/srv/irclogs.ubuntu.com/2013/10/08/#ubuntu-ci-eng.txt

loolrsalveti: w00t?00:02
loolrsalveti: did you want that to go to PPA?00:02
loolrsalveti: it seems to be setup not to00:02
rsalvetilool: well, that's what we asked to fginther00:02
loolrsalveti: I think he setups the upsteam merger00:03
rsalvetito follow the normal daily CI path and such00:03
loolnot the cu2d stuff00:03
rsalvetithat needs to happen then00:03
rsalvetimaybe something cyphermox can give us a hand?00:03
rsalvetiin case you're unable to do so00:03
loolrsalveti: the truth is that there are 2 completely separate systems00:03
loolrsalveti: lp:~phablet-team/ofono/ubuntu still the right branch?00:03
sergiusenslool, what are our grounds on the qtdeclarative grilo plugin?00:04
loolsergiusens: I don't understand the question?00:04
loolsergiusens: you mean is it part of official platform?00:04
sergiusenslool, inside the music app or seeded? and if seeded I guess it would be in sdk-libs or touch?00:05
rsalvetilool: yes00:05
loolsergiusens: hmm I remember mhall tried to push to seed it; I personally think it's fairly internal stuff, so I'd rather err on the side of caution and say to include it for now00:05
sergiusenslool, include in music app you mean?00:06
loolsergiusens: yeah00:09
loolcjwatson: hi there, reading https://wiki.ubuntu.com/DailyRelease/FAQ#Adding.2BAC8-removing_components_to_a_stack it seems I need to poke an archive admin to add a package to the allowed set coming out of cu2d00:09
loolcjwatson: the source package is ofono00:09
loolcjwatson: I've pushed it to the cupstream2distro-config branch; in fact it was there but with daily_release: False00:10
loolrsalveti: preparing ofono in PPA00:11
loolrsalveti: it didn't like the versioning there00:12
loolException: Didn't find a correct versioning in the current package: 1.12+bzr-0ubuntu100:12
rsalvetilool: do you know what would be the correct versioning in there?00:13
loolrsalveti: 1.12+13.10.20131007-0ubuntu100:14
loolrsalveti: note that I think we wont be using upstream tarballs anymore00:14
loolrsalveti: I think it generates tarballs on the fly00:15
rsalvetilool: but thought that the CI script itself would bump the version00:15
loolis that what we want here?00:15
loolrsalveti: it will, but it's checking that it's not screwing the versioning scheme00:15
rsalvetithis branch is bzr bd compatible, so it's fine to not generate the upstream tarball at every upload00:15
loolrsalveti: hmm I suspect cu2d might still generate the tarball at every upload00:15
rsalvetiright, that's not ideal but works00:16
loolrsalveti: you know I think I don't really want this to be cu2d-ized right now; too much of a time waste without didrocks00:16
loolrsalveti: let's visit this post-13.1000:16
loolcjwatson: sorry, cancel that; I think we'll do it post 13.1000:17
rsalvetilool: right, that's fine, should I upload this by hand then?00:17
rsalvetimaybe cyphermox can give us a hand as well on that00:17
rsalvetianyway, bbl00:18
loolrsalveti: are there any testsuites for these two?00:18
rsalvetifor ofono we only have unit tests00:18
rsalvetifor tofono I'm not sure00:19
loolcyphermox: Would you mind testing and landing telepathy-ofono from PPA and ofono from https://code.launchpad.net/~rsalveti/ofono/ubuntu-fixes-online-offline-netreg-mute/+merge/189204 ? it's landing #14000:20
rsalvetisure00:20
rsalvetiofono was tested already, just need to test tofono with it00:20
cyphermoxsure, hold on00:21
loolsergiusens: tested clicks, looked good!00:24
loolsergiusens: I added a slot for another round when you like00:25
cjwatsonlool: ack00:25
sergiusenslool, great, let me prepare them00:27
loolsergiusens: Hmm since it requires coordination to change the 2/3 places, maybe best to stage this for tomorrow afternoon00:36
loolsergiusens: are there 4 remaining ones that we do in one go?00:36
cyphermoxlool: building ofono right now00:40
cyphermoxrsalveti: ^00:40
loolcyphermox: actually ofono was already landing #123; doens't change anything  :-)00:58
cyphermoxwell, you mentioned both01:01
cyphermoxbrb01:01
sergiusenslool, let me just add calendar and notes?01:09
loolsergiusens: Hmm about to kick a build01:12
looljust waiting for dbus to come out01:12
sergiusenslool, ok, lets leave it for the next build01:12
sergiusensI'll bbl then and send you an email with everything01:12
loolsergiusens: do you have the -meta prepared?01:14
loolsergiusens: Or I can just drop them from meta01:14
loolsergiusens: ah that reminds me, we miss the upstart conversion job for the transition to mir01:15
sergiusenslool, https://code.launchpad.net/~sergiusens/ubuntu-seeds/notes_calendar_as_click/+merge/18974901:15
sergiusenslool, and if cjwatson still around he would need to trigger the click_copy to get the clicks in01:15
loolsergiusens: Yeah not sure01:16
cjwatsonI would if I could get to snakefruit01:16
cjwatsonAh, there01:16
cjwatsonsergiusens: http://paste.ubuntu.com/6207493/01:17
cjwatsonenoworky01:17
sergiusenscjwatson, I just ran it, worked fine :-/01:18
sergiusenscjwatson, ah, sorry...01:19
cjwatsonsame when I retry.01:19
sergiusenscjwatson, I'm noticing you are using https://jenkins.qa.ubuntu.com instead of the internal one01:19
sergiusenscjwatson, I wasn't publishing this one, one sec01:19
sergiusenscjwatson, ok, had to rebuild to get it published, can you try now?01:36
cjwatsonsergiusens: done01:37
loolsergiusens: meta is in too01:40
loolsergiusens: anything else missing for this to be picked up?01:40
sergiusenslool, nope, we just need plars to update testing01:40
cjwatsonCan I drop the proposed-migration exception for calendar-app?01:40
cjwatsonIt didn't have one for notes-app, since that's in the primary archive01:41
loolplars: Mind updating tests for notes-app and calendar-app now as clicks?01:41
loolcjwatson: yup01:41
sergiusenscjwatson, almost sure we can01:41
cjwatsonI'll wait until tomorrow morning so you can be really sure01:42
cjwatsonNo huge rush01:42
lool'night01:49
plarssergiusens, lool: back, sorry for the delay. Had some stuff I had to do tonight at home02:23
plarslool, sergiusens: notes and calendar will run as click packages now02:36
cyphermoxlool: it's late, but yeah telepathy-ofono + ofono looks fine to me02:39
cyphermoxoh, I forgot to test the enable/disable02:39
cyphermoxthat doesn't work any better than it did before, but at least I see no regression02:43
plarshmm, who's up building images now?03:31
plarsguessing it's sergiusens03:31
jibelenabling BT from system-settings is broken in #86 (regression from #82 or so)07:10
jibelbug 1236249 with SF and bug 1236705 with Mir07:10
ubot5bug 1236249 in unity8 (Ubuntu) "unity8 crashed with SIGSEGV in UnityMenuModelPrivate::itemState()" [High,Confirmed] https://launchpad.net/bugs/123624907:10
ubot5bug 1236705 in unity8 (Ubuntu) "unity8 crashed with SIGABRT in __gnu_cxx::__verbose_terminate_handler(), thrown from mir::DefaultServerConfiguration::the_connector()" [Critical,Confirmed] https://launchpad.net/bugs/123670507:11
jibeland also playing a song while one is already playing crashes qmlscene with bug 123415207:11
ubot5bug 1234152 in qtdeclarative-opensource-src (Ubuntu) "qmlscene crashed with SIGABRT in __gnu_cxx::__verbose_terminate_handler()" [High,Confirmed] https://launchpad.net/bugs/123415207:11
seb128good morning!08:05
looljibel: Thanks for the heads ups08:12
looljibel: disabling bluetooth works from indicator though?08:12
jibellool, yes, but unity crashes when you re-enable it from u-s-s08:12
seb128jibel, only from u-s-s, not from the indicator?08:18
Laneyyou can't re-enable it from the indicator08:18
Laneybecause it disappears, right?08:18
jibelseb128, you cannot re-enable from indicator08:18
seb128u-s-s does nothing else than using the indicator service though qmenumodel08:18
Laneytry calling the action over dbus08:18
seb128Laney, how did we workaround the setting switch going away when the indicator goes away? I though we didn't solve that issue yet08:18
Laneyin the battery panel, no08:19
* seb128 didn't see the mr going through to fix that08:19
Laneybut in bluetooth it always shows it08:19
seb128oh, right08:19
seb128I keep mixing controls in battery with the individual panels :p08:19
jibelseb128, bug 1233626 and bug 1134375. It didn't mark them duplicates because I don't know if the behaviour should be different on desktop and touch08:20
ubot5bug 1233626 in indicator-bluetooth (Ubuntu) "shouldn't hide itself when bluetooth is turned off" [High,New] https://launchpad.net/bugs/123362608:20
ubot5bug 1134375 in indicator-bluetooth (Ubuntu) "bluetooth indicator is missing when bluetooth antenna is off, giving the user no way to *enable* bluetooth" [High,Triaged] https://launchpad.net/bugs/113437508:20
Laneyseemed to crash unity here08:20
Laneygdbus call --session --dest  com.canonical.indicator.bluetooth --object-path /com/canonical/indicator/bluetooth --method  org.gtk.Actions.Activate bluetooth-enabled \[\] {}08:20
Laneyrun that to toggle it08:20
seb128jibel, desktop and touch are not supposed to behave the same no, desktop is not supposed to hide the icon since there is enough place, touch is more limited in space so the indicator is supposed to only be listed if bluetooth in on08:23
seb128not sure I agree with that, but that's how mpt designed it08:23
jibelseb128, okay, so that's invalidate slangasek's report because it can be re-enable from system-settings/bluetooth on the desktop08:25
LaneyI bet the one with the *emphasis* is his08:26
Laneyoh, how did I guess? :P08:26
seb128jibel, well, on the desktop it shouldn't be hidding08:27
seb128jibel, https://bugs.launchpad.net/bugs/123362608:27
ubot5Ubuntu bug 1233626 in indicator-bluetooth (Ubuntu) "shouldn't hide itself when bluetooth is turned off" [High,Triaged]08:27
jibeloh yes08:27
* jibel leaves the indicators alone and goes find more crashes08:28
loolsil2100: joining us this morning?08:30
sil2100Shit!08:33
looljibel: was the music-app crash on Mir?08:42
jibellool, yes, Mir only08:42
jibelduflu is on it08:42
sil2100My plugin crashed, uh09:01
cjwatsonlool: We have a bunch of arm64 build fixes all of which need to be done fairly urgently and many of which are in core so are on touch images.09:08
cjwatsonlool: Any chance we could have blanket approval for the ones that per common sense don't make behavioural changes?09:09
Mirv_ok I think this should be good when the unity-mir changes are in: http://pastebin.ubuntu.com/6208552/09:09
loolcjwatson: ok, I can watch #ubuntu-relesae for these uploads09:09
cjwatsonThey're mostly autotools updates and such09:09
loolcjwatson: Yes, please upload safe changes09:09
cjwatsonThanks.  We just have very limited time to get this going and I want to minimise round-trips09:09
sil2100Aw come on, again crash09:10
jibellool, bug 1236757 makes WPA dialog unusable on Mir09:15
ubot5bug 1236757 in ubuntu-system-settings (Ubuntu) "UI behind 'WPA Password' dialog receives touch events" [High,New] https://launchpad.net/bugs/123675709:15
jibelthe bug exists with SF and Mir but with Mir the dialog cannot be dismissed09:16
looljibel: that's bad09:18
loolSaviq: this sounds like another inputarea thing?   ^09:18
Saviqlool, jibel, no, it's *the same*09:20
Saviqlool, jibel, you just didn't release unity-mir yet09:21
loolSaviq: Ok; that's exactly what I wanted to confirm09:21
loolSaviq: that it would be fixed with the inputarea fix09:21
Saviqlool, yes09:21
loolSaviq: bear with us, we just came out of our landing/ci team hangout, and my day started an hour ago  :-)09:21
Saviqjibel, lol bug #123521509:22
ubot5bug 1235215 in unity-mir "[Mir]cannot answer call when system settings are in the foreground" [High,Fix committed] https://launchpad.net/bugs/123521509:22
Saviqlool, jibel, you can try packages from http://people.canonical.com/~msawicz/mir-input/ that contain the fix09:22
loolSaviq: we'll be landing all of this in the PPA ASAP09:23
Saviqlool, it's actually there in daily-build ppa already09:23
Saviqlool, got built as part of unity8 stack09:23
Saviqlool, just not published09:23
loolSaviq: perfect, that's what I was about to check09:24
jibelSaviq, thx, I'll test again when it is part of the image09:25
=== vrruiz_ is now known as rvr
=== mandel`` is now known as mandel
sergiusensmorning09:56
sergiusenslool, hey, what's the ETA for next build?09:57
=== davmor2_ is now known as davmor2
asacpsivaa: did the tests got better on latest image when retrying?10:02
psivaaasac: gallery on mako and music on maguro did get better10:03
psivaaasac: i am still working on calendar tests10:03
asacpsivaa: did any crash go away?10:03
psivaaasac: no in the still failing tests, the upstart crash is still occurring10:05
asacpsivaa: but there were tests fixed on retry that had crashes and those dont have crashes anymore?10:06
psivaaasac: no, those that got fixed in the rerun did not have crashes originally10:08
ogra_bah so webapps start now again but dont take any input10:08
ogra_(at least G+ doesnt)10:08
asacogra_: so webapps are annoying right now?10:09
psivaaasac: i could try a re-install and then re-run those tests that still produce crashes if you like10:09
asacogra_: did lool say that dbarth was working on those?10:09
psivaato see if there is any improvement10:09
ogra_asac, well, i only tested G+ and dholbachs new app10:09
asacpsivaa: right.10:09
ogra_asac, G+ doesnt scroll ... and while playing with daniels app maliit apparnetly died10:09
asacpsivaa: lets look at dashboard for a sec10:10
ogra_it takes *some* input but no scroll events and kbd input10:10
asacpsivaa: so those crashes started in 07.110:10
asacthats http://people.canonical.com/~ogra/touch-image-stats/20131007.1.changes10:10
asacogra_: whats the new upstart thing doing there? do you know?10:10
asacpsivaa: maybe try removing upstart-app-launch-tools and see if that fixes the problems10:11
ogra_theer were some bugfixes10:11
psivaaasac: i could try that10:11
ogra_oh10:11
asacpsivaa: e.g. try to reproduce crash of one tests10:11
ogra_that one10:11
asacand see if it goes away10:11
asacpsivaa: but pick a test that didnt crash before 07.1 :)10:11
psivaas10:11
psivaaasac: sure10:11
ogra_asac, upstart-app-launch-tools isnt used anywhere, its scripts that make debugging easier10:11
asacso what causes the reproducible crashes here: http://people.canonical.com/~ogra/touch-image-stats/20131007.1.changes10:12
asac?10:12
asacubuntu-sdk-libs:armhf from 1.083 to 1.085 ?10:12
