[00:01] robru: I see that the silo12 where ralsina was building unity-scope-click is stuck: Can't merge: One package at least is not available at the destination. unity-scope-click (0.1+14.04.20140220.1-0ubuntu1) is in the proposed pocket. [00:02] alecu, yes, that package is stuck in -proposed due to a failed autopkgtest. [00:03] robru: where can I get more info on that failure? Is there any way I can help move it forward? [00:03] alecu, https://jenkins.qa.ubuntu.com/view/Trusty/view/AutoPkgTest/job/trusty-adt-unity-scope-click/ARCH=amd64,label=adt/28/console i don't know if you're able to resolve that. i don't know anything about that package or it's tests [00:03] great [00:03] alecu, you'd normally find that log by visiting http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html and searching for the package name. [00:04] alecu, if you can submit an MP that fixes that crash, I can add it to the landing, rebuild it, then you can re-test it, then we can re-publish it, and hopefully then it'll pass -proposed and make it to distro. [00:06] robru, mmcc has a branch that fixes it [00:06] mandel, oh, do you have a link? [00:07] robru, I've asked for the MP to review, yes [00:07] robru, give us a few mins [00:07] mandel, ok [00:07] mandel, alecu: also it's technically ralsina's job to take the new MP and put it in the list, but if he's not around then i can do it too. [00:08] robru, he is not, left a while back [00:08] mandel, ok no worries [00:08] robru: that will help us greatly, since he's currently on a plane [00:09] robru: thanks a lot [00:09] you're welcome === dobey_ is now known as dobey [00:30] robru, as soon as we have the +1 from alecu we can use https://code.launchpad.net/~mikemc/unity-scope-click/fix-udm-abi-breakage/+merge/207578 [00:30] ok, alecu ping me when you're satisfied [00:31] robru: +1 to that branch, let's land it on the silo and test it [00:31] ok [00:34] alecu, silo num? [00:34] mandel: 12 [00:34] ack [00:35] alecu, mandel, mmcc: just started building, you can watch the log here: http://162.213.34.102/job/landing-012-1-build/14/console [00:36] once that gets to the end you can start testin [00:39] yay! [00:39] I'll test it after dinner [00:39] mandel: you can go get some sleep [00:39] alecu, I want to stay at least until it builds :) [00:39] robru: thanks for working on this [00:40] mandel: I know it builds, because I tested the same steps locally, and both mmcc and you have also [00:40] robru, yes, thx!!! I'm really sorry for forgetting about ABI compatibility :-/ [00:40] alecu, nevertheless.. [00:40] alecu, you're welcome! I'm not even EOD yet... this is why I'm here! [00:40] mandel: and I even got the exact same errors without this fix, so I'm pretty confident it works [00:41] alecu, yes, the errors are clear == ABI because adding virtual methods screws it.. [00:41] alecu, worst thing, I have read this http://techbase.kde.org/Policies/Binary_Compatibility_Issues_With_C++ several times :-/ [00:41] alecu, I was very very very stupid [00:42] mandel: is there a lint tool that can help with these mistakes? [00:42] mandel: if not, food for thought [00:42] alecu, no that I'm aware off.. but you have a point, it would be nice to get a tool that those a diff, checks for changes in headers and screams at you [00:43] surely that must exist? you guys can be the first people to steward an ABI. [00:43] can't [00:44] robru, we are not, unity has done it a couple of times.. [00:44] alecu, robru http://ispras.linuxbase.org/index.php/ABI_compliance_checker [00:45] not ideal.. but is a start [00:47] mandel: why not ideal? it sounds *just* like the thing to avoid these issues [00:48] mandel: it's even been developed for quite some time, and it's also fresh: "last modified on 23 January 2014" [00:48] alecu, I just don't like the xml and I'd like it to be a bzr hook :) [00:48] alecu, but it is a start :) [00:49] mandel: it would have caught this exact issue: "added/removed virtual functions (change of a v-table layout)" [00:49] alecu, yes, it would have, I knew.. so I should have not propose the merge.. [00:50] alecu, I'm going to see if it is packaged, if not, do it and later add an