=== salem_ is now known as _salem === Amy is now known as Guest71968 [04:33] robru: do you know what does that mean? https://ci-train.ubuntu.com/job/ubuntu-landing-011-2-publish/59/console [04:33] kalikiana: ^ look, this is the blocker [04:35] bzoltan_: yeah you guys pushed new commits since the last time the build job was run. You need to rebuild. [04:36] robru: I see.. that I can do [05:35] robru: I am not sure if I get this ^ [05:41] bzoltan_: https://code.launchpad.net/~bzoltan/ubuntu-ui-toolkit/landing_12-6-2015-gles/+merge/261832 MP is wrong, you put the vivid version in, needs to be the wily version. (15.10 instead of 15.04) === Amy is now known as Guest70234 [09:14] robru: thanks, i fixed that [11:06] * sil2100 off to lunch [11:21] robru: sil2100: now the version is on wily, but now it does not like that it is not high enough ... (silo11) [11:22] the wonderful world of -gles [11:23] Uuu [11:24] bzoltan_: yeah, PPAs don't allow same-version uploads, looks like it has been already uploaded once? [11:31] sil2100: it seems that the main build produced the right gles package... so my rebuild was not necessary. The PPA has all the packages now. [11:44] phew [11:44] \o/ === MacSlow is now known as MacSlow|lunch === _salem is now known as salem_ [12:30] kenvandine: jgdx: Silo 3 approved [12:30] rvr, whee thanks! [12:37] o/ [12:38] jgdx, kenvandine: https://code.launchpad.net/~jonas-drange/ubuntu-system-settings/fix-1441192/+merge/255402 unapproved [12:39] :'( [12:39] sil2100, you mean the need fixing from ci? [12:39] Please top-approve it [12:39] ;) [12:40] It needs to be reviewed and top-apptoved by someone, we don't allow releasing non-top-approved merges [12:40] sil2100, the mr you linked to has something to do with silo 3? [12:41] jgdx: it's the only merge in silo 3 [12:42] jgdx: this merge needs to be top-approved before we can release silo 3 [12:42] sil2100, oh, was looking at the wrong silo 3. Okay so https://code.launchpad.net/~jonas-drange/ubuntu-system-settings/fix-1441192/+merge/255402 needs top approve. Right [12:43] it's top approved no? [12:43] Status: [12:43] Approved Edit status [12:43] Approved by: Ken VanDine on 2015-06-11 [12:43] looks like it, but maybe it's counting votes? [12:43] Ah! [12:43] it's rather than it got a commit since it was approved [12:43] that* [12:43] Yes, sorry, was fast with not reading the error message [12:43] sil2100, since the CI stack has broken down for USS, could you hold the phone a bit on that landing? [12:44] jgdx: this branch might need re-building though... [12:44] jgdx, how has "the CI stack has broken down for USS"? [12:44] sil2100, i don't think so [12:44] is that being tracked? [12:44] jgdx: it seems the packages were built and afterwards someone pushed new revisions to that [12:44] seb128, for all long running tests suites [12:44] seb128, yes [12:44] k [12:44] sil2100, ack, thanks [12:44] sil2100, the new revision was just merging trunk [12:44] do you have a bug number about that? [12:45] kenvandine, ^ [12:45] should be safe [12:45] kenvandine, yeah, but seems that revision is not built in the ppa, if it's needed we need a rebuild [12:45] kenvandine: was that a clean merge? [12:45] yes [12:45] seb128, bug v [12:45] bug 1468029 [12:45] bug 1468029 in unity8 (Ubuntu) "unity8 crash breaking autopilot tests entering text with OSK" [High,Confirmed] https://launchpad.net/bugs/1468029 [12:45] jgdx, thanks [12:45] hm, still... [12:45] that one is driving me nuts! [12:45] sil2100, we can do a rebuild [12:45] kenvandine: why was the trunk merged? [12:46] kenvandine, why did we do a rebase merge if there was no conflict/reason to remerge? [12:46] dunno :) [12:46] jgdx, ^^ [12:46] kenvandine, can we wait until AP pass? [12:46] jgdx, i'd rather not, the wifi tests pass [12:46] and there's no eta on when the unity8 crash will be fixed [12:46] kenvandine: since if someone landed something to u-s-s after building the package, we need to rebuild as otherwise it'll not include those commits in the released package [12:46] kenvandine, oh right, they do. [12:46] the silo build includes the same trunk [12:47] the revision that was merged in was from last week's landing [12:47] kenvandine, is anyone looking at that unity8 issue? is it confirmed to be unity8? [12:47] well, we know that crash happens everytime we hit the problem [12:47] it's also affecting webbrowser-app [12:47] kenvandine, I don't understand what was the point of that merge? if somebody went out of his way to do it that's probably for a reason? [12:47] sil2100, i can do a rebuild though [12:47] to be sure [12:47] and if so why do we discard it now? [12:48] jgdx, what prompted the trunk merge? [12:48] kenvandine: you could check the source packages as well and see if the changes