ogra_that were just seed changes iirc10:13
ogra_qtdeclarative5-friends-plugin10:13
ogra_this was dropped from ubuntu-sdk-libs10:13
ogra_(ubuntu-sdk-libs is part of ubuntu-meta now, if you see ubuntu-touch updated it will be there too)10:13
ev!ci10:19
ev#ci10:19
ev^ ignore all that. Testing.10:19
psivaaogra_: is calendar app comes as click package?10:20
psivaacalendar-app-autopilot is not installed as part of it if it's under click packages10:20
ogra_psivaa, i think sergiusens changed it to one yesterday10:20
psivaaogra_: ack10:20
sergiusensyeah, you don't want to install those packages for click10:21
sergiusensthey depend on the app's deb package and would just be an invalid10:21
sergiusenstest10:21
loolsergiusens: next build coming up pretty soon10:23
loolasac: dbarth is working on getting us more testsuites, albeit not one that would test whether webapps work10:23
loolasac: both cordova and unity-webapps-qml testsuites need some love I think10:23
psivaaplars: ^ re: calendar app click package changes in the smoke tests, i used rev:45 of ubuntu-test-cases/touch with 8610:24
loolbut that's kind of low priority on the grand scheme of things10:24
sergiusenslool, hmmm, well I have a notes-app fix10:24
loolsergiusens: ok, can you show it to me?10:24
loolsergiusens: I think it should go in ASAP10:24
loolsergiusens: pulling trunk10:24
sergiusenslool, https://code.launchpad.net/~sergiusens/notes-app/app_domain/+merge/18979810:24
sergiusenslool, problem with the whole localstorage thing10:25
psivaaasac: so, i removed upstart-app-launch-tools and ran clock tests (which did not produce crash before 07.1) and it still produces the upstart crash10:25
loolsergiusens: is anything hardcoding notes-app?10:25
loolsergiusens: icon theme?  other tests?10:25
asacpsivaa: ok...10:26
loolsergiusens: does this affect other packages?10:26
asacpsivaa: so sdk tools?10:26
sergiusenslool, applicationName10:26
asacpsivaa: did you reboot?10:26
asacafter?10:26
loolsergiusens: local_location = "../../notes-app" in tests/autopilot/notes_app/tests/__init__.py?10:26
psivaaasac: yes twice :), i manually rebooted and then the test does it anyway10:26
loolsergiusens: are you trying to keep the .deb working?10:26
sergiusenslool, nah, I ran the tests yesterday, but I had the old ui toolkit10:26
sergiusenslool, yes10:27
sergiusenslool, yes for the deb10:27
sergiusenslool, MRs are deb dependent so I have to anyways ;-)10:27
sergiusenslool, upstream merger hasn't upgraded to click yet10:27
asacpsivaa: try downgrade the sdk-libs10:28
asachttp://people.canonical.com/~ogra/touch-image-stats/20131007.1.changes10:28
asacpsivaa: ubuntu-sdk-libs:armhf from 1.083 to 1.08510:28
psivaaasac: ack10:28
asacpsivaa: otherwise its 100% the click testing10:28
asacsergiusens: anmy idea?10:28
asace.g. why we see those crashes?10:28
loolsergiusens: see review from osomon there10:29
asachttp://reports.qa.ubuntu.com/smokeng/saucy/touch_ro/maguro/86:20131008:20131003.2/4608/ubuntu-clock-app-autopilot/10:29
loolsergiusens: Otherwise, good with me if the tests pass10:29
sergiusenslool, already fixed :-)10:31
sergiusenslool, I'll keep you posted10:32
loolsergiusens: including the .pot?10:32
sergiusenslool, yup10:32
loolsergiusens: ok, please prepare the updated click and ping me when it's up  :-)10:32
loolsergiusens: how can I see it's ready for image build and where's the list of clicks + versions?10:32
loolsergiusens: added to landing plan10:33
sergiusenslool, http://people.canonical.com/~ubuntu-archive/click_packages/click_list10:33
loolrsalveti: could you give a heads up here before uploading a package to archive like android?10:34
loolrsalveti: how do we test it works?10:35
sergiusenslool, the bzr revno is embedded in the manifest for the click package though10:35
loolrsalveti: I mean what's the symptom of the property watcher that is fixed?10:35
sergiusenslool, what rsalveti uploaded was a patch to property watcher, it's grabbing the full list now10:36
loolsergiusens: so you're uploading notes-app to store by some jenkins magic from trunk, then updating that click repo location from store, then manually updating the click_list?10:37
loolsergiusens: do we have to ping ~ubuntu-archive to get an updated click in there?10:37
cjwatsonOr wait for the cron jobs10:37
cjwatsonOr I can bump the frequency ...10:37
loolsergiusens: android property watcher >> I see the fix in the diff, but I dont know in practice how to tell what works; is this to watch android bionic getprop properties?10:37
loolsergiusens: what's the use case for watching them?  adb?10:38
loolcjwatson: ah there's a cron pulling latest version from store?10:38
cjwatsonYes10:38
loolcjwatson: what's the frequency?10:38
cjwatsonlool: I've made it every hour rather than every six hours now10:38
loolok10:38
sergiusenslool, I'm not doing store uploads yet; the store is single user still and I'm afraid of havoc10:38
cjwatson11 * * * *10:38
loolcjwatson: thanks10:38
cjwatsonsergiusens: I would really appreciate it if you could make click_copy.py be quiet on stdout/stderr unless something is wrong10:38
cjwatsonsergiusens: One of the reasons I had it on a lower frequency was the cronspam10:38
sergiusenslool, upstart-local-bridge; init uses the android property system10:39
loolsergiusens: so you're manually building a click from trunk and uploading to store?10:39
sergiusenscjwatson, ack; I'll change10:39
cjwatsonthanks10:39
sergiusenslool, no, built in jenkins10:39
loolok10:39
loolsergiusens: concretely, what behavior changes can I see from the upstart fix?  :-)10:39
loolcyphermox: news of ofono and telepathy-ofono?10:40
sergiusenslool, if you create a job that has a 'start on android-container svc.init.servicemanager=running' it will work on all boots instead of 9/1010:40
loolsergiusens: grep -rl svc.init.servicemanager /etc/init returns nothing; what are we using this for?10:41
sergiusenslool, that was an example10:42
loolI see /etc/init/upstart-local-bridge.conf will emit android-container10:42
sergiusenslool, that's why I said 'if you create'10:42
loolsergiusens: looking for concrete places where we use this that are affected10:42
* lool lunch &10:43
sergiusenslool, no where yet, but it would get all the sleeps out of the jobs if we wanted to10:43
asacpsivaa: did that help?10:43
asacMirv_: the landing 83 is still being worked on?10:45
asacor is it in image/archive now?10:45
psivaaplars: nope it it10:49
psivaait did not10:49
psivaaasac: it did not..10:49
psivaamy fingers are playing havoc :)10:49
loolsergiusens: Ok; that's good motivation (stronger upstart jobs, dropping the sleeps)10:50
loolsergiusens: thanks10:50
loolasac: landing #83 is done10:50
loolasac: it's still up because it's not in a promoted image, and it shows as FAILED #84 because it caused some regressions indirectly, but we want to keep it10:51
loolasac: the step we will take to fix the remaining desktop regression will be added to landing plan once ready (thumbnailer split); the other regression is fixed in #8610:52
loolimage #8610:52
psivaaasac: downgrading ubuntu-sdk-libs:armhf to 1.083 needed installation of dropped qtdeclarative5-friends-plugin10:52
psivaaasac: still it did not help10:52
asaclool: where are the regressions?10:52
asacwhere are the getting fixed?10:53
loolasac: that's what I just said10:53
asaclool: sorry10:53
asacyeah10:53
loolalso covered them this morning, but I understand of information at once10:53
asaclool: so i got feedback from app folks that they hate10:53
asacif we jsut regress them10:53
loolwho is this?10:53
asacthe community folks10:53
loolyou mean if we regress the image?10:53
asacif we regress their tests by updating sdk10:54
asacetc.10:54
loolwe haven't promoted any regressed apps though10:54
asacwe shouldn't let those things in10:54
loolyes, the failing tests have been catched before including the SDK in the image10:54
asacno... its in image10:54
asacand archive10:54
asachence their merge tests fail10:54
asacetc.10:54
looland it /just broke the tests/ not the app, and the tests were fixed /before/ we landed the SDK10:54
asacthats an invalid argument10:54
loolah the current ones in -proposed10:54
asacbut well10:54
asaclool: oh... we fixed the tests?10:54
asacthoguth they are still pending (you said above)10:55
loolasac: well, the first round of regressions from the SDK yes10:55
loolthe ones with the database pathes10:55
loolsee the bug link in the SDK landing you just asked about10:55
loolit has 8 or so merge proposals attached to fix tests10:55
asacright. we shouldnt have let that in anyway without coordinating with all stakeholders.10:55
loolbut it's not related to the failures in #8610:55
loolasac: we have not let itin10:55
asacyeah. just saying10:55
loolasac: this was in PPA only10:55
ogra_we also switched from friends1 to friends2 in the ADK10:56
ogra_*SDK10:56
ogra_which definitely broke one app out there10:56
loolasac: please, let's stop discussing this past stuff, errors were catched before landing; I got pressure the other way around already that the SDK was not getting updated fast enough blahblah10:56
sergiusenslool, those all need to be updated again because they are not using apparmor paths10:56
asaclool: ok. no problem10:56
loollet's keep calm, look at current failures and upcoming landings10:56
asacjust wanted to get the idea10:56
cjwatsonogra_: it wasn't a switch - friends2 was added *ages* ago10:56
loolgot to go to school. be back in a few10:56
cjwatson(it's now explicitly seeded, but it was present earlier anyway)10:56
ogra_ok, well, then we removed friends110:57
cjwatsonyes, which should have been done at roughly the time it stopped being built by its source10:57
cjwatsonthis was tech debt10:57
ogra_(which was scheduled and all ... and discussed witrh the app owner)10:57
cjwatsonright, we didn't cowboy that10:57
ogra_i.e. the process was all fine just pointing out that we did such a change that broke backwards compatibility10:58
ogra_and i dont think we can avoid such stuff before v1.010:58
asacsergiusens: in case you missed: any idea why we see those crashes in app tests after moving to click?10:58
ogra_app devs will have to adapt until we have a really stable API10:58
sergiusensasac, yeah, it's supposed to be fixed in latest upstart-app-launch10:58
sergiusenszeitgeist crashes10:59
ogra_sergiusens, yeah10:59
asacsergiusens: is that already in http://people.canonical.com/~platform/cu2d/results ?10:59
sergiusensasac, I don't know, I'm not following all projects10:59
sergiusensasac, ...  lool told me not to worry10:59
asaclool: is it in there?10:59
asacpsivaa: you could try upgrading that11:00
asacpsivaa: upstart-app-launch (0.1+13.10.20131008-0ubuntu1) saucy; urgency=low11:00
asacpsivaa: thats in daily-build ppa11:00
psivaaasac: ack, will do11:00
ogra_doesnt that also need unity-mir and unity8 updates ?11:00
asacogra_: maybe... there is a new unity-miro as well11:01
asacogra_: i care about the crashes11:01
ogra_i know11:01
ogra_jzust saying we held it back because someone said its a set that needs to land together11:01
ogra_so if you just install upstart-app-launch, yu most likely trigger new issues11:02
Mirv_asac: satoris is fixing the thumbnailer split11:04
sergiusenscjwatson, lp:click-sync -> http://bazaar.launchpad.net/~ubuntu-cdimage/click-sync/trunk/view/head:/click-sync.py11:04
asacMirv_: can you add a comment about the status of UI toolkit landing (the one that failed)?11:04
asaclanding 8311:04
asace.g. what issues are left on latest proposed image11:04
sergiusenscjwatson, same syntax and silent argv[0] is different but feel free to rename if needed11:05
cjwatsonsergiusens: Great, thanks, that's fine11:07
Mirv_asac: it's there in the column I, the thumbnailer plugin split. nothing to be fixed AFAIK on the image, but on the desktop side only.11:07
cjwatsonsergiusens: switched11:07
Mirv_so satoris has now a MP to thumbnailer that according to bzoltan fixes the desktop side issue11:09
Mirv_I asked satoris to join here as well to coordinate on getting the fixed thumbnailer released11:09
Mirv_so it will not need SDK update, only thumbnailer from the services stack11:10
asacsergiusens: how can i find which apps are now shipped as click?11:10
sergiusensasac, http://people.canonical.com/~ubuntu-archive/click_packages/click_list11:10
asacsergiusens: how can i see that on the device? i assume i can use some click cli tools?11:11
ogra_sergiusens, do you think we'll manage to get all core apps to click (i.e. can we drop the PPA before release)11:11
sergiusensasac, sudo -u phablet -i; click list11:11
ogra_clikc list ?11:11
sergiusensclick list --manifest for details11:12
sergiusensogra_, yeah, we will; going to happen today I guess11:12
satorisMirv_: here I am.11:12
asacsergiusens: how do i see which ones are installed by default and which i installed on my own?11:12
Mirv_satoris: so, we (lool, sil2100) mainly need to know when it's merged so that we can relaunch services stack build to have a package to test and then promote to archives11:12
asacwait11:12
asaclet me try L:)11:12
* ogra_ hugs sergiusens 11:13
sergiusensasac, it will be obvious when using --manifest11:13
satorisMirv_: the MR is here and I can approve it any time you choose: https://code.launchpad.net/~jpakkane/thumbnailer/gst-disable/+merge/18981811:13
satorisOnce Jenkins gives it the go-ahead.11:14
Mirv_lool: how far did you understand the issue and does that ^ seem good to merge for you?11:14
Mirv_satoris: so what again were the implications of that on the device?11:16
satorisMirv_: none, really. Embedded album art in MP3 files stops working.11:17
Mirv_satoris: right, just read it from the bug report. that sounds good to me at least, so unless anyone here objects go ahead and approve it when jenkins is happy.11:19
satorisThat is, it won't be shown in the dash.11:19
Mirv_("disable embedded album art provider in thumbnailer, because it is the only part that uses GStreamer and is in the same process (video is done out-of-process)")11:19
satorisThis should not require rebuilding anything as API and ABI are preserved.11:20
satorisMirv_: Jenkins approved, bzoltan approved, top approving now.11:23
Mirv_I also tested it on desktop, and like bzoltan seems fixed11:23
satorisApproved.11:24
satorisMirv_: so you could reproduce it on the desktop? Why is my machine the only one where the issue does not happen? :)11:24