automated check in make check [00:50] mandel: so, a bzr hook sounds right. Also, the xml would be generated by the hook, not by you. [00:50] alecu, one should be added as the supported one or something, right? [00:50] alecu, or.. bzr hook and always generate the one from the previous branch [00:51] revno sorry [00:59] mandel: either a bzr hook, or something that gets run by jenkins [00:59] alecu, jenkins is a better approach [00:59] alecu, since it will work for projects that do not use bzr [00:59] mandel: which checks the current version, and the previous version. And that fails if the package major version is not increased [01:00] alecu, or deb packaged updated [01:00] alecu, and talking about that, the getAll methods will have the same issue, I'll have to do an update break thing for that [01:01] mandel: right. Let's do it at some safe point next week [01:01] alecu, exactly, if it not yet needed, lets not do it this week, is risky [01:01] mandel: we need that for a bug that's not critical, so next week is fine [01:02] alecu, yep, I chat with this with mmcc and he mentioned too that is not urgent, the branch is there so it is a matter of deployment [01:02] alecu, the easiest, use the same silo for both projects and fix the deb control file to state a break [01:03] * mandel needs to read what exactly has to be written there [01:03] mandel: both in the same silo sounds good [01:04] mandel: what about other packages that depend on this client library? system-updates, perhaps? [01:04] alecu, guarantees that abi will be ok at least for the scope, who is the only known client... [01:04] ah, great [01:04] alecu, system-updates uses python, so it goes bia dbus [01:04] via* [01:05] fginther, hey any help here? Build timed out (after 60 minutes). Marking the build as failed. [01:05] fginther, https://jenkins.qa.ubuntu.com/job/autopilot-testrunner-otto-trusty/2971/console [01:06] fginther, but it just dies during env setup it seems [01:08] or robru ^^ [01:09] anyone who knows someone is good :-) [01:09] that's fginther territory, sorry [01:09] I'd just skip otto :-/ [01:12] robru, is fginther the only one that knows otto? [01:14] sergiusens, probably not, but i don't know of anybody else. [01:14] sergiusens, i guess you're supposed to ping cihelp for that [01:18] alecu, robru is the build supposed to take this long? [01:18] mandel, it says it's done to me? [01:18] mandel: it is, yes [01:19] mandel: it just finished! [01:19] alecu, robru yeah, I just mentioned and finished... had I known that I would have said it earlier [01:19] robru, well no need to hurry as if I land this; I'd still need cihelp to change the test runners :-/ [01:19] 40 minutes seems normal to me. certain projects take way longer though [01:19] alecu, cwayne can also test it for you [01:20] * sergiusens is just randomly ranting [01:20] robru, I suppose, is 2:20 am here, so I was anxious [01:20] sergiusens, quejica! [01:21] mandel, me tiene las pelotas llenas! antes tenia mas poder ;-) [01:21] sergiusens, lol [01:27] robru, left a comment in the train sheet; hopefully everything can go in in the AM [01:27] sergiusens, ok [01:28] thanks! [01:29] mandel, mmcc alecu: how's silo 12 going? ready to publish that yet? [01:32] robru: omg [01:32] ? [01:32] gonna kick an image before autopilot lands :) [01:32] oh, good idea. [01:32] ;-) [01:33] hehe [01:33] robru, I'm calling it a night (2:30 am) but alecu can give you the green light [01:33] mandel, ok thanks [01:33] robru, cwayne in #ferret is also testing it [01:33] cyphermox, surely they wouldn't have marked it tested if it wasn't flawless, right? [01:35] I'm reflashing the phone; it will take at least 20 more minutes, and will install the ppa afterwards. [01:35] I'll wait while having dinner [02:47] alecu, any word on those tests? [02:49] robru: I just finished installing, will start testing right now [02:49] ok [02:49] robru: cwayne has been testing it [02:49] I'm past EOD but I'm waiting around just to publish this, then I need to run out [02:50] right, so it fixes the issue of the scope not loading at all, which is good, but installing apps seems to have some issues [02:50] robru: and this