[00:09] doanac: plars: ok pleas shoot me and dirocks etc. a mail what happened. 'night [02:51] Is there anyone around that is able to tell me/find out why I can't see a CI job for this approved MR? https://code.launchpad.net/~veebers/unity8/adding_unlock_emulator/+merge/191328 [02:53] veebers: what do you mean can't see a CI job? [02:54] cjohnston: hmm, actually I might be wrong, I was expecting there to be an autolanding job, but was looking at the wrong one (i.e. -ci) and the autolanding is disabled currently [02:55] ok [02:55] cjohnston: so a different, better, way to ask my question is: What do I need to do to take that MR from Approved to Merged? [02:57] veebers: what is causing all the things to be unstable? [02:57] I'd guess that needs to be fixed [03:02] cjohnston: ack, cheers [07:51] asac, doanac, i think it would be clever to wait until the session is completely up before running systemsettle ... after all we want to test the user experience [07:51] (i.e. make it check if unity8 is there before firing) [07:54] ogra_: do you mind refreshing me on what is systemsettle already? [07:55] didrocks, it takes ten samples from top and waits if the system goes >97% idle [07:55] apparently we run it directly at boot [07:55] ah ok ;) [07:55] running after the session is up seems like a better usecase to me [07:56] yeah, makes sense to have multiple tests, one to know how long we take to have everything started [07:56] and another, to know the idle when the session if fully loaded [07:56] well i want to introduce bootcharts into the tests [07:57] so that part should become more fine grained anyway ... if you want to see the load on boot and how much something takes [07:57] right [08:01] * ogra_ grins seeing lool's comment on the bonita thread [08:02] ogra_: well, for all java/jboss webapps, I don't expect less time even on a server to start TBH :p [08:03] i meant the "luckily we didnt disable swap" :) [08:03] ahah [08:03] we had so many discussions about that before release ... [08:03] yeah [08:03] also, we didn't disable adb [08:03] told you! ;) [08:03] there is a usecase ... for the crazy ones [08:03] didrocks, well, thats just because i ran out of time [08:04] it was actually planned to habe it off by default so you need to enable it from the terminal first [08:04] *have [08:04] ogra_: I know, I just told you we won't have time :) [08:04] true [08:04] ogra_: terminal or system settings? [08:04] i'Äm to optimistic at times === psivaa-afk is now known as psivaa [08:04] terminal [08:05] there was no UI planned for dev mode [08:05] interesting, we should maybe [08:05] (we should do that for 14.10 in any case= [08:05] ) [08:05] err [08:05] 14.04 [08:05] ;) [08:05] so you always live in the future [08:05] and ev always lives in the past (he keeps telling 12.10 :p) [08:06] bah so i uploaded an internet radio app and now my BT speaker inst found :( [08:06] for some games as well it seems? [08:06] and some news sites [08:06] yep [08:06] and a TV program planner :) [08:07] ahah :) [08:07] and we have 2 xkcd apps [08:07] we might end up havfing 10 ... :) [08:07] right ;) still 0 direction/driving apps [08:07] but 2 xkcd apps ;) [08:07] * ogra_ expects the same chaos as googlew play has over time [08:08] well we have OSM ... [08:08] not with firections though [08:08] *directions [08:08] yep [08:08] and google maps works as well [08:08] (with directions) [08:09] webapps are just so easy (creating the click package takes less than 10min if you have a template app ... (of which 9min are used to find a proper icon) [08:09] google maps? [08:09] we do have a wrapper for it? [08:09] or not yet? [08:09] sure, go to maps.google.com [08:09] ah ok [08:10] I'll try to exercise that as a click app as well [08:10] today [08:10] it has some user agent issues so the page flips to desktop at times [08:10] better that we learn our technology :) [08:10] http://daker.me/2013/10/package-your-webapp-for-ubuntu-touch.html [08:11] thats an intresting hack to force the user agent for a page [08:11] (though thats far more effort than i had