cjwatsonasac: click list --manifest shows the directory where they're installed (the _directory key) - the ones you installed on your own are under /opt/click.ubuntu.com, the others are elsewhere11:25
cjwatsonSo if you need to know this for some reason then you can work it out that way11:25
Mirv_satoris: so you tried the Jamie's approach? I don't know.. it hanged without your branch, and launched + worked after installing the thumbnailer from your branch11:25
Mirv_satoris: one option is that if you've played/installed the qtmultimedia-touch packages at some point to your desktop11:26
satorisI most definitely have not.11:27
satorisIt's not installed according to apt.11:27
satorisOh well, the dynamic linker works in mysterious ways.11:29
Mirv_lool: added satoris' thumbnailer change to the landing asks page11:29
loolMirv_: looking11:29
loolsatoris: hey I thought the plan was to split thumbnailer entirely out of SDK11:30
loolsatoris: or at least in its own binary package11:30
satorislool: we fixed the core issue so it is not necessary.11:30
Mirv_I think the talk on sdk channel went something like this is the less risky way than splitting at this point of time11:30
loolsatoris: so from the look of the changes, you've disabled a class of use cases11:31
satorislool: yes, but that is not a regression as it was all new functionality. :)11:31
loolsatoris: a) what are we missing functionnaly with the commented out changes  b) was this tested on desktop and on touch?11:31
=== alan_g is now known as alan_g|lunch
loolsatoris: well, IIUC unity scopes wanted to leverate this?11:32
loolsatoris: if it's new funcitonality that we don't need we should drop it though  :-)11:32
satorisa) mirv_'s comment about 13 minutes ago b) yes.11:32
asaccjwatson: yeah. looks good. thx11:33
loolsatoris: so MP3 album art *stored in ID3vN tags* is the only thing we'd be missing?11:33
satorislool: yes.11:33
loolsatoris: but it still seems to pull in gstreamer, how can that work if there were symbol clashes?11:33
satorislool: that is for video thumbnailing which is done out-of-process.11:33
satorisA standalone binary, basically.11:33
sil2100Mirv_: https://code.launchpad.net/~sil2100/cupstream2distro-config/extra_pkgs_for_phone_again/+merge/189820 <- hello ;)11:34
loolsatoris: so you're saying gstreamer 0.10/1.0 mixup is dropped by this change; there is still a plugin using gstreamer, but it's not used by default and the processes using it only run on touch and work?11:34
satorislool: basically yes, except that the binary is not optional. The reason the clash happened is because libthumbnailer.so linked against GStreamer 1.0. Now it does not anymore. Only the standalone executable does.11:35
loolsatoris: if that's indeed the case, then it seems a better approach indeed11:35
loolsatoris: what's the standalone executable?11:36
satorislool: vs-thumb, a binary that creates a snapshot image of a video file.11:36
loolsatoris: Ok11:36
loolsatoris: yeah, plans look good to me11:36
loolsatoris: so the album art thing was in "imdump"?11:36
loolMirv_: Ok, let's test and land this11:36
satorisOnce desktop moves to GStreamer 1.0 we can re-enable all this.11:36
loolpsivaa: did you manage to test updated unity-mir in PPA?11:37
satorislool: no, those are helper binaries that are not installed. Only vs-thumb is.11:37
Mirv_lool: ok11:37
psivaaasac: so, installing upstart-app-launch (0.1+13.10.20131008.1-0ubuntu1) from daily ppa did make the crash go away11:37
ogra_lool, asac there are two harmless emulator changes we need, i just added them to landing asks line 143 ... or to put them on the landing plan ?11:37
Mirv_sil2100: I actually had that already deployed, but I kind of thought the issue would go away once UI Toolkit hit the archives11:37
ogra_(zero risk xnox is just doing regression testing)11:38
loologra_: I kind of feel like putting emulator on backburner11:38
loolit's noise11:38
loolI've pushed back on the customization stuff for post-release11:38
psivaalool: i only installed the ustart-app-launch.. just to see if it makes the crash go away..i'll install the rest of unity-mir packages to test11:38
loolrealistically, we're not getting any benefit from emulator at this point11:38
loolso let's focus on release on phones11:38
ogra_lool, well, that means we cant change our landing process after release ... i would pretty much like to see all emulator stuff we can land before release11:39
sil2100Mirv_: hmm, not sure, since is that installed in the image by default?11:39
ogra_the emulator is essential for scalable testing11:39
asacpsivaa: interesting11:39
asacpsivaa: how does the phone look like with just that?11:39
asaccan you use it?11:39
Mirv_sil2100: it should be now, since it comes from qtdeclarative5-ubuntu-ui-toolkit-plugin, and it's now in archives11:39
loologra_: we can always update it after release11:39
loologra_: development wont stop at 13.1011:39
Mirv_sil2100: but was it only desktop image that counts?11:40
sil2100Mirv_: sadly, qtdeclarative5-ubuntu-ui-toolkit-plugin is not in the iso manifest11:40
loologra_: what has to go in 13.10 are fixes for real phones11:40
sil2100Mirv_: for daily-build we use desktop images11:40
Mirv_sil2100: right, so because this is touch only.11:40
psivaaasac: i was using the phone in the lab.. i'll try with rest of the unity-mir updates locally11:40
ogra_xnox, sounds like this is a no :(11:40
sil2100Mirv_: so we anyway need this installed in the extra packages - it won't get away otherwise11:40
asacpsivaa: i am intersted in just app-launch without unity-mir11:40
asacbut nevermind11:40
asaci can do it here too11:40
sil2100Mirv_: since cu2d will try fetching it as a dep otherwise, no?11:40
psivaaasac: i'll try without mir first11:41
ogra_lool, do you have any status for the kernel uploads ? i havent seen a trace of them11:41
sil2100Mirv_: so if you can, please approve so that we have it in cu2d from now on ;) I learned the hard way with autopilot that redeploying things without merging it into VCS later just means trouble ;p11:41
* ogra_ wants the ureadahead patches, there is still enough userspace stuff to do to use them 11:42
xnoxogra_: is that no for both, or only initramfs-tools-ubuntu-touch?11:42
Mirv_sil2100: yes, approved11:42
loologra_: no I was told tuesday or so; it would actually be super helpful if you could check where these stands11:42
loologra_: perhaps ping ogasawara or apw?11:42
ogra_yep11:43
Mirv_sil2100: and, since it's needed please check https://code.launchpad.net/~timo-jyrinki/cupstream2distro-config/feedback_thumbnailer_apps_media_sdk_unity8/+merge/189823 as well, those are also already deployed but would get undeployed if it's not merged11:43
xnoxogra_: lool: please comment as such on merge proposals please: https://code.launchpad.net/~xnox/ubuntu/saucy/initramfs-tools-ubuntu-touch/missing-mkdir/+merge/189789 and https://code.launchpad.net/~xnox/ubuntu/saucy/lxc-android-config/add-generic-rules/+merge/18979211:43
ogra_xnox, lool i think the udev rules for emulator are absolutely harmless ... imho they should go in11:43
ogra_but not my call, lool is the gatekeeper :)11:44
apwogra_, ?11:44
apw(as in you needed something)11:44
xnoxogra_: lool: i'm not blocked, i'll just have to hack up locally to modify stuff post-build.11:44
ogra_apw, waiting for aa updates and ureadahead fixes for out touch kernels11:45
loolxnox: are these changes the last ones to get a working emulator?11:45
ogra_s/out/our/11:45
sil2100Mirv_: ACK11:45
xnoxlool: no idea, but it does make a bootable emulator & lxc-android container out of the box.11:45
apwogra_, yep, the kernel parts are pending on the apparmor userspace changes going in first, though there is a FAIL marked in the landing asks page, which i am waiting on jds to wakeup to get clarification on11:45
ogra_lool, ^^^11:46
loolxnox: did you test these on mako or maguro?11:46
xnoxlool: still no, GLES yet nor video output. So I think there will be more changes. No idea if that would be in android package or elsewhere.11:46
* ogra_ wasnt aware that waits on userspace changes 11:46
xnoxlool: i only have grouper.11:46
loolxnox: ok, if we're not getting video output, I think we should defer emulator related landings to post 13.1011:46
xnoxlool: ok.11:46
loolxnox: will update mps11:46
apwogra_, when the kernel goes in i believe it makes things work, which needs the changes for the profiles etc in first else your networking stops at a minimum11:47
apwogra_, but with the mark against it it ain't going nowhere anyhow11:47
ogra_ah k11:47
apwi am expecting the mind with the info imminently11:47
ogra_yeah11:47
ogra_slackish USians ... always sleep into the afternoon11:48
apwi know just rude isn't it11:48
ogra_yeah, they should take the asians as example11:48
xnoxapw: oh, i didn't realise that it's tangeled up like that.11:48
ogra_they are always up early !11:48
apwxnox, it is a tangle and no mistake11:48
asaclool: is there something to test/publish on the input handling front?11:49
loolapw: we've uploaded some apparmor-easyprof-ubuntu changes yesterday; is this what we needed?11:49
* asac reflashes latest proposed11:49
loolasac: you want to upgrade to latest unity-mir + upstart-app-launch from PPA11:49
loolI'd like to start an image build ASAP11:49
loolchecking what's in flight11:49
asaclool: would we copy them to archive?11:49
loolasac: not before we get Mir11:49
asaclool: ok. plesae update spreadsheet11:49
apwlool, my crib sheet says that first, then apparmor, then the kernels, but i may have it wrong but i need the same person to clarify that :)11:49
loolasac: please update spreadsheet for what?11:50
asaclool: ok, then i wait for mir before spending time testing it11:50
asaclool: with wahts inflight etc.11:50
asacwhats11:50
loolasac: please test them *on mir*  :-)11:50
asac13:49 < lool> checking what's in flight11:50
asaclool: ok, will just wait for now. please ping me when the mir that we want to publish together with this is avail.11:51
loolpsivaa, asac: Also looking for someone to run either webbrowser-app or unity8 AP on Mir with latest unity-mir + upstart-app-launch11:51
asaclool: wait. i thought we wait for everything that we believe we need and then test it manually11:52
asacdo you feel that webbrowser-app is alrady fixed with the above?11:52
loolasac: that's what I want to know11:53
loolasac: we have 1/3 merged, 1/3 approved, 1/3 needs a commit message fix11:54
asacwe hsouldnt test if we have no indication that we fixed something :)11:54
asacif you know we fixed something, we can confirm i guess11:55
asacanyway, psivaa can probably do it as he has his device set up anyway11:55
psivaaasac: lool: doing that now11:56
asaclool: so what do we expect?11:56
asachttp://reports.qa.ubuntu.com/smokeng/saucy/touch_mir/maguro/86:20131008:20131003.2/4610/webbrowser-app-autopilot/11:57
asacthat one failed a lot :)11:57
asacwill it fix the hud crash?11:57
asacmaliit surely isnt fixed without the input handling fix?11:57
loolasac: hud updates underway12:06
asaclool: so what about the phone stack12:06
loolMirv_: BTW I added a landing slot for SDK update with the thumbnailer gst fix12:07
asacthere seem to be a few changes that might be ready to take ?12:07
loolasac: you mean from cu2d/results?12:07
asaclool: right. checking what is staged/accumulated12:07
loolasac: ofono and telepathy-ofono have to be landed12:07
cjwatsonI keep reading "hud" as "hurd"12:07
asaclool: we only have ofono :)12:07
asacerr12:07
loolasac: the cu2d/results page is way too long, too much noise; I kept taking stuff from there in last days, but I think it's time to stop12:07
cjwatsontrying to work out why you care :)12:07
asacwe dont have ofono, just the telepathy part12:08
loolasac: no we also have ofono, but it's not ci-ed12:08
loolI almost did that yesterday, but was too painful, will be post-13.1012:08
asaclool: well, we must continue to be fair to all stakeholderes12:08
loolasac: cyphermox has 2 landing asks12:08
loolasac: I'm considering all landing asks fairly12:08
asacthose guys have potential good stuff staged, so lets check with them12:08
asacright12:08
loolasac: I am stopping considering stuff getting into trunk as being needed for 13.10, unless I'm asked about it12:08
asaclool: so we dont have a phone landing ask?12:09
loolasac: outside of the ofono and telepathy-ofono ones, no12:09
asaclool: i think its worthwhile checking with everyone who has something staged given that potentially might stop landing/testing stuff unrelated to mir for a few days12:10
asacat least as long as we dont have the bits for mir together12:10
Mirv_lool: ok.12:10
loolasac: this is what we already did since friday12:12
asaclool: well, we still dont have everything together, so... :)12:12
loolasac: we looked at all cu2d/results and took a lot everyday last week, and aggressively friday and monday12:12
loolasac: it never stops12:12
asacanyway. i will deal with it12:12
loolasac: there are always more things landing12:12
asaclool: sure. thats CI :) ... it never stops12:12
loolasac: I've reviewed it before going to bed yesterday, and picked the indicator-power fix and history-service fix12:12
Mirv_sil2100: lool: FYI DHCP issue again at cu2d, aborting qa stack check job because of that12:13
loolasac: well I think it's the time to stop taking everything though12:13
asaclool: lets just keep going with business as usual12:13
psivaaasac: lool: unity8 crashes when mir enabled after installing the packages from daily ppa12:13
asacwe can take mir when its there12:13
loolMirv_: did you manage to fix?  maybe ping retoaded or IS12:13
asacfolks are still working on stuff etc.12:13
asacno panick is the no. 1 rule12:13
psivaaasac: lool: making the device not usable when mir enabled12:14
loolasac: yes, no panic, and not trying to land everything we can; it's time to be more conservative to allow Mir to get in and to focus on that12:14
asacpsivaa: thanks12:14
loolasac: we can fix most important bugs after Thursday12:14
psivaathat's with 86 + unity daily ppa12:14
asaclool: well, we dont need to stop and be paralysed12:14
asacthey are still working on things12:14
loolasac: we're not stopping, see the landing plan12:14
asacwhy would we stop the rest :)12:14
asacyeah.12:14
loolasac: we have enough landing asks, and we should give priority to that12:15
Mirv_lool: we workaround when spotting, it should have been fixed several times but still pops up every now and then. one notices it by checking why it seems a stack is stalled in autopilot.12:15