solves the empty scope for him, but the install is not working all of the times [02:50] hmmm [02:50] namely, sometimes installing apps causes unity to crash [02:50] so overall, it is definitely an improvement over no scope at all :) [02:50] cwayne, is that a regression or has it always been like htat? [02:50] ah [02:50] robru, i don't know, because the first version with these installation bits in it wouldn't load at all [02:50] just wondering what your acceptance criteria for publishing this is [02:50] so this is the first i've been able to test [02:55] i'd be inclined to say this is 100% an improvement over what's in universe right now :) [02:57] cwayne, alecu : ok, so I'll publish it with the understanding that fixes are coming? [02:58] I'm unable to add the ppa and install from there, so I manually installed the .deb from the ppa. [02:58] the scope crashes after trying to install something here [02:58] and unity8 sometimes dies too [02:59] alecu, so ... i shouldn't publish it? mixed signals ;-) [02:59] robru: I don't like the state, but I agree with cwayne that it's better than not being able to start any app [03:00] alecu, ok so i'm publishing then ;-) [03:00] robru: great, thanks [03:00] also, thanks for staying past your EOD for this [03:01] alecu, no worries. [03:01] alecu, cwayne : ok it's published now. [03:01] robru, thanks [03:02] alecu, I gotta run out and take care of some stuff, but I'll be back in 2-3 hours. if it turns out that you submit a new MP I can assign a new silo for you. After that it has to wait for EU morning, which is ~4 hours from now [03:03] cwayne, alecu : you're welcome! [03:03] robru: don't worry: it's 12am here, so I'll just write a mail so somebody on my team on EU morning time can start working on it [03:04] ok [03:05] ciao! [03:05] * cyphermox signs off too [04:09] and now to wait til it reaches trusty, so that i can kick a build and go to bed... [04:10] cwayne: what are you still waiting to land? [04:11] unity-scope-click [04:12] right [04:13] yeah, its in -proposed now, took awhile last time, for trying to build on arm64 i think [04:18] yeah, before FF [04:18] we usually get tons of uploads [04:19] ah, yeah [04:19] makes sense [04:19] damn [04:23] rsalveti, is there an easy way to see where it is in the queue [04:23] cwayne: rmadison unity-scope-click [04:24] cwayne: that usually tells you if it is already published in our main infra [04:24] right, that shows its in -proposed [04:28] rsalveti, i guess i'm just not well-versed on what actually happens to get it from trusty-proposed to trusty-released [04:29] if you use trusty-proposed you might get additional dependencies/packages as well [04:30] yeah, i don't wanna do that [04:47] hm, still in proposed [04:48] hm, https://jenkins.qa.ubuntu.com/view/Trusty/view/AutoPkgTest/job/trusty-adt-unity-scope-click/ [04:48] auto package test is failing [04:49] that seems less than great [04:49] yeah, not the same [04:49] http://people.canonical.com/~ubuntu-archive/proposed-migration/update_excuses.html [04:50] maybe it might still be running in the private instance [04:50] yeah [04:50] http://d-jenkins.ubuntu-ci:8080/view/Trusty/view/AutoPkgTest/job/trusty-adt-unity-scope-click/33/console [04:51] ah [04:51] so at this point should i just give up and make a new build tomorrow? [04:52] yeah, or just get it published in your ppa [04:52] i dont even think i have upload rights to the ppa, heh [04:52] ill just wake up early and do it tomorrow [04:52] alright [04:54] unless its almost done, but i dont think it is [08:07] robru: thank you for pointing people in the direction of cihelp. We're trying to get to the point where everyone on the team is capable of supporting any part of our infrastructure, and Otto is one area where we've made good progress (because we've had to deal with it a lot ;) ). [08:17] geh, rat emergency, brb [08:46] Back [08:47] sil2100: hey! [08:47] Damn, that rat just HAD to bite off all her stitches [08:47] :/ [08:48] They're sewing her up again ;/ [08:48] can you prevent them in any way to do that? [08:48] I ain't falling for this one again, I'll bandage her up once she's back here, she'll be unhappy