to do for my apps) [08:11] ogra_: excellent! will try that :) [08:11] * didrocks finishes first the morning tasks [09:02] ogra_: didrocks: what time do we want to build the image? [09:02] :) [09:02] now or later? [09:03] asac: as we are unsure that the dashboard is going to pick the right image (and to not confuse result, I would prefer to have the change before we have the image built), I would like to wait for doanac "go" on the dashboard fix [09:03] asac, evening i'D say [09:04] (independently from the dashboard, the more syncs we can get in the better so we have them cross tested) [09:04] didrocks: hmm. its not about the dashboard fx, but making a reasonable soon snapshot to make the set of changes more managable [09:05] asac: well, without have dashboard results, we don't know of the real quality of the AP tests results [09:05] i dont understand why we wouldnt build a new image now and then whenever we feel we need more as well, but *shrug* [09:05] didrocks: we dont. so we produce the snapshots now and then get the results later [09:05] so we can just say "well, it boots" [09:05] didrocks: right. its for post-testeing [09:05] asac: I mean, it will run the tests against 101 [09:05] so we have bi-sect points in history in case that stuff is broken [09:06] didrocks: ah... well, we can surealy ensure that we test the right stuff [09:06] after [09:06] and I'm unsure we have a way to reflash/relaunch a whole set of tests again [09:06] just produce the images now and we can test them once the dashboard is fixed [09:06] didrocks: we have [09:06] if you are sure we can, yeah, +1 for now [09:06] yes i am sure that we even did that in the past [09:06] asac: but I'm sure we are going to screw old results with new ones [09:06] but let's see… [09:06] asac: we relaunched some tests [09:06] what is tricky is testing and image that is not the latest, but that is also possible [09:06] in the past [09:07] not changed the image :) [09:07] ogra_: ok, so let's kick an image, we'll see… [09:07] didrocks: i hope the infra has matured a bit [09:07] * ogra_ goes and kicks [09:07] @ci, with the 13.10 freeze over, would it be possible to get the patches https://code.launchpad.net/~renatofilho/qtorganizer5-eds/fix-match-all/+merge/191080 and https://code.launchpad.net/~charlesk/indicator-datetime/lp-1233176/+merge/190009 in the phone image? [09:07] I need those two patches to verify the working of alarm [09:08] asac, didrocks, build #2 running [09:08] lets see if it even survives :P [09:08] nik90: I think we'll need a landing ask (if not already) for that one [09:09] didrocks: I dont think I have the permission to do so..I will ask popey to add a landing task for it. [09:09] nik90: and we'll start resuming landing tomorrow morning (finishing up transitionning today to trusty) [09:09] thanks nik90! [09:09] didrocks: okay :) [09:09] nik90: didrocks i dont have access either [09:10] * didrocks gives popey power [09:10] noooooooooooooooooooooooo [09:10] ☻ [09:10] popey: powered up! :) [09:12] asac, what about sending a call for MIRing stuff to the MLs. i think we should have people start getting their stuff into main so we dont have to care about that at the end of the cycle [09:13] (and the MIR team can process the requests over the cycle instead of getting them all at the last minute) [09:13] ogra_: MIR team == mterry though [09:13] yes :) [09:14] thats what i mean ;) === vrruiz_ is now known as rvr [09:16] kalikiana: hey, answered on landing ask #185, can you give us more details? [09:16] kalikiana: I wonder why it worked for you, there is surely a way you tested it differently from us and I want that we fix that gap :) [09:22] didrocks: not sure what more detail you're thinking of. you got a bug number and expected fixes [09:22] kalikiana: I mean, you filed a first landing ask [09:22] kalikiana: so I intended it was working for you [09:22] how was it working? did you disable apparmor? [09:22] I didn't disable it that I can tell [09:23] kalikiana: so, you were getting sensors feedback? the toolkit/apparmor regressed you? [09:24] we