
[08:49] <cjwatson> seeds can't do or-ed dependencies, no
[08:49] <cjwatson> doesn't really make sense, their purpose is to specify choices
[08:50] <Mirv> ok then that's not an option
[08:51] <Mirv> bzoltan_: we also have the problem that "SDK" is used as "what we want on images"
[08:51] <bzoltan_> Mirv:  true
[08:51] <Mirv> or I'm not sure if that's a problem, but slightly confusing at least
[08:52] <bzoltan_> Mirv:  that is a problem...  we force the extras on the phone image but do not want to see in on the desktop ... that does not sound good
[08:53] <Mirv> bzoltan_: one more thing: since it's unity-scope-click that's failing the autopkgtest, I wonder why it actually depends on ubuntu-sdk-libs as a whole
[08:53] <bzoltan_> Mirv:  if the oxideqt-codecs-extra is needed on the image then it should be pulled from other place than the sdk
[08:53] <Mirv> alecu: pstolowski: ^ unity-scope-click question
[08:55] <bzoltan_> Mirv:  that is a good question too.. I guess the sdk-libs is just a swiss army knife of solving missing dependencies :D
[08:56] <Mirv> bzoltan_: changelog says " Add runtime dependency on ubuntu-sdk-libs for the frameworks list." <- how this could be achieved otherwise?
[08:57] <bzoltan_> Mirv:  hmm... My guess was right .. let me check
[08:57] <Mirv> original bug filed by mvo in May https://bugs.launchpad.net/ubuntu/+source/unity-scope-click/+bug/1320975
[08:57] <ubot5> Launchpad bug 1320975 in unity-scope-click (Ubuntu) "Shows available click packages even if no framework is installed" [High,Fix released]
[08:59] <Mirv> bzoltan_: holy... that meta package is not meta package, it's true what it says
[08:59] <Mirv> it includes all the framework files in /usr/share/click/frameworks/ :(
[08:59] <bzoltan_> Mirv:  baaaaah
[09:00] <bzoltan_> Mirv:  and that is aaaaal wrong
[09:00] <bzoltan_> llll
[09:01] <Mirv> I don't know the original reasoning, but what about splitting that to ubuntu-sdk-frameworks or something and fixing unity-scope-click dependency...
[09:02] <bzoltan_> Mirv:  at least that one
[09:02] <bzoltan_> Mirv:  that is an ugly crap ...
[09:13] <Mirv> bzoltan_: lool pstolowski alecu: https://code.launchpad.net/~timo-jyrinki/ubuntu/vivid/ubuntu-touch-meta/split_frameworks/+merge/245078 + https://code.launchpad.net/~timo-jyrinki/unity-scope-click/switch_to_frameworks_dependency/+merge/245080 - makes sense?
[09:13]  * Mirv coffee break before hangout
[09:14] <pstolowski> Mirv, click scope depends on sdk libs to get the list of frameworks, and pass them to the store server
=== pstolowski is now known as pstolowski|schoo
[09:14] <bzoltan_> Mirv: SHIP IT
[09:14] <pstolowski|schoo> Mirv, i need to drop now to sort out some school stuff.. bbl if you need more info
[09:14] <bzoltan_> pstolowski|schoo: and now we fix that the way how it should have been done :)
[09:15] <pstolowski|schoo> bzoltan_, hmm ok, make sure to keep alecu in the loop, he may have more to say on that
[09:16] <bzoltan_> pstolowski|schoo: it was a mistake at the first place to overload the sdk-libs with the framework list
[09:18] <bzoltan_> Mirv:  do you have the rights to fix the seeds?
[09:20] <bzoltan_> Mirv:  it is safe to move the frameworks to an other package as no other package in Vivid depends on the ubuntu-sdk-libs
[09:21] <cjwatson> Mirv: the package that contains the framework declaration has to depend on the packages that actually satisfy the framework, otherwise it's utterly bogus
[09:22] <cjwatson> Mirv: so it might as well be part of ubuntu-sdk-libs, it doesn't really help to have it split
[09:23]  * cjwatson nacks that MP, sorry
[09:24] <cjwatson> pstolowski|schoo: that doesn't make sense - surely it should just use whatever frameworks are present on the system
[09:24] <cjwatson> pstolowski|schoo: ubuntu-sdk-libs is always going to be present on phone images; perhaps other images might have something else.  at any rate it is not unity-scope-click's job to specify that
[09:25] <Mirv> cjwatson: ok then.
