=== plars-afk is now known as plars === chihchun_afk is now known as chihchun === forestpiskie is now known as Guest40580 [05:59] Good Morning all [06:00] afternoon DanChapman === Guest40580 is now known as forestpiskie === ev_ is now known as ev === yofel_ is now known as yofel [10:01] jibel or pitti: is it acceptable for a dep8 test to hit http://ddebs.ubuntu.com/ and/or keyserver.ubuntu.com? I think they should be completely standalone? [10:02] rbasak, tests that rely on external resources are usually quite unreliable [10:03] jibel: sorry I intended to point you to the bug. https://bugs.launchpad.net/ubuntu/+source/crash/+bug/1217474 and the proposed debdiff is https://bugs.launchpad.net/ubuntu/+source/crash/+bug/1217474/+attachment/3790784/+files/crash_autopkgtest_v1.debdiff [10:03] Ubuntu bug 1217474 in crash (Ubuntu) "basic autopkgtest for crash" [Medium,In progress] [10:03] jibel: I'm tempted to nak due to the Internet dependency, but would like your opinion. [10:04] A dependency on ddebs shouldn't be too bad I suppose. The keyserver dependency could be fixed by embedding the public key in a heredoc. [10:09] rbasak, OTOH we support testing from PPAs which also requires an access to keyserver.u.c, so these 2 dependencies are not that bad [10:14] OK thanks. I'll test and propose upload then. [10:51] hi balloons :) === zyga is now known as zyga-food [12:27] Good morning [12:27] morning pitti :-) [12:28] rbasak: right, *.ubuntu.com is accessible in principle from the DC [12:28] hey DanChapman [12:31] good morning everyone [12:31] morning thomi [12:32] thomi: yes, I have an i386 sbuild [12:32] pitti: ok, I'll retry the build and see what happens [12:32] thomi: let me try again with current packages; I still have the experimental PPA, is that still good enough? [12:32] yes [12:33] pitti, about bug 1226030, no difference with gdb installed on the device [12:34] bug 1226030 in apport (Ubuntu) "apport refuses to report a crash from another machine if package is not installed locally" [Undecided,New] https://launchpad.net/bugs/1226030 [12:34] jibel: ah, so gdb was missing indeed? that would at least be one reason [12:34] jibel: I'll try myself today, just didn't want to go into major debugging last night any more [12:35] thomi: I'll build and install current ap-trunk first, to make doubly sure I've got the latest stuff [12:35] pitti, np === zyga-food is now known as zyga [12:41] thomi: success on amd64 (local build) === _salem is now known as salem_ [12:51] thomi: i386 still running (taking ages as the "add bamfdaemon" dep isn't deployed yet), but it's well past the point where xvfb would have failed [12:51] thomi: acked your "add test" MP [13:11] jibel, you made changelogs :) [13:12] balloons, yes, faster than waiting for an hypothetical script to land on cdimages :) === pgraner` is now known as pgraner [13:36] * pitti shakes fist at ap-gtk i386 xvfb failure [13:36] thomi: ^ [13:36] works just fine in my i386 schroot [13:37] pitti: :( [13:37] pitti: can you review this for cgoldberg as well please? https://code.launchpad.net/~coreygoldberg/autopilot/tox-and-packaging/+merge/186034 [13:37] the build stuff specifically [13:37] thomi: do these run on real iron? [13:37] pitti: I don't think so, no [13:37] thomi: perhaps there's a stale Xvfb process which didn't clean up? [13:37] fginther: any ideas about that? ^^ [13:38] thomi: we could add a commit to the exp branch which shows the xvfb log, figure it out, and then revert it before final merging [13:39] thomi, pitti, I'll try to take a look, but it may have to wait 30 minutes for my meeting to end [13:39] thomi, fginther: I'll add -a (to avoid failing on stale Xvfbs) and -e to store the log file and cat it [13:39] fginther: thanks [13:46] thomi: ok, pushed robustification and X.org log showing (tested locally with an already running xvbf) [13:47] pitti: awesome, got a MP for me to approve? [13:47] thomi: oh sorry, I just pushed to the exp branch [13:47] as that is already an MP [13:47] so it'll just update https://code.launchpad.net/~autopilot/autopilot-gtk/experimental/+merge/185845 [13:48] check the last 3 commits there [13:49] thomi: one day we'll have autopilot-run-sandbox for that :) [13:49] cool [13:50] now let's wait for the bot run; tests should be much faster now, I added the bamfdaemon dep [13:51] right [13:59] oh, we have a little backup in ubuntu manual tests for reviews [14:00] good morning QAs. [14:01] good morning elopio [14:01] 0/ [14:04] elopio: cgoldberg morning [14:05] morning davmor2 [14:09] elopio: hey [14:09] hey. What's the change you want? [14:09] elopio: so the issue is that autopilot 1.3 trunk now raises an exception instead of returning None [14:09] I'm working on that today, so it's a nice time. [14:10] elopio: but this makes the