loolif I'm twiddling thumbs, I'll ping you to get some work  ;-)12:15
sil2100Mirv_: damn...12:15
asaclool: do we have all failures in current image understood?12:15
asacwill they be fixed?12:15
asaci felt we stopped doing that12:15
asacon "more conservative" ... we should be as conservative as always12:15
asachave decent rules and continue to apply them12:15
loolasac: hold on, weren't you just arguing for landing more crack that you weren't even requested to take?  ;-)12:15
asaclool: right. just want to say that we should try to get a good/green image12:16
loolyes12:16
looland I also want Mir12:16
asacand ignore mir until tvoss comes along saying they are ready12:16
asacyeah, but our hope will not make that happen sooner :)12:16
loolasac: did you look at the 3 Mir MPs?12:17
loolasac: have a look12:17
asaclool: yeah. but tvoss is still working on webbrowser test fixes12:18
asacetc.12:18
loolasac: I've invited you to 2 Mir status standups of 10mn until Thursday that I've setup with kgunn, saviq and dandrea12:18
asacso it will take a bit12:18
asaclool: so lets wrap up a good image and see where we are at then12:19
loolasac: webbrowser is not critical to switch to Mir12:19
asacwhat do we need to take to get a green image?12:19
loolasac: that was the decision from yesterday12:19
Saviqlool, just one in my calendar12:20
asaclool: right. we can look at it later12:20
asaclool: i dont have all the bits to test yet in the ppa12:20
loolSaviq: kgunn added you to them; he might not have picked the early one12:20
asacso we can just work as if it doesnt exist12:20
loolSaviq: you want the one in 30mn too?12:20
loolSaviq: I see youon both actually12:20
loolSaviq: one optional and one optional where you said no12:21
Saviqlool, right, interesting12:21
Saviqlool, ah now I see it - just today, though12:21
asaclool: so going back to start... what do we need before we kick a new image? :)12:21
Saviqlool, the other one... yeah, it's 10pm here when that happens ;)12:21
loolSaviq: correct12:22
loolSaviq: yeah, sleeping or relaxing might be a better option at that time12:22
loolsil2100: do we have history-service?12:23
=== Mirv_ is now known as Mirv
loolsergiusens: did you upload a new notes-app click?12:24
=== alan_g|lunch is now known as alan_g
loolcyphermox: around?12:25
loolcyphermox: any ETA for upload of ofono / telepathy-ofono?12:25
sergiusenslool, it just merged 6 minutes ago...12:25
sergiusensone sec12:25
loolsergiusens: I know, but I know you're fast  ;-)12:25
sergiusenslool, jenkins is slow on MRs ;-)12:26
sergiusenslool, you still get the flaky tests every now and then12:26
loolasac: so I dont think we're getting the upstart fix this time around12:26
sergiusenscjwatson, how frequent is the cron now?12:26
loolasac: after retries, the pass rate is pretty good IMO12:26
cjwatsonsergiusens: 11 * * * *12:26
loolasac: getting the notes-app AP fixes12:26
loolsergiusens: yeah, as before12:27
sergiusenslool, it's not really AP fixes, it's apparmor fixes with localstorage12:27
cjwatsonsergiusens: In fact I've made it 11,41 * * * * now that the cronspam's gone12:27
loolsergiusens: I mean, it will give us a higher AP tests passrate, yes?12:27
sergiusenslool, yes, notes should load ;-)12:28
sergiusenslool, plars doanac  would be good to get syslog or at least output of grep DEN /var/log/syslog in the collected results12:30
loolsergiusens: you have dmesg I think12:30
sergiusenslool, denies don't show up there though12:30
loolsergiusens: albeit I see nothing there https://jenkins.qa.ubuntu.com/job/saucy-touch_ro-mako-smoke-notes-app-autopilot/155/artifact/clientlogs/dmesg.log/*view*/12:30
loolsergiusens: they do12:30
loolit might be a dmesg before the test?12:31
jdstrandsergiusens: I just got here-- can you tell me what the apparmor fixes with localstorage is?12:31
jdstrands/is/are/12:31
sergiusensjdstrand, applicationName12:31
looljdstrand: https://code.launchpad.net/~sergiusens/notes-app/app_domain/+merge/18979812:31
jdstrandsergiusens: right-- that was a change in the sdk. applicationName must be set correctly. apparmor policy has allowed that for ages12:31
sergiusensjdstrand, yeah, but since the aplicationName was 'notes-app' instead of 'com.ubuntu.notes' now that the bug is fixed, it was trying to create in the wrong location12:32
sergiusenss/but//12:32
jdstrandsergiusens: right-- that actually should have broken when the sdk changed12:32
sergiusensjdstrand, first time we retest this since notes-app was built as click 2 weeks ago12:33
jdstrandah, but not for localstorage, right12:33
sergiusensbit rotting apps ftw :-)12:33
jdstrandcause that only just changed in landing #8312:33
asaclool: so what do you suggest?12:35
jdstrandsergiusens: so, all the click apps should have their applicationName corrected, not just the ones with localstorage, otherwise there is a hidden bug if they are updated to do something that needs applicationName properly set12:35
asaclool: wait for the notes-app fix before respin? what exactly will not be fixed due to upstart missing?12:35
sergiusensjdstrand, I'm changing all of them12:35
loolasac: upstart mainly affects mir12:35
jdstrandsergiusens: awesome, sounds like I didn't need to join the conversation :)12:35
loolasac: SF is completely bearable with this issue12:35
jdstrandsergiusens: thanks for the update12:36
asaclool: i am talking about AP results12:36
loolasac: I think current image is pretty good except for notes-app which will be fixed here; also getting a wagon of fixes into one12:36
sergiusensjdstrand, also see that 'friends' and 'usermetrics' are in, so now I can update some jobs12:36
jdstrandyep :)12:36
loolasac: then I think we want to land just after an image build: upstart fix, test input fixes, and consider landing Mir12:36
loolwhich would be in 2 hours12:36
asaclool: so we wait for notes-app fix and then we have a green image?12:36
loolwhen the MP with missing commit message is fixed, and after checking with mir folks, and ourselves playing with it12:36
loolasac: it wont be 100% green, but yeah12:36
asaclool: will our tests regress with just upstart and test input fix?12:36
jdstrandusermetrics in particular is a fun one. more apps shoudl use it-- I think the infographic is cool :)12:37
sergiusensjdstrand, only side effect of this applicationName thing is the translations need to follow it too12:37
loolasac: upstart wont regress anything; it's always been broken on mir and never an issue on SF12:37
asaclool: i am looking for results like 8412:37
loolasac: input will only affect mir and will progress it, but it depends on u-a-l which will regress user experience on SF (perhaps not tests)12:37
asaclool: when you say we wait for MIR, that implies the merge requests that got approved in the bug?12:37
asaclool: u-a-l?12:38
asacah12:38
loolasac: that the last of the 3 merge requests goes in12:38
asacis there a way we can test u-a-l12:38
jdstrandsergiusens: oh? I was unaware of that-- popey and I sent out an email to app developers on how to update their apps for this, but didn't mention translations. can you describe the problem?12:38
loolasac: yes, that's what I asked this morning12:38
looltest on mir with updated upstart-app-launch + unity-mir from PPA12:38
loolunity-mir will soon have another of the 3 mp in PPA12:38
jdstrandhttp://bazaar.launchpad.net/~sergiusens/notes-app/app_domain/revision/20712:38
asaci mean... how can we test the regression on SF12:39
jdstrandsergiusens: ^ that's only for debs though, right?12:39
asacfrom what i undersatnd you propose to land the bits, but not the default switch12:39
sil2100lool: I tested it, all ok, but was waiting for the phone stack to finish12:39
asac(in 88)12:39
sil2100But it's running since 2 hours still12:39
sergiusensjdstrand, https://bugs.launchpad.net/ubuntu-ui-toolkit/+bug/123307112:39
ubot5Ubuntu bug 1233071 in Ubuntu UI Toolkit "Set up the translation domain automatically for click packages" [Undecided,New]12:39
sil2100lool: so I guess I'll publish what is ready12:39
jdstrandno, I guess not, even in the click it should be renamed12:39
asaclool: anyway, lets wrap up the 87 image12:40
jdstrandCLICK_DIR12:40
asaclool: would love to have the crashes fixed though too12:40
jdstranduh oh12:40
asacbut we cant land u-a-l without the rest?12:40
sergiusensjdstrand, from what I think happens, gettex domain is taken from applicationName12:40
* jdstrand looks closer12:40
loolsil2100: is phone stack stuck?12:41
loolcause I updated it yesterday twice12:41
looladded ofono and removed it12:41
sil2100lool: it was waiting on other stacks to finish, but due to the DNS issues it was stuck longer...12:41
loolasac: fixing the crashes is just cosmetic and work intense12:41
asacwell12:41
loolasac: the fix is in upstart-app-launch and it has other changes in trunk that will regress SF12:41
jdstrandok, the patch use a relative path, that seems ok12:41
loolasac: it was confirmed that trunk fixes it already12:41
asacso we can land it on SF12:42
asacwhat is the impact?12:42
loolasac: it does away with the stopping and starting of apps with an URL; because it has no API to resume them + bring them to front, it will fail sending the new URL to e.g. music-app or webbrowser-app or mediaplayer-app12:43
loolso pressing files from scopes will only work once12:43
asaclool: and the crashes we got because of us moving to click, right?12:43
asacthats understood and clear?12:43
loolasac: I dont think all were chased no12:44
asacnote that crashes are not really cosmetic from CI perspective. they create work all the time as AP tests become flaki when crashing etc... so the time saved here is offset at some point by us retrying and praying :)12:44
loolI think the notes-app ones will go away12:45
loolwe have a calendar and a clock regression12:46
looland the usual weather-app flakyness12:46
looloverall, it's good12:46
loolit's just notes-app and crashes polluting this run12:46
asacare calendar and clock understood and worked on?12:46
loolno12:46
asacare both core apps?12:46
loolsorry not calendar12:46
loolcalculator12:47
loolasac: I believe so12:47
loolballoons: would you think you could take a look at the latest clock + calculator AP issues?12:47
loolhistory-service is in proposed12:50
Mirvsil2100: lool: still encountering problems at the 'check' (autopilot) phase in cu2d. I've pinged on the #qa channel on the issues, but if you want stuff to build you may need to manually go to the 'check' job and press the red 'x' to cancel those to allow next stack(s) to build if the problem persists12:51
fginthermorning12:51
sil2100Mirv: thanks for the info12:52
Mirvyeah so since I won't here be for that long, so that you know what to do12:53
Mirvmorning fginther12:53
fgintherMirv, o/12:53
plarsgood morning13:02
plarspsivaa: what's up with calendar app click packages? I didn't get the context from that13:03
psivaaplars: calendar-app-autopilot is not installed after the latest changes, so the tests failed to execute13:04
plarspsivaa: yes, that was intentional13:04
plarspsivaa: you saw the merge proposal last night?13:04
plars<lool> plars: Mind updating tests for notes-app and calendar-app now as clicks?13:05
psivaaplars: yea and according to sergiusens the calendar-app-autopilot is not part of click13:05
plarspsivaa: ^13:05
plarspsivaa: oh, from what I was told last night, calendar and notes were going to click in the next build and needed to change13:05
loologra_: would you be able to prepare a change to set CPU governor to performance as to fix mir display slow downs?13:06
psivaaplars: calendar could be part of click but may be the autopilot part is not13:06
plarspsivaa: sorry, the latest image came in close to midnight for me, and I fell asleep before the tests got far enough to confirm13:06
plarssergiusens, lool: what's up with that? ^13:07
sergiusensplars, they are click13:07
ogra_lool, we can force it after the container is up i suppose ... but thats indeed highgly device specific13:07
sergiusenslool, the fix is in (for domain standardpaths for notes app)13:08
sergiusensplars, I think psivaa is confused on how tests are provisioned13:08
ogra_lool, original governor setup comes from android ... and in case of mako i think performance is broken due to the android "interactive" implementation ...13:08
loolsergiusens: cool13:09
loolsergiusens: click updated in repo?13:09
plarspsivaa: sorry, I think you were out of the loop on this a bit13:09
plarspsivaa: it looks like they passed in the latest run though - you didn't do anything to re-enable the -autopilot package did you?13:09
psivaaplars: i used rev 45 of u-test-cases/touch to make it pass13:10
loolplars: could you update tests for the new clicks?13:10
loolplars: clock and calculator13:10
ogra_lool, for which arches would that be, only maguro ? (i dont realyl see slowdowns on mako)13:10
psivaaplars: not the latest (46) where you've disabled for click changes13:10
plarslool: clock and calculator were already added yesterday13:10
loologra_: all arches13:10
plarslool: they were in the first batch13:10
sergiusenslool, yes13:10
loologra_: it slows down after a while, when powerd gets into some lower sleep state apparently13:10
plarspsivaa: err, so you re-pushed the old revision?13:10
ogra_lool, they will cause probs on grouper and mako13:10
loologra_: https://bugs.launchpad.net/ubuntu/+source/mir/+bug/123519013:11
ubot5Ubuntu bug 1235190 in mir (Ubuntu Saucy) "[mako] Unity8 on Mir got slow" [High,Confirmed]13:11
plarspsivaa: let me go look at what was failing, we need it to *not* install that package13:11
ogra_lool, both use interactive governors13:11
psivaaplars: no, i edited the config of the jobs temporarily.. did not make a MP since not really sure of the history13:11
plarspsivaa: ok13:11
plarslool: be aware, I made the changes for notes and calendar last night but those probably didn't run as is in the current results. i.e. they ran with the autopilot package (and thus the deb for notes/calendar rather than the click)13:12
plarspsivaa: did you revert that change already?13:14
plarspsivaa: I don't see where you made it13:14
psivaaplars: yea i reverted it in the config13:14
plarspsivaa: ok, thanks13:14
psivaaplars: http://10.97.0.1:8080/user/psivaa/my-views/view/Maguro-ro/job/saucy-touch_ro-maguro-smoke-calendar-app-autopilot/125/console you could see i used 4513:15