[09:25] <cjwatson> pstolowski|schoo: you should clearly get the list of installed frameworks from click rather than hardcoding it; but I think you should drop the ubuntu-sdk-libs dep
[09:25] <Mirv> possibly the unity-scope-click could be made not to depend on it anyhow
[09:27] <Mirv> updated https://code.launchpad.net/~timo-jyrinki/unity-scope-click/switch_to_frameworks_dependency/+merge/245080 to drop it instead..
[09:30] <bzoltan_> cjwatson:  by the time this framework list was added (wrongly) to the sdk-libs the click have not yet provided the fw list
[09:31] <sil2100> jibel: hey! How's the testing looking sp far?
[09:32] <bzoltan_> Mirv:  yes, dropping that dependency is good ...I hope pstolowski|schoo can fix the unity-scope-click to use the click API instead
[09:32] <jibel> sil2100, so far it's okay
[09:32] <jibel> sil2100, nearly 60% complete, it should be finished by end of this afternoon
[09:33] <sil2100> Oh my!
[09:33] <cjwatson> bzoltan_: while it didn't have an API for it, I distinctly remember saying "just do the equivalent of readdir(/usr/share/click/frameworks/) until I provide one"
[09:33] <sil2100> You guys are awesome ;)
[09:33] <bzoltan_> cjwatson: yes, I do remember that
[09:34] <cjwatson> bzoltan_: which is basically all that click's API does, so while it would be better for unity-scope-click to use that, it is not a blocker here
[09:34] <bzoltan_> cjwatson:  yes
[09:39] <ogra_> bzoltan_, so when we talked, i asked you to talk to the webapps team to find out if -extras is actually a requitement of the framework ... did you ask them ?
[09:40] <ogra_> if it isnt, moving the package to touch and desktop-next (or only to touch) and just dropping it from sdk-libs is trivial ...
[09:40] <cjwatson> good point
[09:41] <bzoltan_> ogra_: Let me ask them again... I do not remember an exact answer.
=== psivaa_ changed the topic of #ubuntu-ci-eng to: Need a silo? ping trainguards | Need help with something else? ping psivaa | Train Dashboard: http://bit.ly/1mDv1FS | QA Signoffs: http://bit.ly/1qMAKYd | Known Issues: lab move complete. RTM Archive closed, milestone image testing started.
[10:17] <cjwatson> sergiusens: did you run "crontab /srv/cdimage.ubuntu.com/etc/crontab" on nusakan?
[10:19] <bzoltan_> ogra_:  I just had a talk with dbarth about the codecs-extra. So they do need that package on the device image, because a bunch of webapps need it. But they do not specifically need it to be pulled from the ubuntu-sdk-libs. Could we move it to touch seed?
[10:29] <dbarth> ogra_: also, to explain: we don't strictly depend on oxide, since we still have that runtime switch that can take oxide or qtwebkit
[10:30] <dbarth> so at the moment, we still can't fix that with a normal dependency
[10:34] <jibel> Mirv, I had this gallery-app crash again first time I save a photo from the webbrowser. Next tries work fine.
[10:39] <Mirv> jibel: hmmkay
[10:40] <Mirv> bzoltan_: ogra_ is still on vacation, but I offer https://code.launchpad.net/~timo-jyrinki/ubuntu-seeds/ubuntu-touch.vivid_move_oxideqt-codecs-extra_to_touch/+merge/245087 now...
[10:40] <jibel> Mirv, however I am still not sure it is caused by 1113 or it existed in previous versions and is just difficult to reproduce
[10:41] <jibel> I'll dig further
[10:41] <bzoltan_> Mirv:  That would satisfy my needs perfectly. Who's approval do we need to roll it out?
[10:45] <Mirv> bzoltan_: cjwatson rejected the previous one, so perhaps him
[10:46] <cjwatson> yes, I can do that
[10:47] <cjwatson> want me to deal with the upload?