but at least not bleeding to death ;) [08:48] I'll make a mummy out of her [08:50] heh :) [08:50] sil2100: now that you've acted on emergency, there is another one I guess [08:51] sil2100: so, we are seeing a lot of flakyness in the 3 last images [08:51] sil2100: can you use your phone and rerun all AP tests locally? (or at least, some, to confirm you see that flakyness) [08:51] then, I guess we'll have to revert one by one http://people.canonical.com/~ogra/touch-image-stats/20140220.2.changes [08:51] I see 3 potential failures: [08:51] - sdk [08:52] - unity-scope-mediascanner/libmediascanner [08:52] - indicator-datetime [08:57] Ok, let me see [08:57] * sil2100 was trying to resolve one citrain problem in the meantime [08:58] citrain problem? a bug? [09:00] hmm, I need to double-check, but maybe it's a bug in the reconfugire [09:00] Ah [09:00] No, it seems to have been just robru's small error [09:01] didrocks: in the meantime, let me upgrade to latest and do the reverts [09:01] I'll start off with SDK [09:03] sil2100: hum, first, do you mean starting some applications which fails? [09:03] to ensure you can reproduce the issue :) [09:03] mind* [09:03] (without any revert) [09:06] Of course ;) === ev changed the topic of #ubuntu-ci-eng to: Ubuntu CI Engineering Team | Vanguard: ev | CITrain support - US: robru, cyphermox, rsalveti - EU: sil2100, Mirv, didrocks | CITrain support no answer: use mup bot after 30 minutes, but chooose right timezone | Landing instructions: http://goo.gl/8H1Du3 | Known Issues: - [09:12] sil2100: maybe starts with dialer-app [09:12] sil2100: it seems to have failed in the past 2 images [09:12] didrocks: ok, I see webbrowser-app had some problems in one image as well [09:13] yeah [09:13] I mentionned dialer-app because it doesn't have a lot of tests [09:13] whats up ? 199 looks okayish imho [09:13] so, should be easy to restart [09:13] ogra_: look at paul's message [09:13] and the previous image [09:13] there are quite some flakyness [09:14] yes, i see that one [09:14] but it went good again later [09:14] well [09:14] it will mean that it will maybe refail [09:14] Flaky tests are not good in overall [09:14] yeah [09:14] if we ignore, it will strike back [09:14] i didnt say ignore it :) [09:15] hence the "let's bisect it" :) [09:15] I would give some time to the responsible upstream (once we know who it is) to fix the test regression before reverting though ;) Like 2 hours or such [09:16] yeah :) [09:17] first one is to ensure that you have a small amount of tests reproducing the flaky failure [09:18] Righto! [09:19] Upgrade is finishing right now [09:28] Finally upgraded, proceeding [09:32] sil2100: joining? [09:32] sil2100: also, can you join #sdk? [09:32] Aaaa [09:32] didrocks: sorry, got sucked into dialer-app, be right here === zsombi is now known as zsombi|lunch [09:59] didrocks, ogra_: so, after downgrading dialer-app, the test doesn't fail anymore but unity8 still dies and I can't get it up ;) Trying the session revert [09:59] ah, excellent! [10:00] sil2100: so, reverting dialer-app for now [10:01] sil2100: can you get a bug report while the tests are running for dialer-app? [10:01] * didrocks works on the revert meanwhile [10:06] didrocks: filling [10:07] didrocks: just to make sure, don't upload the revert yet - just prepare it [10:07] sil2100: ok [10:24] sil2100, silo for 59 pls [10:24] didrocks, ogra_: so, downgrading ubuntu-touch-session didn't really help, but downgrading dialer-app indeed made the one failure go away [10:24] ohew [10:24] didrocks, ogra_: I also noticed something regarding that unity8 issue [10:24] *phew even [10:25] When the dialer-app test finish, I get a white screen instead of the app [10:25] Now: [10:25] - When I swipe it out, unity8 hangs up and crashes, but then restarts and works correctly [10:26] - When I leave the white screen-app until the screen turns off, unity8 dies and cannot be restarted anymore [10:27] sounds power related [10:28] i.e. muddled up powerd communication or so [10:28] mhr3: one moment ;) [10:29] Anyway, it's really strange [10:30] check the dbus and powerd logs i'd say [10:30] probably there is something obvious [10:31] sil2100: do you have the bug report number? [10:35] didrocks: hah, just filled it in, here it is: https://bugs.launchpad.net/dialer-app/+bug/1282981 [10:35] Launchpad bug 1282981 in dialer-app "Failure in tests.test_logs.TestCallLogs.test_call_log_item_opens_messaging AP test" [Undecided,New] [10:36] * sil2100 is reverting further [10:37] sil2100: so, uploading dialer-app revert [10:38] didrocks: yes, thanks! It's confirmed at least that the revert helps [10:38] ok ;) [10:38] done [10:38] so, next one :p [10:38] you are continuing on the sdk now? [10:39] First I would like to try lightdm, then SDK [10:43] This problem is really strange [10:43] sil2100: hum, lightdm wasn't on that list, you think we were just lucky? [10:45] Because this failure doesn't seem to be related to the SDK, but I'm trying it now [10:46] Need to reboot my device though, as it's hanged again ;/ [10:48] sil2100, two moments passed :) [10:48] mhr3: ;) [10:48] Let me do that while waiting for my device [10:49] Ooooor wait a little bit longer, since we're still bisecting what caused the problems we're seeing, who know's if it's not YOUR PACKAGE ! :D [10:50] i know it isn't! :P [10:50] didrocks: After the findings from sil2100 about dialer app, do you still want me to revert ubuntu-ui-toolkit-theme and qtdeclarative5-ubuntu-ui-toolkit-plugin and run all the tests? [10:51] didrocks: the failure link mail sent to you btw [10:52] psivaa: oh, you didn't start yet? please do [10:53] psivaa: it's only fixing dialer-app, not the rest of the worlds :) [10:53] didrocks: ack, doing it [10:53] mhr3: we need to concentrate on getting back to green first :) [10:53] psivaa: thanks ;) [10:53] mhr3: you're welcome if you can help us btw :) [10:56] didrocks, sorry, my prio is to have pretty mwc image :) [10:56] mhr3: sorry my prio is to keep having a green baseline :) [10:57] ;p [10:59] psivaa: thanks a lot for the detailed email, exactly what I needed! [11:00] didrocks: yw, now installing 199 on a device. (reverting on an earlier device used for smoke might contaminate the dashboard) [11:00] excellent [11:08] sil2100, didrocks: can I get a silo for a very harmless logging branch in row 60? [11:10] rsalveti: hum [11:10] ERROR:root:ubuntu-download-manager is already prepared for the same serie and destination in landing-010 [11:10] ERROR:root:One or more projects are already in use for the same destination and series in another silo (see above) [11:11] and you were the requester :) [11:11] didrocks: didn't know I can't have two silos :-) [11:12] ralsina_: you can't have two silos of the same project [11:12] ok, so now I know [11:12] ok, removing row 60 [11:12] ralsina_: well, otherwise, it means you can have conflicts in your branch [11:13] if you prepare in parallel [11:13] and so, you will have to redo all your work [11:13] it's to prevent shooting yourself in your feet :) [11:13] didrocks: no problem === lool- is now known as lool === vrruiz_ is now known as rvr [11:43] sil2100, around? [11:43] tvoss: hi! [11:50] geeh [11:51] Never before I was so irritated by tests passing [11:52] :) [11:53] I found which test failed first in the webbrowser-app test suite and try to run it in a loop now [11:53] ah ;) [11:54] not sure if it's just running that test which fails [11:54] or the previous one creating that one to fail [11:54] or anything else [11:55] psivaa: anything failing yet? Not sure if we can have another device running the "raw" environment again [11:56] didrocks: running the tests after reverting the ui-toolkit stuff. no failures yet. but the candidate apps for the failures have not started to run [11:57] didrocks: raw environment means? [11:57] psivaa: like rerunning in parallel every tests again with 199 without any change [11:58] didrocks: ok, that can be done on the device used for smoke tests. how soon will there be another image? [11:59] if it's not within another 4 hr, then that's safe to run [11:59] psivaa: not soon [11:59] yes :) [11:59] can you try that at the same time? That would be awesome and give us one more input [12:00] didrocks: ack, will run that again including the install. ( i earlier noticed some apps start to