you merged in from trunk are in there [12:48] SInce a rebuild might trigger a re-test ;p [12:48] kenvandine, failing ap tests, who knows. Want me to uncommit it? [12:48] most likely desperation from failing tests [12:48] that's what i was thinking [12:50] sil2100, confirmed, that merge from trunk only includes the previous landing which was in trunk before the silo was built [12:50] so it won't revert it [12:51] sil2100, the diff from that merge is included in the ppa [12:51] + [ Ken VanDine ] [12:51] + * Activate the accessPoint on click [12:52] kenvandine: ok [12:52] uuuh, we don't have any override flag for this [12:52] ;/ [12:53] sil2100, lets let jgdx uncommit that? [12:53] okay [12:53] sil2100, ok? [12:53] kenvandine: yeah, let's revert that commit and bzr push --override to the branch [12:53] cool [12:53] jgdx, be careful... make sure the last revno is 1386 [12:53] Pushed up to revision 1386 [12:54] cool [12:55] mzanetti, any thoughts on bug 1468029 [12:55] bug 1468029 in unity8 (Ubuntu) "unity8 crash breaking autopilot tests entering text with OSK" [High,Confirmed] https://launchpad.net/bugs/1468029 [12:55] mzanetti, i'm seeing a unity8 crash while running the settings AP tests, and after the crash all text input from the OSK fails [12:55] restarting unity8 fixes it [12:56] mzanetti, not sure if the bug is actually in unity8, but that's where the crash is [12:56] Publishing :) [12:56] mzanetti, not sure who to harass about it, just want to make sure someone looks at it [12:56] we can't get passing CI right now :/ [12:56] sil2100, thx [12:57] kenvandine, do you know when it started? [12:57] jibel, i think i first saw it monday [12:58] but oSoMoN saw it before then [12:58] jibel, i can reproduce it 100% of the time, when running the full test suite [12:58] smaller test runs work fine [12:58] kenvandine, did you try to downgrade unity8 to the previous version and see if that resolves it? [12:58] once it fails... i have to restart unity8 to get any of the tests to pass [12:59] i had some dependency issues doing that [12:59] downgrade those as well ;-) [12:59] I don't think we had a transition involving half the archive [12:59] kenvandine, on vivid and wily or only wily? [12:59] wily and vivid [13:00] unity8 has had dual landings [13:00] vivid is stable though, we didn't get any unity8 sru [13:00] vivid overlay [13:00] oh, that's on the device [13:00] right [13:00] jibel, oSoMoN has seen it on krillin with vivid too [13:00] did you try on a desktop? [13:00] so not just mako [13:00] couldn't reproduce it on the desktop [13:00] it's only on the device [13:00] well if it's an osk issue... [13:00] k [13:01] and it happens when unity8 crashes [13:01] seb128, I meant overlay, sorry. There has been 2 uploads in a week [13:01] everytime it happens, there's a unity8 crash [13:01] oSoMoN, do you have an idea when you first saw it? [13:01] kenvandine, unity8 segfaulting doesn't mean it's not a qt or lib bug [13:01] seb128, right... it's just as low as i was able to bisect it [13:02] would be useful to downgrade to nail down the buggy component [13:02] and once it happens i can't get even a single test to pass using the OSK until i restart unity8 [13:02] can you open the osk manually? [13:02] yes [13:02] and that works [13:02] weird [13:02] just not from autopilot [13:02] right... i thought the same thing [13:02] weird [13:02] very easy to reproduce [13:03] let me phablet-test-run on my bq [13:03] well... with an hour long AP test run :) [13:03] autopilot send input event and don't really use the osk [13:03] it raises the OSK when entering text though [13:03] maybe a focus issue? [13:04] keyboard.write doesn't work once it gets to that state [13:04] or whatever that function is :) [13:04] brendand, have you seen bug 1468029? [13:04] bug 1468029 in unity8 (Ubuntu) "unity8 crash breaking autopilot tests entering text with OSK" [High,Confirmed] https://launchpad.net/bugs/1468029 [13:04] i spent a bunch of time narrowing it down this far, but had to move on to something else [13:05] jibel, i don't think i've encountered it personally [13:18] kenvandine, I don’t remember when I first saw it, no, but I would say at least a couple of weeks [13:22] Need to jump out to the vet [13:22] kenvandine: hey! Could you keep one eye on the train for me in the meantime? :) [13:25] pete-woods: Hi. The merge proposal of silo 35 hasn't been reviewed. The silo is blocked for QA. [13:25] pete-woods, I'll verify silo 35 since I was affected by the issue but could you have someone to review the MR and top approve it? [13:26] jibel: lol [13:26] :) [13:31] jibel: yes, will do [13:32] sil2100, i need to run out for a few minutes myself, but i can look in on it when i get back [13:32] i won't be long [13:33] Same