[10:47] <bzoltan_> cjwatson: thank you for your help
[10:48] <bzoltan_> Mirv: ^^
[10:57] <Mirv> cjwatson: yes, that'd be welcome
[10:57] <cjwatson> ok, on its way
[10:58] <Mirv> thanks! I'll kick the autopkgtests after it has migrated and update the bug once again
[10:58] <cjwatson> you probably won't need to kick them manually
[10:58] <cjwatson> after all, this situation arises in part because packages depend on ubuntu-sdk-libs; so the automatic system should suffice
[10:58] <cjwatson> but I guess we'll see
[11:04] <Mirv> right probably it will sort out automatically
[11:07] <cjwatson> https://launchpad.net/ubuntu/+source/ubuntu-touch-meta/1.207
=== pstolowski|schoo is now known as pstolowski
[11:10] <sergiusens> cjwatson: yes :(
[11:13] <cjwatson> d'oh
[11:13] <cjwatson> don't suppose you have crontab -l output in history
[11:13] <cjwatson> failing that I think we resurrected most of it ...
[11:13] <cjwatson> the real problem here is IMO that system-image shouldn't be running as the cdimage user
[11:13] <cjwatson> if it were separate, it could more easily have its own crontab
=== _salem is now known as salem_
[11:24] <sergiusens> cjwatson: no I don't, I foolishly trusted the first lines in the crontab that said that it should be in /srv/cdimage.ubuntu.com/etc/crontab if you wanted your changes to survive :-/
[11:25] <cjwatson> That doesn't quite imply the converse :-)
[11:44] <jibel> Mirv, there are 2 other problems with the gallery-app, cropping is broken for several testers, and the rotate action disappeared
[11:55] <Mirv> jibel: sounds bad
=== psivaa_ changed the topic of #ubuntu-ci-eng to: Need a silo? ping trainguards | Need help with something else? ping cihelp | Train Dashboard: http://bit.ly/1mDv1FS | QA Signoffs: http://bit.ly/1qMAKYd | Known Issues: lab move complete. RTM Archive closed, milestone image testing started.
[12:45] <Mirv> ^ robru sil2100: the big vivid migration problem resolved, at least for now, via ubuntu-touch-meta package update. possible further actions in unity-scope-click & oxide-qt, but they are not critical to landing other stuff.
[12:48] <sil2100> Mirv: thanks!
[12:48] <sil2100> jibel: geh, are thosw promotion blockera? Can we somehow revert latest gallery from the store?
[12:50] <jibel> sil2100, so the new crop issue seems to be limited to wrong coordinates when you first crop an image. So the result is wrong, this is a regression. Crop not working at all has already been reported in 173
[12:51] <jibel> sil2100, the frequent crashes of gallery-app are more worrying. We should probably prepare for a revert of the gallery
[12:51] <sil2100> Ok...
[12:52] <sil2100> Its a click app so we need bfiller best to revert in trunk
[12:52] <sil2100> Or, hm, maybe theres a revert function in the store?
[12:52] <jibel> sil2100, the changelog doesn't say anything about an update of the crop feature
[12:58] <Mirv> jibel: because how the rtm branch was set up, in addition to the latest changes the branch had unreleased commits 1102-1111, of which 4 had actual changes other than translation updates. those don't seem to touch cropping either - only album changes in all of those.