plarspsivaa: just a brief rundown on what's changed...13:18
plarspsivaa: for apps that are now getting installed as click packages rather than deb, we want to remove the dependency on installing the -autopilot deb package because it would depend on the deb for the app13:18
plarspsivaa: the ci scripts now run phablet-click-test-setup, which is something sergiusens set up to look at the manifest data for all installed click packages and pull the bzr branch for the version we have installed13:19
plarspsivaa: then we run the autopilot tests and have them point PYTHONPATH to /home/phablet/autopilot to ensure they can find the tests for that app13:20
psivaaplars: ohh ok, i did not see the second bit13:20
plarspsivaa: so it looks like the tests actually ran ok for calculator, or tried to at least, but there was some sort of output processing issue once it got to the yaml. I'm installing locally so I can debug it easier13:21
ogra_lool, asac, you are aware that switching to performance will kill batteries en masse ?13:21
loolpsivaa: hey13:22
psivaaplars: from what i could see, the tests did not run13:22
loolpsivaa: so we're reinstating the manual mir testing spreadsheet13:22
loolpsivaa: we've also landed updated unity-mir + upstart-app-launch + qtubuntu13:23
loolpsivaa: could you try updating qtubuntu-android libupstart-app-launch1 upstart-app-launch upstart-app-launch-tools libunity-mir113:23
looland run tests there?13:23
psivaalool: ack will do13:23
loolpsivaa: I've added a template for current revisions of packages13:25
psivaalool: is that on image 83?13:28
psivaalool: current devel-proposed is 86 so a little confused13:29
loolpsivaa: on devel-proposed13:30
loolpsivaa: sorry my bad13:30
loolpsivaa: fixing13:30
psivaalool: nw, i already have 86 :)13:31
plarspsivaa: ahh, I see the problem13:35
plarssergiusens, psivaa: calendar-app-autopilot apparently depends on python-dateutil13:35
sergiusensplars, ok, let me look at that13:36
plarssergiusens, psivaa: we can work around it by specifying that the test needs to install python-dateutil as a prerequisite if needs be13:36
psivaaplars: ahh ok, that probably explain line 20 error.. sorry i was kind of in a hurry to make it pass :)13:37
loolok, about to build an image13:37
loolhistory-service is in13:37
lool== Building #87 ==13:38
plarspsivaa: the error gets sort of hidden because we use autopilot list to generate the list of tests to run, so it was trying to run one with that in the name13:38
sergiusensplars, please do, and I'll add that and python-mock to phablet-click-test-setup13:38
plarspsivaa: so the easy way to see it is to run the autopilot list by hand13:38
psivaaplars: ack13:40
plarspsivaa, doanac, cjohnston, josepht: could one of you ack https://code.launchpad.net/~pwlars/ubuntu-test-cases/ubuntu-system-settings-online-accounts-autopilot/+merge/189753 for me, I'd like to get it in before the next build lands13:42
josephtplars: looking13:42
loolasac: binaries: qtubuntu-android libupstart-app-launch1 upstart-app-launch upstart-app-launch-tools libunity-mir113:42
loolpsivaa: ^13:42
loolnot sure where qthybris goes13:42
psivaalool: running unity8 tests with the above on 8613:44
loolcool13:47
loolhmm image build returned, but that's too short13:48
loolpsivaa: you installed said packages?  (updated first column)13:50
asaclool: i guess we should create one sheet for each build id13:50
asacso we know if the results are up to date :)13:50
loolasac: I've put the details on the left13:51
asaclool: so with the stuff installed, the unity8 autopilots still fail miserably13:51
loolimage id + package revnos13:51
asaci see that13:51
looland I've hidden older runs13:51
asacbut still we could jkust keep the 83 result and create a new for whatever build you want to see tested13:51
asacanyway13:51
asacunity8 APs fail miserably still afaict13:51
loolasac: the 83 results are still there13:51
loolsergiusens: heya13:52
asacright. but where is the 86? :)13:52
loolsergiusens: so dont convert the apps which have URL handlers to click yet13:52
loolsergiusens: calendar, music, gallery, mediaplayer13:52
loolsergiusens: cause they need updated upstart-app-launch13:52
sergiusenslool, calendar has them?13:52
loolsergiusens: there are some calendar:// links yes13:53
sergiusenslool, darn, I knew about music, gallery and camera13:53
sergiusenslool, so when can that happen?13:53
sergiusenswe did calendar last night13:53
loolsergiusens: reviewing mp13:53
loolsergiusens: after mir13:53
sergiusenslool, isn't that really risky?13:53
sergiusensunless mir happens like now13:53
loolsergiusens: what do you want to break?  :-)13:54
ogra_asac, lool, http://paste.ubuntu.com/6209409/ but i would really rather not do that just to hide a not properly researched bug ...13:54
loolsergiusens: hey I think I need some of your help to prepare an upstart job to switch to mir for people upgrading wtih system-image13:54
sergiusenslool, I prefer url handlers as a bug than switching when it's too late to fix any unknown13:54
loolsergiusens: if upgrading from build <= xyz, then touch .display-mir; done13:54
asacogra_: yeah. at least get a good bug for that13:54
asac:)13:54
asacthanks13:54
loolsergiusens: so we want something shippable by thursday; then we have one week to land safe bug fixes13:55
asacogra_: i thought there was a bug already, check with lool and tvoss13:55
loolsergiusens: but the deadline is that thursday needs to have mir and be good13:55
loolasac: I passed ogra_ the bug link already13:55
asaclool: can you give it to me as well?13:55
loologra_: isn't there standard infrastructure to set this?  /etc/init.d/ondemand or such?13:56
loolasac: 15:11 < lool> ogra_: https://bugs.launchpad.net/ubuntu/+source/mir/+bug/123519013:56
ubot5Ubuntu bug 1235190 in mir (Ubuntu Saucy) "[mako] Unity8 on Mir got slow" [High,Confirmed]13:56
asacso with latest from ppa unity8 still crashes with unity8 crashing13:57
loolsergiusens: Mir should be tonight13:57
loolbut we need to lock asac in a closet first13:57
asacunity8 AP fails with unity8 crashing i mean13:57
ogra_lool, no, that sysv script is desktopish and built to force ondemand13:57
ogra_lool, we dont want to abuse it for a temporary hack on touch13:57
loologra_: I know, just felt it was handling this kind of settings13:57
ogra_definitely not13:58
asaclool: the thostr_1 hud fix ... is that in the ppa already?13:58
loologra_: I thought there was some config file for this13:58
ogra_this is one of the grossest hacks we have done yet13:58
loologra_: or wouldn't e.g. sysctl.d/*.conf work for this?13:58
ogra_i will definitely not add it to some sane script13:58
sergiusensplars, fwiw, just seeing that calendar tests are broken com.ubuntu.calendar-app is supposed to be com.ubuntu.calendar13:58
psivaalool: the versions are 0.1+13.10.20131008.1-0ubuntu1, the sheet says the source code versions i suppose13:58
loolasac: no, it's at bottom of landing asks, waiting on code13:58
ogra_lool, i dont think sysctl covers cpufreq13:58
asaclool: thats the one that tvoss said might fix our unity8 crashes, right?13:59
asacwasnt there a MP13:59
asac?13:59
loolpsivaa: Yes; you be able to confirm from changelog13:59
ogra_lool, this is a hack we should remove asap (and i still think it is wrong to do it at all)13:59
psivaalool: ack, will confirm that13:59
loolpsivaa: or by looking at PPA .changes13:59
ogra_lool, so the lxc-android-config package is the proper place to not affect desktop13:59
loolasac: yes13:59
loologra_: Ok; could you test it?13:59
ogra_will soon14:00
plarssergiusens: confusing... is someone fixing that? The module for autopilot will still be calendar_app though right?14:00
asaclool: ok, so we need that for mir i guess. thanks14:00
ogra_just testing another fix14:00
loologra_: perhaps how battery life is affected / does device get hot when not in use / when in use14:00
sergiusensplars, yeah, just the package name is wrong, I'll fix14:00
sergiusenslool, music app as click is working fwiw14:01
ogra_lool, right14:01
loolcjwatson: Hmm I dont see where my livefs build log went14:01
ogra_lool, battery testing takes time though14:01
cjwatsonlool: The public site mirrors hourly14:01
ogra_the hack is small enough that we can quickly add and remove it though14:01
loolcjwatson: ok thanks14:02
cjwatsonlool: Is there a problem, or just curious?14:02
thostr_1thostr14:04
thostr_1asac: lool: no hud fix landed yet.... we're still testing making sure it really works14:05
asacthostr_1: ok, but you have one that looks promossing??  will that fix the unity8 autopilots?14:06
cyphermoxlool: pushing the buttons now to land ofono and telepathy-ofono, unless you tell me now that I should still wait14:06
asacwith mir?14:06
thostr_1asac: yes, that fix adds/completes hud support for mir14:06
asacthostr_1: so you say you have run unity8-autopilot with mir enabled and it succeeds? wow. cool!14:07
cyphermoxrsalveti: do you normally do the uploads for ofono yourself or can I just take the code and upload it?14:08
thostr_1asac: no, I said we're testing this. I'm not saying yet it fully works14:08
asacthostr_1: kk. please check unity8-autopilots. and sorry for asking so details, want to get a better picture of whats coming today14:08
sergiusensplars, this is sort of ugly, but gets the job done https://code.launchpad.net/~sergiusens/phablet-tools/click_testing_python_modules/+merge/18974414:10
plarssergiusens: if we just need python-mock also, I'm happy to add that on our side14:15
plarssergiusens: which one needs that? It doesn't look like it's calendar14:15
plarssergiusens: just filemanager and music-app?14:16
sergiusensplars, terminal and/or filemanager14:16
sergiusensplars, well if we do it in phablet-click.... there's no need to go into rw14:17
cyphermoxlool: ok to push?14:17
plarssergiusens: this is true. When are those slated to arrive as click?14:17
sergiusensplars, today14:17
plarssergiusens: not in 87 though right?14:18
sergiusensplars, well music right after mir14:18
sergiusensnope14:18
plarsok14:18
plarsso we should have what we need for 87, except there could be problems because of s/calendar-app/calendar right?14:18
sergiusensplars, I'm running ap right now to test the fixes for calendar14:19
loolcyphermox: yes14:25
loolcyphermox: you'll need to manually upload ofono14:25
loolcjwatson: my build failed, but no idea why14:26
cyphermoxyeah14:26
loolcjwatson: or it was way too short14:26
loolcjwatson: not seeing the build log still14:26
cjwatsonlool: You didn't ask it to do a livefs build14:27
cjwatsonSo it reused the existing one14:27
loolcjwatson: right14:27
cjwatsonlool: I advise copying and pasting from the crontab14:27
loolcjwatson: that's what I do 99% of the time; first time I didn't do it, crap14:27
cjwatson(i.e. you forgot the --live option, it seems)14:27
loolactually I ran for-project ubuntu-touch cron.daily-preinstalled --live I think14:28
cjwatsonYou didn't14:28
loolweird14:28
loolmaybe a paste error14:28
cjwatsonThe log is in http://people.canonical.com/~ubuntu-archive/cd-build-logs/ubuntu-touch/saucy/daily-preinstalled-20131008.1.log14:28
cjwatsonNote the difference at the top between that and 2013100814:28
loolcjwatson: indeed, I've missed14:28
loolcjwatson: did you run it?14:28
loolor can I run one14:28
cjwatsonI left it to you14:28
loolthanks14:29
loolrunning14:29
loolseb128: are you working on the bluetooth crasher in settings?14:30
seb128lool, no, I've no clue about bluetooth, charles_ owns it14:30
charles_lool: I'm working on bluetooth now14:31
seb128lool, what bug is that? I'm not even sure we have it/its a settings bug14:31
asacpsivaa: do you have a mako as well?14:31
psivaaasac: no, only maguro14:32
charles_lool, do you have a bug # on that?14:32
loolcharles_: cool14:34
loolcharles_: I've given a heads up to olli about, sorry if he pings you again (reviewing high prio crashers with jfunk)14:34
loolcharles_: yes will give you bug id14:34
loolcharles_:  * bluetooth crasher in settings when turning back on (LP #1236705)14:35
ubot5Launchpad bug 1236705 in unity8 (Ubuntu) "unity8 crashed with SIGABRT in __gnu_cxx::__verbose_terminate_handler(), thrown from mir::DefaultServerConfiguration::the_connector()" [Critical,Confirmed] https://launchpad.net/bugs/123670514:35
loolcjwatson: and sorry for the confusion BTW14:36
cjwatsonnp14:36
loolI didn't remember where the cdimage logs were14:36
charles_lool, seb128: at first glance it looks like a mir crash14:37
=== charles_ is now known as charles
loolcharles: could you escalate this?  is it some UI thing you're doing that's crashing it?  minimal test case + reassign to Mir perhaps?14:38
charleslool, let me test & see if I can reproduce it14:38
charlesthis is kind of in the area that I'm working on, with interaction between the bluetooth indicator & settings panel14:39
charlesbut I haven't seen a crash14:39
charleslool, seb128: I flashed my nexus 7 about 30 minutes ago with devel-proposed and followed the recipe in https://launchpad.net/bugs/1236705, but am not seeing a crash14:44
ubot5Ubuntu bug 1236705 in unity8 (Ubuntu) "unity8 crashed with SIGABRT in __gnu_cxx::__verbose_terminate_handler(), thrown from mir::DefaultServerConfiguration::the_connector()" [Critical,Confirmed]14:44
loolcharles: are you using Mir?14:44
loolcharles: not sure that works on nexus714:44
sergiusensplars, lool fixes for calendar coming in14:47
charleslool, erm, good question. let me check14:48
rsalvetilool: to test the android package you need to rebuild the image, that's why it was reviewed and tested by sergiusens before the upload was done14:48
asaclool: so you said that we have the fix for enabling hwdecode in browser ready?14:48
asacrsalveti: ?14:48
rsalvetiit's a one liner that fixes the index of the loop, you can easily see that it was wrong14:48
rsalvetilool: and property watcher wasn't working, it's now :-)14:49
rsalvetiasac: ?14:49
sergiusensrsalveti, I already explained btw ;-)14:49
rsalvetisergiusens: yeah, just explaining it again :-)14:49
asacrsalveti: browser has no hwdecode was said you had a fix for14:49
rsalvetiin case lool itself wants to test everything14:49
rsalvetiasac: has hwdecode, missing support for software rendering, same as thumbnail, which is in progress14:49
charleslool, ps isn't showing unity-system-compositor running on the nexus... so, I guess not. :/14:50