here [13:33] kenvandine: anyway, thanks! ;) [13:49] sil2100, i'm back :) [14:20] sil2100, is there anything to discuss in the landing meeting tonight? [14:54] trainguards, ubuntu/landing-004's changes are top-approved now and ready to publish [14:57] kenvandine: I think sil2100 run out for a few minutes was longer than yours :) [14:57] davmor2, indeed :) [14:58] * davmor2 prods sil2100 repeated till he gets annoyed and nolonger pretends to be away ;) [14:58] davmor2, something i can help with? [14:58] charles, thx [14:58] kenvandine: no [14:58] kenvandine: I just like to annoy sil2100 :) [14:59] charles, published [14:59] lol [14:59] kenvandine, thanks === fgimenez__ is now known as fgimenez [15:36] jibel: top approved now [15:59] davmor2: I can hear & see you just fine [16:03] jibel: ping [16:29] sil2100, pong === alan_g is now known as alan_g|EOD === dpm_ is now known as dpm [18:49] robru: hi, I'd like to request access to the CI Train sheet so I can put Telegram in [18:49] karni: ok one sec [18:51] Is there any template for "Test plans to run"? We already have one, but it seems clock and reminders-app are using their 'own'/different layouts, so I guess that's fine? [18:51] karni: I would just copy & paste an existing one but then change the details to be specific to your program [18:52] karni: ok I've given you edit rights on the spreadsheet and also put you in the lp group necessary for triggering builds. [18:52] karni: did you read the details about how the train works? [18:52] robru: thank you [18:52] robru: I think that would be difficult, as our testing sheet is quite extensive https://docs.google.com/spreadsheets/d/12ZcjGNJuWCjT4cRSGrwkK10bWbbdAy9S2p3cPkVDW9k/edit#gid=1 [18:53] karni: oh well if you have a document that explains how to test it, just link that URL in the "test plans" field. [18:53] Long story short, I want a single row to be addressed only, but I'll leave that for the notes section [18:53] great, thank you robru [18:53] karni: you're welcome [18:54] robru: where can I read up on 'how the train works'? [18:54] karni: also note that the spreadsheet will be replaced in 1-2 weeks so don't spend too much time learning how to use it ;-) [18:54] robru: ok, that's ok :) [18:54] karni: https://wiki.ubuntu.com/citrain/FAQ is a good start with links to other docs [18:54] *oh, that's ok - haha [18:54] robru: thanks! [18:54] karni: you're welcome! [18:55] karni: any questions, feel free to ping me, or 'trainguards' to get somebody timezone-appropriate. [18:55] karni: and when in doubt -- ask for help before making a mistake ;-) [18:55] aha, I see the chan title now. great :) [18:55] yessir [18:56] karni: the train isn't as terrible as it used to be, but it still has a lot of bespoke gotchas. you'll learn them over time, no worries. [18:56] ok :) [19:26] trainguards: what should I put in "Additional source packages to land" (column G of "tarballs and clicks" sheet) in CI Train sheet? [19:26] oops.. not sure if I should have used that word (I hope only some folks have highlight for it) [19:27] robru: if you're around, question ↑ [19:29] karni, that's only if you have packages to land that aren't included in bzr branches [19:29] like a manual upload [19:29] or a sync from a different series [19:30] kenvandine: I'm sorry, I still don't get it :( I have a click, it's a non-citrain upload to the store. I have a click I want to have QA'ed. What should I put in there? [19:30] robru: sil2100: so what blocks the silo11 from landing? [19:31] karni, you have a merge proposal right? [19:31] i'm actually not familiar with building clicks with citrain [19:31] yes, to a library that our app (in a different branch) is using [19:31] but i'd imagine you have bzr branches [19:32] kenvandine: karni: the train doesn't build clicks. [19:32] well.. the build is more complex than bzr branch and click build TBH. we have a dependency [19:32] so you skip the addition packages [19:32] okay :) [19:32] but i guess it isn't what you want anyway :/ [19:32] bzoltan_: did you miss my last message? I said you have to fix the version number in your MP, you put the vivid version when it needs the wily version. [19:33] karni: basically you have to indicate in the notes not to assign a silo because that's meaningless, then you need to build your click yourself and link to it from the MP field. then ping QA about it. [19:33] karni: basically this is a giant hack and totally unsupported by any infra. but the spreadsheet replacement will come soon and handle this better. [19:33] robru: sorry, I am confused. The PPA seems to have right packages: https://launchpad.net/~ci-train-ppa-service/+archive/ubuntu/landing-011 [19:34] robru: gotcha, thanks! [19:34] kenvandine: thank you for your help anyway :) [19:34] np [19:34] bzoltan_: wtf... looking [19:36] bzoltan_: I don't know what's going on, did you delete revisions from the branch or something? [19:36] robru: I have not done anything special... after your message I have updated the branch and pushed a new build... what failed with the message you can see on the silo dash [19:37] bzoltan_: as far as I can tell from what's in the MP and what the build logs indicate, it looks like you built -gles correctly with the correct version number, then broke the version number, then built again and got the error that the version number is wrong, then changed the MP back to the correct version number, but didn't rebuild again. [19:38] bzoltan_: if you look at the build logs, there's a successful one an hour before the most recent failed one, and they both reference the same revision number. [19:38] robru: so what should I do? [19:38] bzoltan_: I'm going to try a watch-only build and see what happens. worst case it might need another rebuild. but the MP looks correct at this point. [19:39] robru: All right. Just ping me if I need to rebuild the main and bump the gles version [19:40] bzoltan_: can you smoke-test the silo contents since they were rebuilt? [19:40] robru: only tomorrow morning :) I am on holiday and soom 10pm [19:41] * bzoltan_ hides from the wife already [19:41] bzoltan_: ok, watch_only seems to have worked, but my concern is that the packages were rebuilt since QA approved it, so there's potential for regression there. [19:41] robru: is there a friendly QA dude around here who could take a quick look? [19:42] jibel: any qa'ers that can quickly recheck silo 11? [19:43] davmor2: ^ [19:43] ToyKeeper: ^ ;-) [19:45] Oi, what changed in the rebuild? [19:45] UITK isn't generally a "quickly recheck" sort of deal. [19:45] ToyKeeper: basically nothing, just need a quick smoketest to make sure it didn't explode. [19:46] ToyKeeper: i think it's just a version bump because of reasons, but it needed a rebuilt [19:46] When it was tested before, it took 5 hours. [19:47] ToyKeeper: http://bazaar.launchpad.net/~bzoltan/ubuntu-ui-toolkit/landing_12-6-2015/revision/1536 here's the newest commit that got rebuilt, just translations. [19:47] Fortunately, it's almost always solid. [19:47] ToyKeeper: yeah we don't need the full 5 hours, if you could just poke at it for 10-20 minutes to confirm it turns on that'd be super. [19:48] I'll take a look... [20:09] robru: I don't see anything obviously broken... [20:10] ToyKeeper: great, thanks! [20:10] A little confused why it didn't land three days ago after being approved though... or did it? [20:12] ToyKeeper: no it didn't. what happened was they pushed new commits since the approval so the train refused to publish because there were new commits that hadn't been built [21:30] cjwatson: bah, I broked copy2distro on snakefruit, can you help me troubleshoot? http://people.canonical.com/~ubuntu-archive/cicopy.log says "not found" but it shouldn't be that bad... [21:34] infinity: ping? ^ [21:48] robru: I'm wandering by in passing between a nap and a nap, but lemme look. [21:48] infinity: so i modified that file, but i didn't move/delete it, not sure why it would say "not found" [21:48] infinity: a previous copy of that log had a traceback I was working to fix. [21:49] infinity: can you paste me a copy of /home/ubuntu-archive/cu2d/run.sh ? [21:49] robru: Because you switched from python2 to python3, and snakefruit no haz. [21:49] infinity: can we make it have python3? [21:51] robru: If 3.2 is good enough for you, you could ask IS nicely. If you happen to need various extensions ported, though, and you mistakenly believe snakefruit is a recent release, you'll be disappointed. [21:51] (it's precise) [21:51] bah [21:51] infinity: I have backports of stuff I need prepped, but it's all for trusty [21:51] Are you familiar with the saying "write to the infrastructure you want, not the infrastructure you have"? [21:52] If not, that's because no one says that. [21:52] infinity: well 99% of what I'm dealing with is trusty, so what's what I write for. then snakefruit stabs me in the eyes out of nowhere. [21:53] Make it both 2 and 3 compatible, and use a 2 shebang for now until snakefruit is upgraded? [21:53] I mean, assuming you need stuff that's not built for 3 in precise, but that seems pretty amazingly likely. [21:53] Especially if you use, say, lplib. [21:53] infinity: ok, will do. it's just that I dropped python 2 support from a bunch of stuff recently thinking I had all my deps covered for trusty. [21:55] infinity: snakefruit has python-six right? I think I asked for that a while ago [21:57] robru: https://pastebin.canonical.com/134046/ [21:57] infinity: goodie, thanks === salem_ is now known as _salem [22:23] there we go