[12:59] <Mirv> http://bazaar.launchpad.net/~phablet-team/gallery-app/rtm-14.09/changes
[13:00] <Mirv> that deducting from the previous version's version number including "1101", before rtm branch existed
[13:04] <Mirv> thanks cjwatson for making vivid + bzoltan + etc happy, all seems fine for now and I've noted on the bug on the possible further actions for unity-scope-click & oxide-qt but which are less critical
[13:06] <cjwatson> Mirv: cool
[13:08] <cjwatson> Mirv: I've clarified a couple of points
=== chihchun_afk is now known as chihchun
[13:11] <Mirv> thanks, better
[14:08] <jibel> Mirv, the crash of the gallery app could be due to the thumbnailer upload on the 16th
[14:09] <jibel> Mirv, om26er can reproduce it easily and will try previous version
[14:21] <Mirv> jibel: ok. the diff was this interesting http://launchpadlibrarian.net/191646180/thumbnailer_1.3%2B14.10.20141020-0ubuntu1_1.3%2B15.04.20141106~rtm-0ubuntu1.diff.gz together with qtubuntu-camera from rsalveti / jhodapp / satoris
[14:25] <Mirv> the qtubuntu-camera was huge: http://launchpadlibrarian.net/192809040/qtubuntu-camera_0.3.3%2B14.10.20141001-0ubuntu1_0.3.3%2B15.04.20141208.is.0.3.3%2B15.04.20141204~rtm-0ubuntu1.diff.gz (publishing history seems to confirm it's correct: https://launchpad.net/ubuntu-rtm/+source/qtubuntu-camera/+publishinghistory )
[14:25] <Mirv> I've updated the bug #1376500 which the landing fixed, which was also not updated to indicate it would have been fixed in the first place
[14:25] <ubot5> bug 1376500 in thumbnailer (Ubuntu RTM) "thumbnails in metadata generated incorrectly" [Undecided,In progress] https://launchpad.net/bugs/1376500
[14:27] <Mirv> the trello comment on that landing does not seem to indicate the "silo-diff" (from brendand) was used to actually check the rtm diff, as it only mentions the silo fixing two bugs
[14:28] <Mirv> jibel: crash was also mentioned in the trello: https://trello.com/c/zpBMvnsD/443-ubuntu-rtm-landing-012-qtubuntu-camera-thumbnailer-jhodapp-satoris
[14:32] <jibel> bfiller, hey, the crash of the gallery app seems to be caused by the thumbnailer. brendand and om26er both confirmed that reverting this package fixes the issue
[14:32] <jibel> Mirv, bfiller what would be the impact of a revert of this package?
[14:33] <Mirv> jibel: see above for the diff:s. it would help if only thumbnailer needs to be reverted.
[14:34] <bfiller> Mirv: can you paste the diff link again
[14:34] <bfiller> Mirv, jibel: I would think we'd need to revert the qtubuntu-camera and thumbnailer together but need to see what changed
[14:35] <Mirv> bfiller: http://pastebin.ubuntu.com/9560672/
[14:35] <jibel> bfiller, http://launchpadlibrarian.net/191646180/thumbnailer_1.3%2B14.10.20141020-0ubuntu1_1.3%2B15.04.20141106~rtm-0ubuntu1.diff.gz
[14:36] <Mirv> both of the landings were some weird landings anyhow wrt changelogs
[14:36] <Mirv> qtubuntu-camera http://launchpadlibrarian.net/192809040/qtubuntu-camera_0.3.3%2B14.10.20141001-0ubuntu1_0.3.3%2B15.04.20141208.is.0.3.3%2B15.04.20141204~rtm-0ubuntu1.diff.gz
[14:38] <bfiller> jibel: is there a stack trace for the thumbnailer crash?
[14:38] <jibel> bfiller, no there is nothing
[14:39] <bfiller> jibel: how do we know it's thumbnailer then?
[14:39] <jibel> bfiller, there are several uploads to errors.u.c but none of them has been retraced successfully
[14:40] <jibel> bfiller, we started seeing the crash on 181 and the crash doesn't happens if you downgrade thumbnailer to 1.3+14.10.20141020-0ubuntu1
[14:40] <bfiller> jibel: if the crashes can't be reproduced with the old thumbnailer I would be in favor of reverting it. The change there is to use the thumbnail out of the exif file rather than generate it on it's own
[14:41] <bfiller> jibel: reverting it shouldn't have any major negative effect, should be like before. The thumbnail generation will be slow and memory intensive but that is what it's been up until now anyway
[14:41] <bfiller> I'd say this is a risky fix for not a huge gain right now
[14:41] <bfiller> +1 for revert
[14:41] <bfiller> I'm just not sure about qtubuntu-camera
[14:42] <om26er> bfiller, maybe ask jhodap about that ?
[14:42] <bfiller> om26er: yeah don't see him online
[14:43] <om26er> ugh
[14:43] <jibel> brendand, om26er Mirv okay to revert thumbnailer?
[14:44] <om26er> yes, makes sense to me.
[14:49] <Mirv> jibel: om26er brendand bfiller: I'm past my EOD and already late for a meeting, but landing line 40 / silo rtm-000 has revert of both. land either one or both, and please double check with silo-diff they are good to go of course + test. I've also filed bug #1403906 about it including the IRC discussions as description.