asacrsalveti: i thought we dont have that in browser working. then i have bad info it seems14:50
asacik14:50
loolsergiusens: I've approved the upstart-app-launch work to get the URLs working with click14:53
loolin dispatcher14:54
loolsergiusens: but will take some time to land14:54
rsalvetiasac: the issue is the same as with the thumbnail, you'll see a corrupted screen14:54
rsalvetibut it's decoding the video using the hw decoder14:54
loolasac: we will NOT get hwdecode in browser14:54
loolasac: software only14:54
sergiusenslool, which MRs?14:54
asaclool: we will not get it for 13.10?14:54
rsalvetilool: we *do* get hwdecode14:54
rsalvetiwe don't hw *rendering*14:54
loolah, sorry for mixing the two14:54
asacbut will we be able to view 1080p h.264 video?14:55
asacrsalveti: ?14:55
loolin the browser?14:55
asacyeah14:55
rsalvetiasac: yup, but not with 0-copy yet14:55
rsalvetiour webkit version doesn't support texture streaming14:55
asacrsalveti: thats a GL extension, right?14:55
rsalvetiasac: not actually, the browser would just need to create a egl texture and use that when rendering it's own engine14:56
rsalvetithat's happening in upstream, but qt also dropped support for webkit14:56
rsalvetiso I expect this to happen just when they move away from webkit14:56
rsalvetiand switch to blink14:56
ogra_lool, asac, so after using mako on performance for about 30min it drained 10% battery ... and this is without SIM, only wlan14:57
asacrsalveti: ok, so no youtubre without draining battery :)14:57
asachehe14:57
rsalveticyphermox: please, just use bzr bd and you should have something valid to upload14:57
asacrsalveti: fine. wonder why we didnt know before though :)14:57
loologra_: but with screen on?14:57
asacogra_: we have jfunk testing that14:57
asacand giving us an assessement14:57
loologra_: if you leave it unattended / unplugged, does it remain hot, and does it drain?14:57
ogra_lool, yes, 30min browsing and playing around with apps14:57
loolthat's expected to drain it though  ;-)14:58
ogra_it is constantly warm when on14:58
rsalvetiasac: because we didn't technically investigated it until a few weeks ago, we were still focusing on getting hw decode to work14:58
ogra_lool, well, 12h vs 5h is quite a hit14:58
loolrsalveti: android >> it's not that I want to test every single thing, but it's a touch specific package and I didn't understand the change and it went in without a heads up14:58
rsalvetilool: because it's not even used yet14:58
ogra_lool, asac, also note that this change doesnt seem to have any influence on maguro performance14:58
loolrsalveti: a) couldn't tell  b) can still break stuff14:58
ogra_the UI is still stuttering14:59
rsalvetiI wouldn't push stuff that would affect the image without pinging you :-)14:59
rsalvetithat's why I didn't push ofono14:59
ogra_rsalveti, do you know anything about why the omabfb driver sends out uevents for each vsync ? does PVR use it ?15:00
rsalvetiogra_: well, not that I know, but I believe such event might be currently dropped in android15:00
rsalvetiogra_: you could ignore it in udev and see what happens15:00
loolcharles: can you try a bit harder?  it seems the bug can sometimes be triggered on SF15:00
loolcharles: I'm also asking for someone to teston Mir15:01
rsalvetiotherwise we can try to ignore it in upstart, to avoid it sending a dbus message for every vsync15:01
ogra_rsalveti, lol, no, fire up dbus-monitor --system on maguro and see yourself :)15:01
ogra_rsalveti, right, we have a patch for upstart-udev-bridge15:01
ogra_but still need one for udev15:01
cjohnstonasac: did now work for you for a 1:!?15:01
cjohnston1:1 too15:01
rsalvetiyou probably just need one for udev15:01
charleslool, ack15:01
* ogra_ has a quiet dbus over here already, but also a udevd that constantly eats 5% CPU15:01
ogra_rsalveti, well, i was hpoing we could just drop the code from the driver15:02
ogra_rsalveti, and was trying to understand why it is there at all15:02
rsalvetiogra_: oh, ok, I might be able to take a look15:02
rsalvetilet me clone the maguro tree15:02
asaccjohnston: this was suppoed to be moved over to ev15:03
cjohnstonahh15:03
asaccjohnston: maybe an oversight15:03
asaccjohnston: seems its because you hadf it on your calendar... can you move over?15:03
asacthanks15:04
cjohnstonsure.. I'll talk to ev and see when he wants to chat15:04
ogra_rsalveti, the function around line 2303 in drivers/video/omap2/omapfb/omapfb-main.c15:04
evmsm seems to have missed that one when moving the rest over. Sorry guys15:04
evcjohnston: happy to chat after this call15:04
cyphermoxrsalveti: yeah almost done15:04
rsalvetiogra_: thanks15:04
ev~25 minutes15:04
rsalveticyphermox: great15:04
evif that works for you?15:04
cjohnstonsure15:05
cyphermoxlool: rsalveti: telepathy-ofono and ofono done15:08
cyphermoxlool: btw, landing asks is out of date... I saw you mentioned I had two asks, I see only one and that one has landed already15:09
cyphermoxhowever, there will still be NM 0.9.8.4, which isn't in asks yet if you want to add it or something15:09
rsalveticyphermox: great!15:10
rsalvetithanks15:10
Saviqfginther, we always only had one mako for s-jenkins?15:11
sergiusensfginther, i just had to mannually launch the saucy VMs; I launched two to get my jobs run and noticed others taking the queue meaning they weren't being launched for a while15:12
sergiusensfginther, did that ever happen to you?15:12
fgintherSaviq, we're fighting a temporary config problem and need to reload the configuration soon to fix it15:12
Saviqfginther, k15:13
fginthersergiusens, i think it's related. jenkins accepted a bad none name and lots of things stopped working15:13
fginthersergiusens, s/none/node/15:13
cyphermoxrsalveti: https://code.launchpad.net/~mathieu-tl/ofono/1.12+bzr6836-0ubuntu1/+merge/18988415:13
loolcyphermox: check the list, there are two15:17
loolcyphermox: you have two landing slots in the main plan15:17
cyphermoxah, in plan15:18
cyphermoxyeah, so 111 is landed akready15:18
cyphermoxand yeah 123 is just uploaded15:19
loolcyphermox: yup, updated, thanks15:22
sergiusensfginther, thanks15:24
cyphermoxlool: can you add asks / plan for NM 0.9.8.4 ?15:28
loolcyphermox: what's this?15:28
loolnew upstream?15:28
cyphermoxlool: NM bugfix release, yeah. a new minor version15:28
asaclool: the new image is ready for pick uyp?15:32
asacpsivaa: plars: ^^ 8715:32
asaclool: we want 87 right? or was it 88?15:33
plarsasac: yeah, testing has kicked off on it15:33
asaccool15:34
rsalveticyphermox: thanks15:36
psivaalool: asac: do you want me to run the mir testing on 87 then?15:40
asacpsivaa: might be good to see if anything improved at all15:41
asacafaik there are still some things in the ppa15:41
psivaaasac: ack15:43
popeyasac / lool we're not expecting webapps to work in #87, right?15:53
asacpopey: would be nice if you could get dbarth more involved here15:54
asacsorry, but cant answer your question i think15:54
popeyk15:54
asacplars: how do sneak results look like?15:55
plarsasac: not much to see yet, install and default have run on mako with no problem, install still running on maguro15:55
asaclool: so notes-app didnt make it?15:56
asacfor .2?15:56
asaclool: seems we didnt build the new revision from trunk yet even15:56
Mirvas the hangout replaces our weekly irc chat I'm joining but again in a noisy place because of the evening15:58
psivaaplars: in running the AP tests for one of the click packages i get. http://pastebin.ubuntu.com/6209898/16:01
asaclool: coming?16:02
psivaaplars: do you think i should do anything else16:02
plarspsivaa: you need to be the phablet user16:02
plarspsivaa: you should just run with jenkins.sh from the touch branch16:03
psivaaplars: ok, ill try them. thanks16:03
plarspsivaa: if you installed with provision.sh, then all you need to do is ./jenkins.sh -s <your serial> -a <app>16:03
psivaaplars: i have used phablet-flash but i'll try those with provision16:04
loolthostr_1: https://code.launchpad.net/~indicator-applet-developers/hud/support-unity8-window-stack/+merge/18981016:43
loolthostr_1: a lot of people believe that fixing hud will trigger less unity8 crashes and that it's important while switching to mir16:43
loolthostr_1: the first of the two Hud branches is approved, this one is needs review, is review already assigned on this one?16:44
loolthostr_1: (thanks!)16:44
thostr_1lool: ted and pete are working together so they can cross review16:45
loolthostr_1: perfect, thanks16:46
loolthostr_1: FYI, I've acked two MPs (one was already acked) to allow click apps to be launched by url-dispatcher to allow switching remaining apps to click (calendar was bogusly already switched in #86 or so)16:47
loolthostr_1: this was something I wanted to defer to post-13.10, but we need it so...16:47
loolthostr_1: this is url-dispatcher code BTW16:47
thostr_1lool: ok, ted wanted to sync with you that anyway at some point... now hud got into way16:48
loolthostr_1: I think we did earlier today16:49
loolbut perhaps that was another sync16:49
* lool updates to #8716:49
thostr_1no, that might have been it... it's already 3 hours I talked to him16:49
Mirvlool: if there's pressure to get the thumbnailer in then it should be good to go, but thostr_1's mediascanner crash report made me doubt enough so that I'd like to double-check the thumbnailer with Satoris16:49
lool50 MB update again16:50
loolgosh16:50
loolMirv: is mediascanner crashing with new thumbnailer?16:50
Mirvlool: it's a good question, https://bugs.launchpad.net/mediascanner/+bug/1236940 - I don't see that, but I also can't evaluate whether it's a new crash or would happen also with the old thumbnailer16:50
ubot5Ubuntu bug 1236940 in mediascanner "Mediascanner crashes" [Undecided,New]16:50
thostr_1Mirv: I think that crash would also happen with the old one...16:51
loolthostr_1: ah thanks, just what I was asking in the bug16:52
thostr_1Mirv: but I'm seeing other strange behaviour.. have you tried adding new videos after a while? When I did that no video is shown any more16:52
loolthostr_1: wit old mediascanner or old thumbnailer?16:52
thostr_1lool: I only upgraded thumbnailer16:52
loolthostr_1: so you were testing the mediascanner from image?16:53
Mirvthostr_1: the problem is that I haven't seen video thumbnails for some time with or without the thumbnailer update, which is I guess also known16:53
loolMirv: yeah; I saw /some/ between crashes with updated mediascanner + mediascanner-scope16:53
thostr_1Mirv: yeah...16:53
loolbut it was too crash, filed a bug on that one16:53
ogra_Mirv, yeah, thats an issue of the MM stack16:53
loolthostr_1: did James H look into the double-free?16:53
thostr_1lool: that is fixed16:54
loolcool16:54
thostr_1lool: https://bugs.launchpad.net/ubuntu/+source/mediascanner/+bug/123663716:54
ubot5Ubuntu bug 1236637 in mediascanner (Ubuntu) "r389 fails to build with double free error" [Critical,New]16:54
thostr_1lool: I think I might have found something: CRITICAL critical: gst_amc_format_free: assertion 'format != NULL' failed16:55
thostr_1that comes from gstreamer when thumbnailing... seems like a missing or broken codec16:55
loolhttps://code.launchpad.net/~ted/hud/fix-hud-activation/+merge/189869 looks really fishy actually16:56
loolthostr_1: yeah16:57
loolrsalveti: ^ does this ring a bell?16:57
Mirvlool: ogra_: so IMHO the new thumbnailer does not regress anything and would fix the SDK on desktop, but I just feel I don't have the complete picture because there are so many issues elsewhere. if you think it seems ok I'd like to release it, but otherwise I'll recheck in the morning with Satoris.16:57
loolrsalveti: the CRITICAL critical: gst_amc_format_free: assertion 'format != NULL' failed16:57
loolMirv: I think you can land it then16:58
ogra_++16:58
Mirvok, thanks16:58
loolMirv: this one is relatively safe, I know the feature we will be lacking is completely not release critical16:58
rsalvetilool: hm, in bug 1236637?17:02
ubot5bug 1236637 in mediascanner (Ubuntu) "r389 fails to build with double free error" [Critical,New] https://launchpad.net/bugs/123663717:02
rsalvetiwe shouldn't be using our stack when building stuff in the archive17:03
=== alan_g is now known as alan_g|EOD
=== psivaa is now known as psivaa-afk
dobeylool: hey. there's a package pulled into the touch image via hard dependency, ubuntuone-client-data. i'm about to upload a new version into ubuntu, but there's no source change to it. it's only a version bump. do we need to do a special landing request for it or anything, or it's ok and will just get pulled in the next image builds once it's in the archive?17:13
rsalvetilool: how did you get that message?17:14
ogra_rsalveti, thostr_1 had it17:14
ogra_trying to generate thumbnails17:14
thostr_1rsalveti: I copied the video from http://www.doi.gov/news/video/thisweeksept202013.cfm to device17:15
loolrsalveti: mediascanner on some files (thostr gets it) with thumbnailer17:15
looldobey: will check after dinner17:16
thostr_1rsalveti: see also https://bugs.launchpad.net/thumbnailer/+bug/123687417:16
ubot5Ubuntu bug 1236874 in Thumbnailer "No thumbnailing even though image is playable on phone" [Undecided,New]17:16
rsalvetithostr_1: right, the lack of thumbnail is what we're currently working on17:17
rsalvetiit uses the hw decode elements, but it fails to copy it over for sw rendering17:18
thostr_1ah, that perfectly explains this17:18
thostr_1can you update/reassign the bug accordingly17:18
rsalvetithostr_1: yes, thanks17:18
thostr_1rsalveti: thank you!17:19
=== psivaa-afk is now known as psivaa
rsalvetidone17:22
asaclool: how is the spin of the final things we wanted to pull goinmg?17:23
asacdid we pick up half-hud as well?17:24
asace.g. in ppa for testing17:24
rsalvetilool: is bug 1236355 still happening?17:24
ubot5bug 1236355 in unity-scope-mediascanner (Ubuntu) "Crashes when generating thumbnails for some videos" [Undecided,New] https://launchpad.net/bugs/123635517:24
asacrsalveti: is there a landing ask that would fix that?17:25
rsalvetiasac: the crash loop issue was in theory fixed in the thumbnailer17:26
asacrsalveti: the bug is pretty empty on this regard17:27
rsalvetiright, that's why I'm asking lool to see if it's still happening with latest17:28
plarsthe notes-app tests seem to be getting stuck and timing out17:42
plarseven on sf17:42
popeyasac: is #88 ETA today?17:51
ogra_kind of17:51
ogra_it would only be released tomorrow morning though17:52