pass on the second run onwards) [12:00] psivaa: excellent, thanks! [12:00] yw :) === zsombi|lunch is now known as zsombi [12:04] sil2100: yeah, no failure for me as well on first run [12:05] didrocks: I was even running 3 subsequent tests (the last one which failed) in a loop -> no failures after 20 runs, now re-running whole webbrowser-app suite again, maybe I'll get lucky [12:06] I wonder if we would be able to reproduce it at all [12:06] that's why I guess the "raw" rerun from psivaa will help to say if we should ocntinue or not investigating [12:07] Indeed [12:10] hi [12:11] this is to get a silo reconfiged to add an MP to it [12:11] hey dbarth_, I think there is a silo to reconfigure, sil2100? ^ [12:11] ppa-005, i've added the MP to the pending tab already [12:11] sil2100: I would say, let's stop investigation for now, let's assign silos, but not publish [12:11] until we get more results from psivaa [12:11] sil2100: this MP covers friends-app which needs to be updated to work with the OA changes [12:11] sil2100: everything passed here as well [12:12] sil2100: i haven't seen other friends-app landings in parallel, so should be safe and won't lock other changes [12:12] Ou [12:12] Looking [12:12] didrocks: right, my re-run has finished, no failures ;_; [12:12] sil2100: yeah, third rerun here as well, nothing :/ [12:13] didrocks: sil2100:tests running on 2 devices one raw 199 and the other 199 with ui-toolkit stuff reverted. btw we did not have webbrowser app failures on 199 though [12:13] dbarth_: reconfiguring [12:13] didrocks: can we land an updated MP again? [12:13] psivaa: yeah, but there was no change for it since the failure on previous updates [12:13] thostr_: so, ok to attribute silos, we don't publish to distro yet though [12:13] until we get the results we are talking about ^ [12:14] psivaa: excellent that you can even use 2 devices! :) [12:14] (it's 2 mako?) [12:14] didrocks: ack, yes [12:14] perfect ;) [12:14] didrocks: meaning, MP was landed but then we figured it caused issues and updated the MP. So can I now just get a new silo with same MP (but different rev)? [12:15] (old silo was already cleaned) [12:15] thostr_: just rebuild the silo by mentioning the component you updated [12:15] Ah [12:15] thostr_: hum, so the MP was "merged [12:15] thostr_: yeah, it's not an issue [12:15] hm [12:15] you can get a new silo with that [12:16] (just note that if the MP commit is the same, it will just reput the same one in the changelog) [12:16] didrocks: can I get one right away then for line 60? it's urgent for mwc demo... [12:16] thostr_: ok [12:16] dbarth_: 005 reconfigured [12:16] sil2100: handling the other one? [12:16] didrocks: I'll assign for line 60 [12:16] excellent, thanks [12:20] thostr_: silo assigned [12:20] mhr3: for you as well [12:20] thx [12:20] sil2100: thanks [12:21] sil2100: that was fast. thanks! [12:21] thostr_, wait for the icon mp? [12:22] mhr3: is that ready? [12:22] thostr_, i asked alecu to add it to the screenshot one [12:22] oh, but that I already added to the silo [12:23] mhr3: that is https://code.launchpad.net/~alecu/unity-scope-click/add-screenshot/+merge/207487 [12:24] yea, although he seems afk [12:24] i'll make a new one [12:24] ne wmp [12:25] mhr3: ok [12:32] thostr_, https://code.launchpad.net/~mhr3/unity-scope-click/proper-icon/+merge/207636 [12:32] someone should review though [12:32] mhr3: yeah, let's get it reviewed first [12:33] and then land separately... need to get the scope fix (blocks right now) in ASAP [12:40] sil2100: didrocks: if you'd like to see the raw 199 results live: http://ci.ubuntu.com/smokeng/trusty/touch/mako/199:20140221.1:20140115.1/6736/ [12:40] sil2100, did you get anywhere with your research ? [12:41] * ogra_ holds back two uploads for the unity8 landing from silo 16 [12:41] i'm sure you'd ignore the queued ones. [12:41] psivaa: ah great, but what about the seconde device? [12:41] second* [12:42] ogra_: no... we didn't find anything ;/ In the end the dialer-app crash was something different, and we couldn't reproduce the flaky issues on any of our local re-runs [12:42] hmm [12:42] ogra_: we're waiting for smoke tests [12:42]