[14:49] <ubot5> bug 1403906 in thumbnailer (Ubuntu RTM) "Crash using gallery with new qtubuntu-camera and thumbnailer" [Critical,New] https://launchpad.net/bugs/1403906
[14:49] <Mirv> I can peek back before going to sleep in 3-4h or so.
[14:50] <bfiller> Mirv: thank you
[14:50] <jibel> Mirv, thanks
[14:54] <jibel> Mirv, before you leave who can rebuild an image?
[14:54] <Mirv> jibel: I don't know, I think that was when we talked slangasek should be pinged if we don't know any better. for me it has always been just og_ra who handles it.
[14:54] <jibel> Mirv, ack
[15:23] <cjwatson> jibel: in general: intersection of ~ubuntu-cdimage and ~canonical
[15:24] <jibel> cjwatson, intersection of ~ubuntu-cdimage and ~canonical and ! on holidays :)
[15:24] <cjwatson> well yes
[15:24] <cjwatson> gonna be "nobody" pretty soon :P
[15:25] <cjwatson> I can do it for about the next two hours
[15:25] <jibel> om26er, ^ we need silo 000
[15:26] <om26er> jibel, I am in the process of installing it
[15:44] <jibel> brendand, reverting also fixes the problem with cropping in the gallery
[15:44] <brendand> jibel, reverting thumbnailer?
[15:45] <jibel> brendand, yes
[15:45] <brendand> jibel, that must have been the issue - the only time it worked for me was when i used a photo from interwebz
[15:51] <jibel> om26er, I checked camera, gallery, mediaplayer, photo and video scopes. checked that thumbnails are correctly generated, that apps don't crash on launch or switching between photos, and that editing features of the gallery work fine.
[15:55] <om26er> jibel, I can't crash it either
[15:56] <jibel> om26er, k, I marked it pass for upstream
[15:57] <jibel> om26er, do it for QA when you think it's good
[15:57] <slangasek> jibel, Mirv: fwiw I told sil2100 that I'm reachable by SMS if images need respun
[15:57]  * om26er resets multiples times just to be sure
[15:58] <jibel> slangasek, thanks, we'll need an image in less than an hour when silo 000 is verified and in the archive.
[15:59] <slangasek> jibel: ok; just ping me here or by phone when you're ready for the build
[15:59] <cjwatson> I'll still be here in an hour and am not yet on holiday
[15:59] <cjwatson> FWIW
[15:59] <slangasek> or cjwatson, since as he said he's around for the next hour
[15:59] <slangasek> right
[15:59] <slangasek> jibel: ^^ so just give it to cjwatson, and I'll go find my pillow again ;)
[16:00] <cjwatson> it's down the back of the sofa no wait
[16:00] <jibel> slangasek, heh, enjoy your pillow :)
[16:06] <sil2100> I'm back home now and waiting for my system to boot
[16:07] <sil2100> What component will be reverted? Gallery or something else?
[16:08] <sil2100> i.e. how much retesting will be needed on the new image?
[16:08] <jibel> sil2100, thumbnailer and qtubuntu-camera
[16:08] <jibel> sil2100, it's in silo 0
[16:08] <jibel> sil2100, they also broke cropping in the gallery app
[16:09] <om26er> jibel, sil2100 approved.
[16:09] <sil2100> oh, hm, so it was the landing that Ricardo was meddling with versions
[16:09] <sil2100> Ok, I'm on my PC now
[16:10] <sil2100> Let me publish
=== chihchun is now known as chihchun_afk
[16:11] <sil2100> jibel, om26er: ok, everyone sure I should publish 000?
[16:11] <jibel> sil2100, why wouldn't we?