asacpopey: some stuff still coming in afaict17:58
asacbut we try to do it before EOD17:58
asaclool knows all17:58
sergiusenslool, do I have headways to switch more apps?17:58
asacafaik we havent seen the crashes getting fixed in click apps yet18:00
asacso, i prefer to do that later :)18:00
ogra_as long as it is before release :)18:01
* ogra_ wants to get rid of the PPA18:01
sergiusensasac, doesn't that mean I have to wait for mir?18:01
asacyeah18:02
sergiusensmeaning potentially not landing or landing too late?18:02
asacright. everything has to stand back now18:02
sergiusensasac, that's just hiding a problem; click apps are being used daily; crashes are just happening everywhere18:02
asacwho knows18:03
asacso fix that first18:03
* sergiusens avoids discussing again18:03
jdstrandfyi, landing ask line 124 "apparmor: fix AF_UNIX..." is now "Code ready"18:17
ogra_jdstrand, thats the one that needs to go in together with the kernel fixes ?18:18
loolsergiusens: URL apps or non-URL apps?18:18
loolsergiusens: URL apps needs u-a-l to land first18:18
jdstrandogra_: yes18:18
ogra_yay18:18
jdstrandogra_: apparmor + kernel. lool said I can upload apparmor-easyprof-ubuntu now (in landing page for mir)18:19
jdstrandso that will be handled shortly18:19
looljdstrand: Yup, shoot18:19
ogra_great18:19
* ogra_ is waiting for ureadahead fixes in the kernels that were held back for this 18:19
looljdstrand: so the kernels are also getting ready?18:19
looljdstrand: will likely be post Mir then18:19
jdstrandlool: the kernel team and cking tested those, yes18:20
jdstrandsorry, the security team and the kernel team (cking in particular)18:20
sergiusenslool, terminal and filemanager where my thoughts18:20
loolrsalveti: apparently this was a grilo bug: https://bugs.launchpad.net/ubuntu/+source/mediascanner/+bug/123663718:21
ubot5Ubuntu bug 1236637 in mediascanner (Ubuntu) "r389 fails to build with double free error" [Critical,New]18:21
loolwill check18:21
jdstrandlool: apparmor is ready and is also tested. it can go any time. please contact apw for the kernels18:21
jdstrandapparmor-easyprof-ubuntu uploaded18:21
* jdstrand heads over to #ubuntu-release18:21
looljdstrand: thanks18:21
loolsergiusens: Let's just wait for apparmor-easyprof-ubuntu to be approved, then you can go; let me add a slot18:21
loolsergiusens: note that clicks dont work on mir without new apparmor18:22
=== Ursinha is now known as Ursinha-afk
loolsergiusens: I've rated this medium because terminal is mission critical for some folks  ;-)18:23
sergiusenslool, ack, also is zeitgeist crash already landed?18:24
sergiusenslool, and last but not least; is 87 from earlier today or fresh?18:24
loolsergiusens: no, it's in PPA, going in with Mir tonight18:25
loolsergiusens: looking18:25
loolsergiusens: oh image #87 is from this afternoon18:25
loolsergiusens: like 2pm UTC18:25
loolor 3pm rather18:25
sergiusenslool, hmmm, why didn't my notes and calendar fixes make it in :-/18:26
loolbfiller: You added an ask for notes-app, but sergiusens already uploaded it tihs morning  :-)18:26
loolsergiusens: what's the expected notes version?18:27
loolsergiusens: this is in: com.ubuntu.notes        1.418:27
loolcom.ubuntu.calendar     0.418:27
sergiusenslool, give me the click list --manifest18:27
loolhttp://paste.ubuntu.com/6210481/18:27
bfillerlool, sergiusens : don't see a release in the changelog of notes-app trunk, bunch of things were merged today18:27
sergiusensbfiller, it's click now18:28
loolsergiusens: r206 is there18:28
sergiusenslool, 206 has my fixes :-/18:28
bfillersergiusens: how can you tell when a click was released?18:28
bfillersergiusens: need r20718:29
loolpsivaa: in case you're still trying Mir stuff manually, leave it some time until we land more input focus fixes; the results you have posted are bad enough to confirm we want them; thanks a lot for listing them up there18:29
sergiusensbfiller, on the device run click list --manifest; com.ubuntu.notes will have a bzr revno in there18:29
sergiusensbfiller, it's just missing ugo's qmlscene fixes which isn't used in click anyways18:30
loolsergiusens: what's the problem is r206 has your fixes and r206 is there?18:30
loolwe just need to do another one for bfiller's fixes in r20718:30
loolthat's alright18:30
bfillersergiusens: is there a bzr branch for the click packaging?18:30
loolbfiller: it's the same one18:30
sergiusensbfiller, it's all in code18:31
loolsergiusens: good idea on README.click BTW18:31
sergiusensbfiller, there's a readme in there18:31
ogra_++18:31
bfillerlool, sergiusens: would be nice to add a commit or tag or something so I could look at trunk and see what rev has been released as click as I did with deb18:31
loolbfiller: yeah18:32
loolbfiller: currently parts of the relesae process is manual (sergio uploads to appstore)18:32
loolbfiller: perhaps we can do that manually too?18:32
loolsergiusens: ^18:32
loolsergiusens: until that's automated that is18:32
loolsergiusens: so were you worried because the tests aren't passing?18:32
loolthey aren't failing either, they are somehow in limbo18:32
loolplars: mind checking what the notes-app click tests are doing?18:33
sergiusensbfiller, lool there is no fun dch in click, but I can hack something if you want18:33
sergiusenslool, yes, it isn't loading for some reason18:34
sergiusenslool, the clock failuers I tracked down to clock app using dconf, which isn't allowed18:34
plarslool: yeah, I'm looking18:34
loolsergiusens: just tag it appstore-0.4 or something?18:35
loolsergiusens: uhoh18:35
apwjdstrand, whenever the apparmor is up let us know, we have the kernels pending ready to copy18:35
loolsergiusens: what is it for?  do we need it unconfined, or is this internal?18:35
sergiusenslool, to store settings I think, being discussed in #ubuntu-touch18:36
loolapw: apparmor is going in to fix mir, but we'd like to hold kernels til tomorrow to focus on landing mir tonight, unless anything dramatic happens18:36
loolapw: could we coordinate this tomorrow?18:36
loolsergiusens: thanks for passing that on18:36
loolsergiusens: let me know when the fix comesin18:36
plarslool: all I know right now though is that they seem to either be taking forever, or getting stuck at some point and timing out18:36
plarslool: I'm setting things up to run them locally so I can watch though18:37
sergiusensplars, I think it slows down once apport gets in18:37
jdstrandlool: to be pendantic-- apparmor-easyprof-ubuntu is going to fix mir, and is uploaded. the kernel also needs an apparmor upload to happen at the same time or before the kernels18:37
plarssergiusens: could be18:37
jdstrandapw: ^18:37
jdstrandlool: that is tested and ready to go at any time. just let me know18:38
loolplars: ok18:38
looljdstrand: Oh right _apparmor_ itself, I remember it from the ask18:38
looljdstrand: if you want to push apparmor now too, that's fine18:38
jdstrandlool: yeah-- seems like they go together, but they don't. apparmor-easyprof-ubuntu is just click policy. apparmor is the whole shebang18:39
jdstrandso we separated them18:39
loolbfiller: so do we need r207 of notes-app?18:39
jdstrand(updating policy is less disruptive than apparmor uploads)18:40
looljdstrand: sure, I understand; had just forgotten that apparmor itself also needed an upload18:40
bfillerlool: yes for it to work properly with MIR I believe so, sergiusens is this still true?18:40
looljdstrand: did you test the mir change with new or old apparmor?18:40
jdstrandlool: new18:40
looljdstrand: ok, then it might as well go in now18:40
jdstrandlool: is that an ok to upload? :)18:41
sergiusensbfiller, lool 207 is just for desktop/deb mode http://bazaar.launchpad.net/~phablet-team/notes-app/trunk/revision/20718:41
loolfginther: https://code.launchpad.net/~dandrader/unity8/fix_lp1236773/+merge/189821 seems bad; do we have a spare mako there?18:42
sergiusenslool, bfiller that said, you can daily release it and nothing will break18:42
looljdstrand: yes18:42
looljdstrand: definitely18:43
looljdstrand: updated the landing slot18:43
jdstranduploading18:43
* jdstrand heads to #ubuntu-release18:43
fgintherlool, we're down to two at the moment, one was taken offline a little bit ago because it was failing everything. I'm trying to get it back up18:44
fgintherlool, that MP is running now on mako and maguro18:44
loolfginther: awesome18:45
psivaalool: ack, was having a bit of difficulty running click package app tests..18:49
loolpsivaa: these are actually all failing on Mir18:52
loolpsivaa: and we just uploaded a fix18:53
loolpsivaa: apparmor-easyprof-ubuntu18:53
loolpsivaa: minutes ago18:53
psivaalool: ack, thanks :)18:53
loolbfiller, sergiusens: right, so I'll abstain on the notes-app thing; bfiller I've pointed your ask as sergiusens' upload18:54
loolbfiller, sergiusens: we don't need to fix notes-app.deb18:54
loolSaviq: can unity8 tip with all merges go in without unity-mir/upstart-app-launch/qtubuntu?19:06
loolSaviq: I wanted to try running these tests alone19:07
lool(would rather land things one by one than as a group)19:07
Saviqlool, there's no hard requirement, but don't expect fireworks if you don't get it all19:07
Saviqlool, as in, unity8 only had a handful of fixes - that don't work without the other things19:08
asaclool: are all the bits we want to check now built?19:10
loolSaviq: yes, my goal is to take unity8, test it, mark it good, move to the next one19:12
loolSaviq: probably qtubuntu next19:12
loolthen upstart-app-launch + unity-mir as one19:12
loolasac: it's all merged now!19:13
loolsince 1mn19:13
loolnow looking at getting it in PPA19:13
Saviqlool, and why won't you just make sure all works in concert?19:13
Saviqlool, and in case of regressions we'll do the long route?19:13
Saviqanyway, your call ;)19:14
loolSaviq: it's a bit of a gut feeling19:15
loolwe usually land these one by one most of the time19:15
loolsometimes 2 together19:15
Saviqk19:15
loolit's still some 7 merges in unity819:15
asaclool: hmm... so still not done before the call19:16
asaclet me know19:16
Saviqlool, huh? 3 are in queue, and they're all minor19:24
dobeylool: re: ubuntuone-client-data landing?19:26
=== thomi_ is now known as thomi
loolSaviq: oh right sorry, was checking bzr log and skipped over a changelog update19:30
looldobey: what's the goal of the version bump?19:31
dobeylool: version consistency across the ubuntuone packages.19:32
looldobey: ok, shoot19:32
looldobey: I'll update spreadshhet19:32
dobeylool: ok, thanks19:32
loolSaviq: Ok, I'll try to update everything; the qtubuntu changes are tiny and unity8 changes are smaller than I thought, it's not worth delaying testing too much19:35
loolpackage being uploaded...19:35
loolupgraded19:37
loolrebooting19:37
loolCache invalid, retry (internally handled)19:39
loolwtf19:39
loolsensorservice is pegging the CPU during boot like crazy19:39
loolcouple of crashes: _usr_bin_maliit-server.32011.crash _usr_lib_arm-linux-gnueabihf_hud_hud-service.32011.crash19:40
loolcan enter passwords19:40
loolin wifi setup19:40
loolok, launching webbrowser tests19:43
loolkgunn: so I wanted to run the webbrowser-app-autopilot, these stop unity8 and start it again immediately, but it doesn't seem to come up19:46
loolkgunn: like can't unblank display thing19:46
loolit came back after the test19:47
kgunnlool: that's weird...the webbroweser ap tests stop unity8 ??19:47
kgunnlool: is there a reason for that ?19:48
kgunnthomi: ^ just in case you know19:48
loolkgunn: to turn on "testability"19:48
loolwhat's odd is that I dont see the failure19:48
loolI see the display is actually on, albeit black19:48
loolmaliit-server uses 100% CPU19:48
asaclool: webbrowesr AP tests stop unity? is that new?19:50
asacthought we only use -n for unity819:50
loolit's actually me being stupid19:50
loolasac: right19:50
loolasac: I forgot -n was doing that19:51
Saviqlool, stop unity8?19:51
Saviqlool, right - that19:51
Saviqlool, yeah, and that's the cause for unity8 test failures, btw - only the first test will ever start19:51
Saviqlool, 'cause the others get the unblank issue19:51
Saviqlool, or well, you can `sudo stop powerd` to help19:52
loolasac: joining?19:52
loolSaviq: ok interesting19:52
loolSaviq: in the past I was getting the socket things when restarting unity8, but not anymore19:53
loolneed to find out why19:53
Saviqlool, it was improved in mir in the mean time19:53
plarssergiusens: so, trying to revert to just running notes_app tests with phablet-test-run, I get 'initctl: Unknown instance: com.ubuntu.notes_notes_1.4'20:05
plarssergiusens: anything special I need to do with phablet-test-run for running click tests?20:06
loolgetting timeouts in the browser tests20:07
sergiusensplars, I'm almost sure that's the loop waiting for the app to come up20:12
plarssergiusens: it seems to be trying to launch, maybe that's the crash20:14
asacfginther: hi20:15
asacfginther: so the latest idea is to turn off the upstream merger for everything20:16
asacand revisit tomorrow morning what to turn on if anything20:16
fgintherasac, that was going to be my question20:16
asacfginther: can you do that and send a mail to ci mailing list? CC olli.ries and lool20:16
plarssergiusens: yeah, upstart-app-launch crashing20:16
fgintherasac, ack20:17
asacfginther: "in order to ensure that the default mir switch does not cause confusion due to temporarily failing upstream merger jobs, we20:17
asacdecided to turn of upstream merger and autolanding until further notice ...20:18
plarsasac: hopefully image 88 will fix that?20:18
asacplars: fix what?20:18
fgintherasac, thanks20:18
asacplars: the app tests? yes20:18
plarsasac: the upstart-app-launch crashes20:18
plarsasac: ok, I *think* that might be the source of a lot of our problems20:18
asacfginther: tell them we plan to revisit the state of the dashboard tomorrow and might enable CI participants who have green tests again, but that they should stay tuned20:19
plarsgiven what I'm staring at locally, I'm sort of amazed anything is working properly20:19
asacand that them waiting will give the mir team a less noisy environment to focus making this key feature happening in time for saucy!20:19
asacfginther: ^^20:19
asacthats the full story i guess :)20:19
fgintherasac, thanks20:19
asacfginther: CC olli.ries@ and asac@ and send to ubuntu-phone@ and ci@20:20
fgintherasac, should that go out to ubuntu-engineering as well?20:20