try to identify what you differently than us to get sensors feedback working when you requested the landing ask, so that we can write better guidelines on testing [09:27] my assumption was apparmor changed, though jdstrand told me it didn't have a policy for the vibration before - the next thought would be android layer changes… but I don't know much about it [09:28] is it possible that the container stuff changed permissions? [09:28] kalikiana: well, can you check with jdstrand? because we tested in the same image than you did I guess (it was the day you added the landing ask) and didn't get it working [09:29] so maybe it's an application test diff, some confined, some not confined, but would be great that we identify it [09:29] didrocks: I repeat I wasn't able to test with that particular image to begin with [09:29] I had an older one when I tested before the ask [09:30] I ran into tooling problems preventing me from testing anything useful but finding unrelated bugs [09:30] kalikiana: hum, so, before filing the landing ask, maybe you can check with us if this reproduce? [09:30] as we were able to test, we can give advice maybe :) [09:32] didrocks and/or other CI people: Why indicators linked to bug #1241539 haven't been released into trusty yet? =) [09:32] bug 1241539 in indicator-sound (Ubuntu) "ubiquity-dm is missing keyboard, input, sound, system indicators" [High,Triaged] https://launchpad.net/bugs/1241539 [09:32] well Mirv was testing also [09:32] I guess you would have prefered not to add it to the sheet before either of us confirms it with the latest image [09:32] kalikiana: right, it's the deal to have a fast landing flow [09:33] xnox: well, we are deploying the switch to T. See the emails by asac, we'll get first landing tomorrow [09:33] didrocks: on which mailing list? [09:34] ubuntu-devel? [09:34] asac: do you remember where you did send that? ^ [09:34] hum, seems only ue-leads, that would explain… [09:35] xnox: ue-leads [09:35] didrocks: ^^ [09:35] didrocks: i see one to ubuntu-phone just now. [09:35] you can forward [09:35] xnox: thats one without timeline. [09:36] xnox: forwarded you the initial mail [09:36] * asac will stop using ue-leads for high level allignment as managers seem to not forward such info to their teams :) [09:38] It'd be better to use a public list, really [09:38] where possible etc [09:38] asac: i think ubuntu-engineering@l.c.c is a good one and/or ubuntu-phone@ even better. [09:39] cause everyone is subscribed to ubuntu-engineering@ [09:41] Everyone in UE at Canonical [09:48] asac: didrocks: the smoke is now running trusty image now. the results are starting appear in the dashboard [09:48] psivaa: really? wowo!! [09:48] with image 1 [09:48] psivaa: excellent! [09:48] ncie one [09:48] http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/1:20131021.1:20131015/4783/ [09:48] psivaa: will maguro also go ahead? [09:48] asac: i'll look at that [09:50] plars: doanac: so if i run top -b -d 3 i constantly get better than 97.5 [09:50] plars: doanac: so if i run top -b -d 5 i constantly get better than 98.5 even [09:51] plars: doanac: so maybe we should revisit our parameters used [09:51] plars: doanac: e.g. closer to the ones i initially proopsed than the very quick/fast we have now [10:00] sil2100: Mirv: so, we can build all trusty now? [10:02] Mirv: did you redeploy everything? (I did qa) [10:03] Let's keep our fingers crossed that we didn't miss anything ;) [10:04] load average: 2.19, 1.01, 0.39 [10:04] hmpf [10:04] i'd really like to know why our load is so high all the time [10:05] sil2100: ok, autorun launched [10:05] Mirv: damn, my qa merge didn't get in (but all is redeployed) [10:05] (there is no obvious process in top consuming anything) [10:12] didrocks: sil2100: yes, and let's see how it goes now [10:14] Mirv: huh, I can't redeploy the SDK stack for saucy it seems - ERROR Can't set the target branch ~ubuntu-sdk-team/qtcreator-plugin-ubuntu/trunk-2.7 for lp:qtcreator-plugin-ubuntu/2.7 [10:15] Mirv: I'm not a member [10:15] I guess we need to ask them to add all