[16:12] <om26er> if the revert diff is exact then yes ;)
[16:12] <sil2100> ...published
[16:12] <sil2100> qtubuntu-camera in the end is an revert that reverts another revert
[16:13] <jibel> sil2100, yeah, they should really try harder to land small fixes rather than this big syncs from vivid
[16:14] <sil2100> I think landing syncs that pull in more than one version should be automatically considered risky as well
[16:17] <sil2100> Ok, packages should migrate any minute
[16:18] <sil2100> jibel: will this mean we need to re-do the whole regression suite? Or will you only re-test the delta for those? (so anything with thumbnails and camera)
[16:18] <jibel> sil2100, anything with camera, thumbnails and which uses exif metadata
[16:21]  * sil2100 watches rmadison
[16:24] <sil2100> jibel: remember that LP is always a bit fast with migration info, so even if the queuebot says it migrated then there's still a few moments that we need to wait for it to be 100% true
[16:24] <jibel> sil2100, I'm checking with rmadison
[16:24] <sil2100> So let's wait for rmadison to properly register it
[16:30] <sil2100> Ok, qtubuntu-camera is in, now just thumbnailer is left
[16:35] <sil2100> jibel: ok, it seems that both packages migrated now, rmadison is happy, update_output as well
[16:36] <sil2100> cjwatson: hey! Can we ask you for a new ubuntu-rtm image? :)
[16:36] <cjwatson> sil2100: running
[16:36] <sil2100> cjwatson: thank you!
[16:36] <cjwatson> np
[16:36] <jibel> sil2100, cjwatson thanks
[16:37] <sil2100> jibel: so we should have a new image in ~1.5h - is the rest of regression tests finished?
[16:37] <sil2100> jibel: will you only have to re run the delta now?
[16:37] <jibel> sil2100, sanity tests + delta
[16:38] <sil2100> Fingers once again crossed that this revert won't cause any regressions
[16:40] <imgbot> === IMAGE RTM 184 building (started: 20141218-16:40) ===
[16:40] <dobey> cihelp: if the autopilot-testrunner-otto-vivid job is running tests with phablet-test-run, is the lxc the tests are being run in, creating a mir surface somehow and running adbd in the lxc?
[16:42] <dobey> tail: '/var/lib/lxc//vivid-amd64-20141218-1841/run/delta//home/ubuntu/.xsession-errors' has become accessible
[16:42] <dobey> hmm, i guess that's not mir
[16:42] <fginther> dobey, autopilot-testrunner-otto-vivid uses autopilot directly to run the tests. And it's not doing anything with mir, it essentially uses the vivid desktop ISO which would be unity7
[16:42] <dobey> fginther: oh ok
[16:45] <dobey> fginther: thanks. i'm trying to understand what's different there, versus running the autopilot tests via autopkgtest with adt-run under qemu (which is how the ubuntu autopkgtests jobs in jenkins run the autopkgtests)
[16:45] <dobey> because the autoplit tests seem to be very unstable for unity-scope-click at least
[16:46] <dobey> and of course it doesn't help that qemu seems to be getting SIGTERMed randomly for me locally
[16:47] <fginther> dobey, I've never had good luck with qemu, it's very unreliable IMHO
[16:48] <fginther> dobey, or at the very least, there are some things that just can't be done there (threads being one of them I think)
[16:48] <dobey> fginther: well, i think i've got the tests equally unreliable in both the autopilot-testrunner-otto-vivid job, and in my autopkgtests branch running under qemu now (save for the qemu being killed bit)
[16:50] <dobey> though, creating the user session or starting Xvfb sometimes fails for no reason in the qemu case too. but when the tests actually run, they seem to fail in the same way for me in both places, now, which is at least good
[16:50] <fginther> dobey, otto has it's own problems... It was built to quickly fill a need and did well, but we've discovered numerous things that just don't work well under lxc
[16:50] <dobey> but there's no way we can land my branch :(
[16:50] <dobey> fginther: yeah, lots of things don't work well there i'm sure
[16:51] <fginther> dobey, we've had to disable otto testing for some projects in CI simply because that environment doesn't work anymore
[16:51] <fginther> dobey, for the most projects, the phone is still the better test environment (Just not sure that applies in your case)
[16:53] <dobey> fginther: well, my task was to get autopilot tests running in autopkgtests for unity-scope-click
[16:53] <dobey> the problem is we can't actually run things in autopkgtests that aren't reliable and passing, because they will block things landing in archive, if they break
[16:54] <cjwatson> fginther: threads are unreliable under qemu-user-static; I've never heard of them being unreliable under full-system emulation
[16:54] <cjwatson> fginther: if they were then all virtualised PPAs would be basically entirely screwed
[16:54] <dobey> and our tests have been unreliable in the otto job as well for various reasons
[16:56] <dobey> but right now in this branch, i seem to have the tests themselves in the same state of unreliability in both places
[16:56] <dobey> which is both a good sign and a bad sign :)
[16:57] <dobey> unfortunately, i don't know how to improve them further at this point, or why qemu is being killed :-/
[16:58] <fginther> cjwatson, thanks for the clarification
[17:02] <cjwatson> I thought qemu was being killed because the autopkgtest nodes are overcommitted
[17:02] <cjwatson> IIRC anyway
[17:03] <davmor2> sil2100, jibel: meeting?