asacfginther: i dont think so... if at all it should be ubuntu-devel@l.u.c20:21
asacfginther: CC olli ... he will take the internal share role20:21
asacand forward where needed20:21
fgintherasac, ok20:21
asacfginther: use ubuntu-phone@... we can still forward as we like20:21
loolSaviq: it's odd, I can ran the tests with stop powerd, but getting:20:34
loolRan 22 tests in 983.110s20:34
loolFAILED (failures=21)20:34
loolSaviq: looks like AP can't verify the widget data20:35
Saviqlool, only first test passed, I imagine?20:35
Saviqlool, what's in stderr / stdout?20:36
loolSaviq: yeah20:36
loolSaviq: I see stuff like:20:36
looldebian phablet-click-test-setup phablet-config phablet-demo-setup phablet-dev-bootstrap phablet-flash phablet-network phablet-test-run phabletutils repo setup.py tests (process:12162): CRITICAL **: Unable to get session bus: Error calling StartServiceByName for com.canonical.hud: GDBus.Error:org.freedesktop.DBus.Error.TimedOut: Activation of com.canonical.hud timed out20:36
loolfile:///usr/lib/arm-linux-gnueabihf/qt5/qml/Ubuntu/Components/MainView.qml:257: TypeError: Cannot call method 'hasOwnProperty' of null20:36
loolprocess-return-code: {{{-11}}}20:36
loolthat's SEGV I think20:36
Saviqlool, yes it is20:38
Saviqlool, CRITICAL **: Unable to get session bus: huh?20:39
loolSaviq: It might be the same issue url-dispatcher has20:39
Saviqlool, actually that's hud's fault20:39
Saviqlool, you should have a unity8 crash file, then20:39
loolSaviq: apparently upstart fails to set DBUS env properly in some cases20:40
Saviqlool, apport-bug it?20:40
loolSaviq: I do have it20:40
loolI see /usr/lib/arm-linux-gnueabihf/hud/hud-service crashed with no DBUS in env20:41
popeyhmm, mediascanner is eating cpu on #83 here20:43
popey 2274 phablet   20   0  210m  22m  15m S 100.2  1.2   6:44.65 mediascanner-se20:44
bregmafginther, the upstream merger freeze is not aplying to the Unity7 stack is it?20:47
popeyhttps://bugs.launchpad.net/mediascanner/+bug/123706520:48
ubot5Ubuntu bug 1237065 in mediascanner "mediascanner consumes a lot of cpu for a long time" [Undecided,New]20:48
fgintherbregma, that specifically wasn't discussed. does any of that go into touch?20:48
fgintherasac, lool ^20:48
bregmafginther, not that I know of, but 13.10 final freeze is effectively tomorrow for projects with autolanding, it would really hurt to lose any builds at this point20:50
fgintherbregma, http://bazaar.launchpad.net/~cupstream2distro-maintainers/cupstream2distro-config/trunk/view/head:/stacks/saucy/unity.cfg20:50
fgintherbregma, looks like a few scopes and a lens are on touch20:51
fgintherbregma, will try to get some clarification20:51
asacfginther: nothing is auto20:56
asacpeople have to ask if they want to get desktop only stuff built etc.20:56
fgintherbregma, ^. so if you have specific unity 7 centric branches, let me know20:58
fgintherbregma, -ci jobs are still enabled20:59
=== Ursinha-afk is now known as Ursinha
looltesting boot thing21:15
sergiusensplars, I think I found the cause of notes-app failing; and reason is in syslog21:15
loolWhat's the best way to test for this or that device?21:17
loolkgunn: any good way to detect whether a device is supported by Mir?21:17
loolkgunn: or should I hardcode mako || maguro?21:17
thomiHi All - what's the current problem running unity8 AP tests in the image builds?21:21
loolthomi: a) we have to stop powerd before running it, but not related to AP   b) somehow not getting the widget values; processes crash and/or not getting the data over dbus or something21:21
loolthomi: usually just the first test succeeds21:21
thomihmmm21:22
thomilool: so are those tests just disabled? I notice that the dashboard says 0 tests were run21:22
loolthomi: depends which21:22
loolthomi: but try e.g. webbrowser-app-autopilot ones21:22
thomilool: I specifically need the unity8 output, but that's OK, was more curious than anything else21:24
sergiusensplars, yup, works fine... I'll prepare an MR21:24
plarssergiusens: what's the deal?21:24
sergiusensplars, test needs an emulator param and app needs network access21:25
sergiusensplars, the initctl: Unknown instance: com.ubuntu.notes_notes_1.4 is just initctl status looping on the app_id until it finds the app to start the test21:26
plarssergiusens: I did see notes-app blink up quickly then die when I got those messages21:26
plarsdoanac: I recreated the jobs for ro and mir21:27
sergiusensplars, if you want I can create a test app for you21:27
plarsdoanac: note: we have a new test now too21:27
plarssergiusens: a test app?21:27
sergiusensplars, a testing non release notes click app21:28
plarssergiusens: for what purpose though? just trying the mechanism of upstart, click starting apps, etc?21:29
kgunnthomi: fginther ....just fyi, we are trying to land this on our trunk https://code.launchpad.net/~mir-team/mir/fix-1236912/+merge/18991621:29
sergiusensplars, oh, I just thought you were curious to try; it's ok not to ;-021:29
kgunnthomi: fginther basically moves the mir_socket to a diff dir, per security team21:29
kgunnunderstand there might be some hardcoded tests checking /tmp/21:30
kgunnjust a heads up21:30
thomikgunn: thanks21:30
thomikgunn: did the fix to remove stale sockets get landed?21:30
doanacplars: nice. thanks21:31
loolthomi: BTW the output I was getting from unity8 was like:21:39
lool22:36 < lool> debian phablet-click-test-setup phablet-config phablet-demo-setup phablet-dev-bootstrap phablet-flash phablet-network phablet-test-run phabletutils repo setup.py  tests (process:12162): CRITICAL **: Unable to get session bus: Error calling StartServiceByName for com.canonical.hud:  GDBus.Error:org.freedesktop.DBus.Error.TimedOut: Activation of com.canonical.hud timed out21:39
lool22:36 < lool> file:///usr/lib/arm-linux-gnueabihf/qt5/qml/Ubuntu/Components/MainView.qml:257: TypeError: Cannot call method 'hasOwnProperty' of null21:39
lool22:36 < lool> process-return-code: {{{-11}}}21:39
thomiO.021:39
loolthomi: do you have a Mir capable device?21:42
thomilool: yes21:43
loolthomi: would you want to reproduce?21:43
loolthomi: basically upgrade to unity8, upstart-app-launch, unity-mir, qtubuntu from ppa:ubuntu-unity/daily-build21:43
thomilool: I'm knee deep in debugging something else right now. If I get that done I can take a look, but I don't think I'm the right person for this - it looks to me like some odd security issue perhaps?21:45
thomip $esp21:45
thomioops, wrong window21:45
sergiusenslool, plars fwiw https://code.launchpad.net/~sergiusens/notes-app/click-fixes/+merge/18998022:47
sergiusenslool, plars this needs to land for notes to have passing tests22:47
loolsergiusens: mind reviewing https://code.launchpad.net/~lool/session-manager-touch/lightdm-depends-on-lxc-android-config/+merge/189981 ?22:49
loolsergiusens: empty line in .desktop is ok?22:50
loolsergiusens: please upload any time, landing #15222:51
sergiusenslool, upstream merger is disabled though22:53
loolsergiusens: session-manager-touch is upstream merged?  don't think so22:54
loolsergiusens: or you meant notes-app?22:54
sergiusenslool, oh, I meant notes app22:55
sergiusenslool, let me test this22:55
loolsergiusens: if you dont mind, mind testing https://code.launchpad.net/~lool/ubuntu/saucy/lxc-android-config/set-display-mir/+merge/189986 afterwards22:58
sergiusenslool, done with https://code.launchpad.net/~lool/session-manager-touch/lightdm-depends-on-lxc-android-config/+merge/189981 (I can't upload though, remember?)23:02
sergiusensrsalveti, can you ^^23:02
loolsergiusens: I can23:03
loolsergiusens: was looking for peer review23:03
rsalvetifine if lool can't take it23:03
sergiusensrsalveti, it's his MR ;-)23:03
rsalvetioh, sure then :-)23:03
rsalvetilool: sergiusens + and started lxc-android-config)23:04
rsalvetishouldn't be and started android?23:05
rsalvetilxc-android-config emits android23:05
sergiusensrsalveti, started on  is for proper upstart jobs23:05
rsalvetiright, not started on23:05
rsalvetibut on android23:05
rsalvetianyway, should work, but I know that job emits android as well23:05
loolrsalveti: whatever you think is cleanest, but I think on android will prevent lightdm restarts23:06
loolsince android is emitted once23:06
loolnot 100% sure23:06
rsalvetiright, but I don't think you ever want to restart lxc-android-config23:06
rsalvetirestarting container is probably a no-go23:06
loollightdm23:07
rsalvetiright, but why depending on lxc-android-config wouldn't?23:07
loolah right, it might try to restart the container23:07
loolrsalveti: didn't seem to restart lxc-android-config here23:09
looldon't ask me why23:09
sergiusenslool, what is the purpose of XXX in https://code.launchpad.net/~lool/ubuntu/saucy/lxc-android-config/set-display-mir/+merge/18998623:09
rsalvetilool: good then23:09
rsalvetilool: why a job for that?23:09
loolsergiusens: it's to be fixed if we fix boot hooks23:09
rsalvetishouldn't be part of livecd-rootfs?23:09
loolrsalveti: this will not work on upgrades23:09
sergiusensrsalveti, for people doing upgrades23:09
rsalvetiwhy not?23:09
rsalvetioh, it's home23:09
rsalvetithat's why23:09
sergiusensrsalveti, yeah, any writable location would had that issue23:10
sergiusenshave23:10
rsalveti29+start on boot-hooks and starting lxc-android-config23:11
rsalvetiwill this be enough for lxc-android-config to always get that file when starting itself?23:11
rsalvetiit should I believe23:11
sergiusensrsalveti, lxc-android-config won't start until this task is finished23:11
rsalveticool, upstart is that smart then23:12
sergiusensrsalveti, although not a task as the job says23:12
sergiusensrsalveti, that's the magic of 'starting'23:12
* lool uploaded ubuntu-touch-session_0.77_source.changes23:12
looloh crap23:12
rsalvetilooks fine then23:14
rsalvetiwe can revert this after we have a released image with it included by default right?23:14
rsalvetino, until we move that file somewhere else, anyway, this should be good23:14
asacplars: 87 is not getting better?23:17
asaclool: ?23:17
asacplars: notes app in particular23:18
loolasac: didn't look at 8723:18
asachmm23:19
asacnot good23:19
loolasac: I ran unity8 testsuite on SF and it passed here23:20
asaclool: unity8 is not the problem23:20
asacnotes-app23:20
asacthat was supposed to be fixed23:20
loolasac: it wasn't enough, sergiusens did another fix23:21
loolasac: https://code.launchpad.net/~sergiusens/notes-app/click-fixes/+merge/18998023:21
asaclool: why didnt we know?23:21
asacthats not really promissing23:21
loolsergiusens: I'm assuming notes-app wont be automerged, would you merge it by hand?23:22
asaclool: what is really bad is that we regressed our image without even landing mir23:22
loolasac: Yeah, so I don't know how to test the new click APs personally, I think we should do that before landing other clicks23:23
loolinstall the click by hand, run the tests23:23
loolplars, sergiusens: Is there an easy way to run the AP tests on clicks?23:23
asaci dont understand why we even move to more click apps if we cant test it23:23
asaccan we revert notes-app?23:23
asaci am sure its still in the ppa?23:23
asaci just dont want more click app transition noise23:23
loolasac: it's well isolated23:24
asaci dont know23:24
loolasac: I think it's good to decouple the image from the apps with this move23:24
asacfeels very wrong to be so lax about things.23:24
asacwith what move?23:24
asacmir?23:24
asacmir didnt even land :)23:25
asacthe thing is that we introduce something that will now create regressions on apps23:25
asacso we dont have notes-app in the game for mir team to look at23:25
asacsame for all the other apps that we regressed23:25
asac(still not sure what change exactly started our crashers and why we kept going, but well.)23:25
asacanyway23:26
lool== Published qtubuntu (platform stack) ==23:27
sergiusensasac, the crashers were still there23:28
sergiusensasac, just not exposed23:28
sergiusensso I don't understand how you can be against that23:28
sergiusenswe had a bad image that just looked good on the dashboard23:29
sergiusensasac, if no regressions is the case, mir is a big one23:29
asacwe had a missing test for click packages i feel23:29
lool== Publishing unity8 (unity8 stack) ==23:29
asacthat doesnt mean we have to put all the other apps in the spotlight as well23:29
sergiusensasac, it's a transition; not a thing forever23:30
asacanyway, i see a few with crashers23:30
asacand thats fine... notes-app is just still broken23:30
sergiusensasac, and by doing this we are finding a bunch of fallacies in our tools23:30
asacsergiusens: you can prep transitions outside23:30
sergiusensasac, not if you merge above23:30
sergiusensasac, not when I need support from the infra23:30
sergiusensasac, if it's not in the image, it doesn't get exposed so people won't fix btw23:31
asacthats exactly the attitude we try to fix23:32
sergiusensasac, how?23:32
asacits also not really true for things that have priority23:32
asacwe have seen lots of landings that we didnt do because they were not ready, and that could very well be prepped outside23:32
asacbreaking the image is not a mean to get priority23:32
asacif thats the case, we should fix that23:32
sergiusensasac, the image isn't broken btw, you can write notes still, it's just the tests23:33
asacwell, if the APs are broken you cant merge anymore23:33
sergiusensok23:36
cjwatsonlool: It's supposed to be phablet-click-test-setup23:39
cjwatsonI haven't actually tried it myself23:40
cjwatson(I think, anyway)23:40
loolok23:40
sergiusenscjwatson, lool yes, that's how I got the results in the MR btw23:40
loolasac: I think we have a serious issue with new upstart-app-launch23:41
sergiusensafter installing you need to run 'phablet-config autopilot --dbus-probe enable'23:41
loolcan someone please test upstart-app-launch in PPA and check whether you can open OS updats from settings?23:41
loolI'm trying to Mir23:42
loolfrom23:42
loolso it worked in Mir, and it worked on reboot23:45
loolok, upstart-app-launch works for me23:49
loolgood news is that we get click appearing on instal now23:49
looldobey, ralsina: Seems I can't uninstall preinstalled click again; looks like an UI thing though23:50
loolsergiusens: do you have a fixed notes-app.click I can try with phablet-click-test-setup?23:52
lool== Publishing upstart-app-launch (misc stack) ==23:58

Generated by irclog2html.py 2.7 by Marius Gedminas - find it at mg.pov.lt!