of us, the ubuntu-unity team as members [10:24] sil2100: yeah sdk team was made more strict lately because the trunk was messed up. I'll add individuals. [10:28] ogra_: hum, I click install my own package and don't see it in the installed app [10:29] is there any trick? (I already tried rebooting) [10:30] ah, click register surely [10:31] hum, not apparearing after that, let's reboot [10:32] ah, a reboot did it :) [10:33] webbrowser app test failing? is that a known flaki one? [10:33] http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/101:20131018:20131015/4782/ vs. [10:33] http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/1:20131021.1:20131015/4783/ [10:34] didrocks: unity8 also as results not meeting what we expect, right? [10:34] http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/1:20131021.1:20131015/4783/unity8-autopilot/ [10:34] asac: no, we didn't release any other ones [10:34] asac: see my answer to doanac :) [10:35] so it's known to be screwed until we release one [10:35] didrocks: release == new unity8 fix? [10:35] yep [10:35] ok. so webbrowser? [10:35] those are supposed to be good [10:36] didrocks: please don't use click install (unless you're very careful with its arguments), use pkcon install-local [10:36] didrocks: the click manual page has advice [10:36] cjwatson: ah, ok will do then :) [10:36] didrocks: yeah. http://reports.qa.ubuntu.com/smokeng/trusty/touch/mako/1:20131021.1:20131015/4783/webbrowser-app-autopilot/ [10:36] didrocks: (I'd point to manpages.ubuntu.com but it doesn't have 13.10 yet ...) [10:37] But "This is a low-level tool; to install a package as an ordinary user you should generally use pkcon install-local PACKAGE-FILE or some higher-level user interface instead, which take care to use the correct set of options. (Do not use sudo when invoking pkcon, as it needs to know the calling user.)" [10:37] cjwatson: I guess that we should have the tutorial as well pointing to us (I only used --help TBH) [10:37] to pkgcon* [10:39] didrocks: the help output does point to pkcon [10:39] didrocks: http://paste.ubuntu.com/6282270/ [10:39] cjwatson: I just click --help [10:40] and figured out the parameters then [10:40] didrocks: which points to "click install --help" for more on that specific command, which points to pkcon [10:40] cjwatson: right, but I didn't think that care was needed, so I just click --help and then click install [10:41] sigh [10:41] OK, I'll add a note to the top level as well [10:41] cjwatson: well, maybe I'm just stupid thinking that click install was safe and didn't need to read something else than click --help [10:44] didrocks: + * Adjust top-level "click help" entry for "install" to point to pkcon. [10:45] cjwatson: thanks ;) [10:52] didrocks: maliit is crashing :) [10:52] still; [10:53] asac: yeah, not news, we knew it, this a crash somewhere in the stack, the keyboard guys were pinged about it [10:53] ok so no regression... kk [10:55] ogra_: i think maguro flashing is taking longer than usual with trusty image 2. looks like more than 25 mins. i'll run once more to find out the exact time [10:55] ok [10:55] well, it is usually between 20 and 30 for me [10:56] ogra_: you are counting download time for you, right? [10:56] didrocks, no plain flash times [10:56] ogra_: can you see if the unity8 upstream merger a) run their AP and b) if it passes currently? [10:56] ogra_: waow… [10:56] err [10:56] didrocks: ^^ [10:57] ogra_: sorry [10:57] asac: upstream merger? we are not upstream for maliit [10:57] http://people.canonical.com/~ogra/touch-image-stats/20131022.changes [10:57] didrocks: talking about unity8 itself ... sorry :) [10:57] diff between 1 and 2 [10:57] (FYI) [10:57] didrocks: was back on the "unity8 is failing, but fix is in trunk - most likely" topic [10:57] didrocks: i figured that if its fixed, the upstream merger should give green on those [10:57] but now i dont know if folks have again turned the AP off :) [10:58] Saviq: https://code.launchpad.net/~unity-team/unity8/trunk [10:58] Saviq: I don't see your fix btw, you told me on