[17:06] <fginther> the plan is transition as much as possible to nova VMs, non x86 is still a bit of a question mark though
[17:07] <dobey> cjwatson: i don't know about that on the launchpad side. i'm running locally. i don't think the node being overcommitted is a problem for running one test suite locally
[17:07] <dobey> it's not like it's using all 8 of my cores or anything
[17:07] <cjwatson> dobey: (Launchpad doesn't run autopkgtests)
[17:07] <cjwatson> Indeed, though, I don't know what that would be
[17:07] <cjwatson> nova VMs> still qemu under the hood surely
[17:08] <davmor2> jibel: sil2100: we gave up on the meeting
[17:08] <dobey> well, s/launchpad/jenkins/ but not relevant to local running qemu
[17:08] <cjwatson> nova is making some progress on other arches (we need it for LP), so that's probably the right direction
[17:19] <dobey> oh well
[18:00] <imgbot> === IMAGE RTM 184 DONE (finished: 20141218-18:00) ===
[18:00] <jibel> sil2100, 184 is ready, we restart the sanity tests and the delta
[18:00] <imgbot> === changelog: http://people.canonical.com/~ogra/touch-image-stats/rtm/184.changes ===
=== alan_g is now known as alan_g|EOD
[18:25] <iahmad> imgbot, help
[18:25] <imgbot> I am the firendly system-image watchbot !
[18:25] <imgbot> I know the following commands:
[18:25] <imgbot> help, stop, status, map, stunt
[18:25] <imgbot> for questions please mail ogra@ubuntu.com
[18:25] <iahmad> imgbot, status
[18:28] <iahmad> imgbot status
[18:30] <ogra_> iahmad, status needs an image number
[18:30] <ogra_> imgbot, status 180
[18:30] <imgbot> Status: succeeded, Started: 2014-12-16 03:02:06 UTC, Finished: 2014-12-16 03:50:48 UTC
[18:30] <imgbot> Build URL: https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu-rtm/14.09/ubuntu-touch/+build/14095
[18:30] <imgbot> Changelog: http://people.canonical.com/~ogra/touch-image-stats/rtm/180.changes
[18:30] <ogra_> imgbot map 180
[18:30] <ogra_> imgbot,  map 180
[18:30] <ogra_> grr
[18:30] <ogra_> imgbot, map 180
[18:30] <imgbot> krillin rtm version: 180 maps to mako version: 151
[18:30] <imgbot> krillin rtm version: 180 maps to generic_x86 version: 145
[18:31] <iahmad> imgbot status 184
[18:31] <iahmad> ogra_, thanks
[18:31] <iahmad> imgbot,  status 184
[18:31] <ogra_> (i need to improve the pattern matching, needs the exact command (with comma and only one space)
[18:31] <iahmad> imgbot, status 184
[18:31] <imgbot> Status: succeeded, Started: 2014-12-18 16:36:36 UTC, Finished: 2014-12-18 17:27:28 UTC
[18:31] <imgbot> Build URL: https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu-rtm/14.09/ubuntu-touch/+build/14358
[18:31] <imgbot> Changelog: http://people.canonical.com/~ogra/touch-image-stats/rtm/184.changes
[18:32] <ogra_> if you want vivid results just append vivid ...
[18:32] <ogra_> imgbot, map 57
[18:32] <ogra_> err
[18:32] <imgbot> krillin rtm version: 57 maps to mako version: 136
[18:32] <imgbot> krillin rtm version: 57 maps to generic_x86 version: 130
[18:32] <ogra_> imgbot, map 57 vivid
[18:32] <imgbot> mako ubuntu version: 57 maps to krillin version: 64
[18:32] <imgbot> mako ubuntu version: 57 maps to generic_x86 version: 58
[18:32] <ogra_> imgbot, status 58 vivid
[18:32] <imgbot> Status: succeeded, Started: 2014-12-18 02:02:06 UTC, Finished: 2014-12-18 02:59:01 UTC
[18:32] <imgbot> Build URL: https://launchpad.net/~ubuntu-cdimage/+livefs/ubuntu/vivid/ubuntu-touch/+build/14289
[18:32] <imgbot> Changelog: http://people.canonical.com/~ogra/touch-image-stats/58.changes
[18:37] <oSoMoN> trainguards: can I have a silo for line 42, please?
[18:38] <robru> oSoMoN: vivid 6
[18:38] <oSoMoN> robru, thanks man!
[18:39] <robru> oSoMoN: you're welcome!
[18:58] <john-mcaleely> imgbot status 5
[18:58] <davmor2> john-mcaleely: you forgot the comma
[18:58] <john-mcaleely> imgbot, stunt
[18:58]  * imgbot rolls on its back and purrs
[18:58] <john-mcaleely> imgbot, status 5
[18:58] <imgbot> Error: There does not seem to be any build with the number 5
[18:58] <john-mcaleely> davmor2, :-)
[18:58] <john-mcaleely> imgbot, status 10
[18:58] <imgbot> Error: There does not seem to be any build with the number 10
[18:59] <john-mcaleely> :-)
[18:59] <john-mcaleely> imgbot, map 183
[18:59] <davmor2> john-mcaleely: I'm assuming it only points to -proposed too
[18:59] <imgbot> krillin rtm version: 183 maps to mako version: 153
[18:59] <imgbot> krillin rtm version: 183 maps to generic_x86 version: 147
[18:59] <john-mcaleely> davmor2, I guess so
[19:00] <john-mcaleely> cool though
[19:17] <robru> boiko: https://ci-train.ubuntu.com/job/prepare-silo/3556/console need MP, not branch
[19:18] <boiko> robru: oups, let me fix that
[19:20] <boiko> robru: fixed, sorry for that
[19:21] <robru> boiko: no problem, vivid 12
[19:36]  * Mirv back to check situation
[19:37] <Mirv> it seems the revert was done
[19:37] <Mirv> oh, and image too!
[19:38] <Mirv> thanks jibel sil2100 robru whoever involved in getting that publishing of my silo done and let's hope the image is now PERFECT! :)
[19:38] <jibel> Mirv, right revert of qtubuntu-camera + thumbnail even if the thumbnailer only was responsible for the crashes
[19:38] <robru> Mirv: don't thank me. I guess sil2100 and cjwatson handled that while I slept
[19:38] <jibel> Mirv, we are testing 184
[19:38] <Mirv> jibel: the qtubuntu-camera update was scary, so getting that back to more tested version sounds sane
[19:45] <Mirv> sil2100: qtubuntu-camera had a weird landing that sounded like revert but was really not. both of these reverts were practically speaking just "simple" reverts from the previous landings, but both had manual uploads and weird changelogs as part of the landing a few days ago
[19:48] <Mirv> sil2100: so this http://is.gd/jkvo7Z was really what landed the last time - a combination of actually a vivid sync and manual upload done to the landing PPA..
[19:49] <Mirv> sil2100: as seen from the publishing history https://launchpad.net/ubuntu-rtm/+source/qtubuntu-camera/+publishinghistory
[19:49] <dobey> anyone in here familiar with qemu and adt-run?
[19:53] <Mirv> I wish I was, I didn't get my setup working
[19:59]  * Mirv sleep, good luck
[20:25] <jibel> sil2100, bfiller, Mirv testing of 184 is almost done, we are waiting for the results of the battery drain test. No blockers found so far. I'll do a review of the bugs tomorrow morning and I think it's all good for this milestone.
[20:38] <bfiller> jibel: thanks
[20:40] <oSoMoN> trainguards: can silo 6 be published, please?
[21:10] <oSoMoN> thanks!
[21:17] <sil2100> jibel: great news then, thanks!
[21:18] <sil2100> jibel: excellent work, I think promoting tomorrow sounds feasible