Thursday you merged it, right? [10:58] didrocks, fix for? the autopilot tests? [10:59] Saviq: yep [10:59] didrocks, http://bazaar.launchpad.net/~unity-team/unity8/trunk/revision/473 [10:59] didrocks, asac [10:59] oh, it was at Christopher's name [10:59] but we need unity-mir released first [10:59] and then drop unity8-setcap.conf from lxc-android-config [10:59] or well, not for image testing [11:00] Saviq: ? yeah, we are only talking about the tests [11:00] Saviq: what does UNSTABLE: http://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/2564 mean? [11:00] test failed? [11:00] why do you merge anyway? :) [11:00] Saviq: https://jenkins.qa.ubuntu.com/job/generic-mediumtests-touch/3025/console [11:01] unmet dependency on the one I picked [11:01] didrocks, look up [11:01] i cant really parse this ... https://code.launchpad.net/~veebers/unity8/ap_env_fix_1240801/+merge/191567/comments/440420 [11:01] asac, we're not [11:01] let's look at a MP older then [11:01] so we have some SUCCESS: ... on armhf -> SUCCESS: http://jenkins.qa.ubuntu.com/job/generic-mediumtests-builder-saucy-armhf/3023 [11:01] didrocks, we never got mediumtests-touch to a working state with mir yet [11:01] ah thats building [11:02] asac: saucy is desktop AFAIK [11:02] Saviq: it was on the image, apart from 1 failure, right? [11:02] didrocks, yes [11:02] Saviq: well, the idea is that after your merge [11:02] it works [11:02] didrocks, the failure was a random crash, too [11:02] so you shouldnt merge it if it doesnt :) [11:02] 8 is more though [11:02] asac, really, please stop [11:02] ah [11:02] ok [11:02] asac, I'm telling you what's the prerequisite for unity8 even *starting* [11:03] and that is unity-mir release [11:03] Saviq: juts to understand on that one: https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/2575/? [11:03] it's your fix [11:03] didrocks, asac please stop [11:03] I guess it was ran without this unity-mir dep [11:03] and listen [11:03] yeah lets listen [11:03] sorry Saviq. go ahead [11:03] we need to release unity-mir, so that we can drop unity8-setcap.conf from lxc-android-config [11:03] sil2100: unity-ubuntu now part of the sdk team (the 11 team members were ones that I would have added individually as well) [11:03] that will let us install unity8 in mediumtests-touch [11:04] and only then we can start looking into actually fixing those runs [11:04] I trust ~ubuntu-unity members not to push --overwrite to ui-toolkit repo :) [11:04] ok sounds good [11:04] Saviq: do we have all pieces ... in theory? [11:04] and looking into them if they didn't pass [11:04] Saviq: I still don't undestand why having just the code executed in https://jenkins.qa.ubuntu.com/job/generic-mediumtests-runner-mako/2575/? (so your fix), I was able to have all tests (but 1) passing on the image [11:04] asac, yes, unity-mir just needs a release [11:04] and lxc-android-config? [11:04] without any unity-mir/dropping unity8-setcap.conf [11:05] didrocks, those failures were intermittent - now fixed in trunk [11:05] Saviq: the 8 of them? [11:05] didrocks, yes [11:05] we just got really unlucky? [11:05] ok [11:06] didrocks, intermittent as in we fixed them [11:06] ok [11:06] didrocks: guess we should make an ask for this and do this as one of our first things. seems a few components are involved that require coordination :) [11:06] asac: indeed, I would be interesting if such ask was made to make it a priority [11:06] lxc-android, unity8, unity-mir [11:06] knowing that will dep on a mir ABI bump which is already in the ppa [11:06] well, we didnt tell folks if they shyould continue doing asks [11:06] so we'll need mir as well [11:06] now the mail is out [11:06] so... :) [11:06] as it's built against it [11:07] and so platform-api for this mir abi bump [11:07] as well as u-s-c [11:07] didrocks: lets create this one for saviq [11:07] asac: ? [11:07] I disagree [11:07] well, your call :) [11:07] otherwise, there is no value, we just add more paperwork to our own soul [11:07] for the sake of it [11:08] and we aren't sure to have all pieces listed [11:08] nor the impact [11:08] Saviq: can you add an ask about fixing unity8 APs and give details about all the bits and pieces so we can coordinate getting all in? [11:09] thanks [11:18] didrocks: ogra_ i am seeing some issues with flashing with trusty image 2. [11:19] i get 'ERROR:phablet-flash:Installation is taking too long or an error occured along the way.' [11:20] the installation on my local maguro is hung on boot screen - Google and Padlock [11:21] psivaa, doing an OTA upgrade here atm ... lets see how that behaves [11:21] psivaa: interesting (and worrying) [11:21] ogra_: there was no hybris change right? [11:21] http://people.canonical.com/~ogra/touch-image-stats/20131022.changes [11:21] so can't be an android/ubuntu mismatch? [11:21] are we planning to stop this landind-asks business at some point please? [11:21] i dont see one [11:22] Saviq: see asac's email [11:22] Mirv: thanks! [11:22] didrocks, lightdm and libpam are new ... [11:22] as well as lxc [11:23] didrocks, yeah, didn't find an answer to that question (other than "we'll discuss in OAK") [11:23] they look like the most possible candidates for the session not starting [11:23] Saviq: if there was a decision before the discussion, that wouldn't be a discussion [11:23] well said [11:23] effective diff to pam is pretty small [11:24] cjwatson, yeah, let me test first to see whats up here [11:24] lightdm or lxc seem more probable [11:24] (neither was an autosync) [11:24] both just are obviously involved in session startup [11:25] knowing that lxc dropped a fix we still needed yesterday for otto [11:25] not sure if this one was useful for touch [11:25] * ogra_ twiddles thumbs watching the android with rotating guts [11:25] - 0006-add-pstore-to-container-fstab.patch [11:25] we really should switch these animations to something ubuntuish :) [11:25] ogra_: we needed still that one for otto yesterday ^ [11:26] not sure if it's relevant to you :) [11:26] didrocks, i thought that was upstreamed [11:26] ogra_: oh possible (we ship our own fstab for otto, that's why we noticed) [11:26] * didrocks looks at the diff [11:26] yep, in -Index: lxc-1.0.0~alpha1/templates/lxc-ubuntu.in [11:27] ok,, seems the container is up fine [11:27] psivaa, confirmed btw [11:27] ok, so not lxc :) [11:27] right [11:27] * ogra_ tries a manual start lightdm [11:27] ogra_: thx [11:27] I'd stop autosyncs but they aren't going to run for 4.5 hours anyway. Let me know if it's needed [11:28] [+0.34s] DEBUG: Seat: Session authenticated, running command [11:28] [+0.34s] DEBUG: Registering session with bus path /org/freedesktop/DisplayManager/Session0 [11:28] [+0.35s] DEBUG: Session pid=884: Running command /usr/sbin/lightdm-session ubuntu-touch-session [11:28] [+0.36s] DEBUG: Session pid=884: Logging to .xsession-errors [11:29] the lightdm log looks fine as well [11:29] root@ubuntu-phablet:/# ps ax|grep 884 [11:29] 884 ? Sl 0:00 lightdm --session-child 10 13 [11:29] and lightdm runs [11:30] * ogra_ digs into session logs [11:30] asac, didrocks, ask added [11:30] phablet@ubuntu-phablet:~$ start unity8 [11:30] start: Job läuft bereits: unity8 [11:30] thanks! [11:30] oh wow [11:30] thanks Saviq [11:30] phablet@ubuntu-phablet:~$ ps ax|grep unity [11:30] 1565 pts/3 S+ 0:00 grep --color=auto unity [11:30] ogra_: FYI, your job is running away to unity 8 =) [11:30] lies ! [11:31] * xnox somehow finds technical translations to other languages amusing =) [11:31] ogra_: status unity8? [11:31] terminate called after throwing an instance of 'boost::exception_detail::clone_impl >' [11:31] what(): Could not open hardware module [11:31] all xnox' fault !!! [11:32] ogra_: right, let me check if unity/mir was rebuild with new boost. I thought I did. [11:32] xnox, well probably you did but it wasnt let out of CI [11:32] didrocks, ^^^ [11:33] ogra_: I guess xnox is talking about directly uploads [11:33] didrocks: ogra_: well mir was rebuilt, but not unity-system-compositor. [11:33] u-s-c isn't used on touch [11:33] didrocks, no idea i thought mir and unity were not direct [11:34] ah, then no idea. [11:34]