ubuntu ui toolkit emulator tests fail [14:10] oh, right. [14:10] elopio: which means didier won't release the newer AP to distro [14:10] cgoldberg: https://code.launchpad.net/~coreygoldberg/autopilot/tox-and-packaging/+merge/186034 reviewed [14:10] thomi: I can change that easily on the emulators. The problem is that it will make many other tests fail on the apps. [14:10] thomi: perhaps we should back that out and move to 1.4? [14:11] elopio: there's only thee failures: https://launchpadlibrarian.net/150474284/newautopilot.txt [14:11] pitti: I think we can just update the emulators [14:11] pitti: but that's certainly an option. what do you think elopio? [14:11] thomi: are you completely sure? That would be a nice surprise. [14:11] thomi: if the failures are only on the emulators, +1. I'll change it in an hour. [14:12] but I'm not sure those results you are showing me are for all the suites of all the apps. [14:12] elopio: when you propose the merge, please let me know and I'll update the landing SS and email the right people [14:12] elopio: no, just the toolkit [14:12] hmmm [14:12] maybe pitti's right [14:12] gah [14:12] this is bad [14:13] I estimate it will take a couple of days to get all the apps to expect this behavior. Maybe balloons can do a better estimate. [14:13] elopio: OK, I think that, in that case, we have no choice but to postpone the change to 1.4 [14:13] I can update them, but as we are running with the testing, maybe pitti's suggestion is the right one [14:13] * thomi swears [14:13] uh, seems a simple merge from the 1.3 branch into trunk pulls in a gazillion commits (I wanted to forward-port the recent changes, but seems you don't usually work that way) [14:14] pitti: should work just fine [14:15] * balloons is only slightly following [14:16] elopio: OK, forget I asked - I'll remove them from 1.3 and add those changes to 1.4 instead [14:16] cgoldberg: FAOD, my only real objection is about including debian/changelog; the "run full tests" is just a suggestion, and we can always add that later on [14:17] thomi: sounds wise. Lets do it the week after the release, because I really like that change. [14:17] pitti, hi.. thanks for the review! I only included changelog there because it was already being referenced in setup.py already [14:18] elopio: lets talk about it then, yeah [14:18] cgoldberg: ah, it parses out the version from there [14:18] ugh. conflicts [14:18] thomi: yeah, seems 1.3 hasn't been merged into trunk for quite a while [14:18] pitti: yeah. I'll work through the changes [14:19] leave it with me :) [14:19] pitti, yea... referencing debian/changelog in setup.py seems odd though [14:19] pitti: we should have donen the merge forwards before the experimental branches landed [14:19] that would have avoided all this [14:19] cgoldberg: we do that so we don't need to maintain three separate version numbers, just 2 [14:19] thomi: you can unwind your local checkout to that, merge 1.3, store the diff, and rewind to trunk [14:20] pitti: yeah... if this gets tricky I might do that [14:22] * pitti tosses some candy jenkinswards to lure it to re-run the ap-gtk MP test [14:51] thomi: hm, does https://code.launchpad.net/~autopilot/autopilot-gtk/experimental/+merge/185845 need some prodding to get re-tested? I thought pushing new commits was enough, but curiously they have disappeared from the comment flows (they are still there in the diff) [14:54] afternoon all [14:55] afternoon slickymaster :-) [14:55] DanChapman: Hi Dan, hope evrything's fine with you [14:55] pitti, needs to be top-approved to be re-autolanded; not sure if it'll be tested in "CI-build" once approved--also consider pinging fginther and CI team about :) [14:55] alesage: oh, ok; I'll do that then [14:56] slickymaster: all good here thanks :-) [14:56] alesage: yes, that's what thomi did before (re-approve to trigger a new test), but I thought as I pushed new revisions the tests would run without that [14:57] pitti: it should work [14:57] pitti: probably the build queue is too long [14:57] pitti: ask the CI guys I guess [14:57] pitti: don't top-approve it unless you want it to land [14:57] ack [14:58] the new tests will get picked up :) [14:58] thomi: well, I do want it to land, but it should run the tests again [14:58] pitti: I thought you wanted to back out at least one of those revisions? [14:58] pitti, hey! when do you release new apport ? [14:58] thomi: no, I think we should keep them in [14:59] thomi: can't hurt to show the X log on failures [14:59] om26er: I'm still working on bug 1226030 to complete bug reporting from phone [14:59] bug 1226030 in apport (Ubuntu) "apport refuses to report a crash from another machine if package is not installed locally" [Undecided,In progress] https://launchpad.net/bugs/1226030 [14:59] om26er: so at some point today [14:59] I got the others fixed [14:59] pitti, ok, cool [14:59] pitti: oh, well in that case you can top-approve if you want, but that's not needed to get the CI system to run [15:00] perhaps fginther could tell us what's going on - whether it's just a very long build queue, or something else? [15:00] thomi: ok; I'll check after the next LinuxCon talk, if it still didn't run I'll prod CI [15:00] cool [15:02] thomi, pitti, build queue == long [15:02] fginther: ack, thanks for checking [15:02] fginther: :( [15:02] did some fiddling, should run better for the moment [15:03] * thomi thinks we should turn jenkins up to 11 [15:03] more like 12 [15:03] http://xkcd.com/670/ ! [15:05] pitti, :-p [15:09] fginther: ah, so it ran; now it failed on arm [15:09] eek [15:09] fginther: https://jenkins.qa.ubuntu.com/job/autopilot-gtk-saucy-armhf-ci/25/console [15:09] fginther: that doesn't look healthy, is that a glitch? [15:10] thomi: anyway, i386 succeeded now, so I guess there's some stale xvfb process somewhere [15:10] pitti, ugh. that's a transient error I've starting seeing in the past week [15:11] pitti, I'll re-trigger [15:11] fginther: thanks [15:11] pitti, although if the other builds pass, I would just consider it ok and retest with autolanind [15:12] fginther: that's what I intend; I did set it to approved [15:12] pitti, thanks! no sense in adding to the build queues :-) [15:32] hmmm, found a packaging problem in autopilot 1.4 [15:35] thomi, what problem? I've been doing a crash course in AutoPilot packaging :) [15:36] cgoldberg: for some reason you can't have python_autopilot >= 1.4 and libautopilot-qt >= 1.4 installed together at the same time [15:36] cgoldberg: I think it's due to the version numbers build in the experimental PPA [15:41] oh nuts [15:41] I accidentally pushed directly to trunk :-/ [15:41] * thomi undoes quickly before anyone notices [15:44] fginther: failed again with the same 401 error :( [15:45] pitti, do you have a link, I want to make sure it's not the same machine having an issue [15:46] http://jenkins.qa.ubuntu.com/job/autopilot-gtk-saucy-armhf-autolanding/14/console [15:46] fginther: previous one (even if the number suggests otherwise) was http://jenkins.qa.ubuntu.com/job/autopilot-gtk-saucy-armhf-ci/25/console [15:46] pitti, they were run on different hosts [15:46] pitti: cgoldberg, elopio, veebers: could I please get you guys to eyeball this? https://code.launchpad.net/~thomir/autopilot/forward-port-changes/+merge/186082 [15:47] this is the forward port of the 1.3 changes into 1.4 [15:47] fginther, it is on the calxeda box [15:47] I've run all the tests locally, and they pass [15:47] balloons, hi :) [15:47] jibel, yes, both failures were on calxeda [15:47] WebbyIT, hello! how are you? [15:47] thomi: queueing (talk now) [15:47] balloons, fine, thanks :) And you? [15:47] cool [15:48] fginther, are all the keys and creditials being copied to this box? we had something similar on daily-release a while ago [15:48] credentials [15:49] jibel, I was getting ready to check that. Other builds are working there, but perhaps they don't hit launchpad the same way [15:51] thomi: sure, give me a moment. [15:52] jibel, the problem is intermittent. Is this the same issue that daily release has been fighting? [15:52] WebbyIT, good.. thank you. What have you been working on recently? [15:53] balloons, I'm finishing the calc app, and I have a question for you ;) There is a bug that is assigned to you [15:53] bug 1207593 [15:53] bug 1207593 in Ubuntu Calculator App "[Autopilot] test_save_calculation_between_reload needs a better solution" [Medium,Triaged] https://launchpad.net/bugs/1207593 [15:53] balloons, what's about it? [15:54] WebbyIT, good question, I hadn't seen that [15:55] fginther, no we've been fighting with networking issues, here it is clearly an access denied [15:55] jibel, thanks [15:56] WebbyIT, is there a disabled test or was it completely removed? [15:58] balloons, dunno, boiko rewrote a lof ot test [15:59] elopio: got time for another review? This one's important: https://code.launchpad.net/~thomir/autopilot/remove-incompatible-changes/+merge/186086 [16:00] WebbyIT, well, I guess we should ask om26er his thoughts about https://bugs.launchpad.net/ubuntu-calculator-app/+bug/1207593 [16:00] Ubuntu bug 1207593 in Ubuntu Calculator App "[Autopilot] test_save_calculation_between_reload needs a better solution" [Medium,Triaged] [16:01] balloons, I assume we need, a) a way to kill the app and then start it again in the same test [16:02] om26er, ohh is that the test.. gotcha.. yes, we can't cleanly exit the application then start it again.. [16:05] WebbyIT, while I have you though, do you mind reviewing https://code.launchpad.net/~nskaggs/ubuntu-calculator-app/disable-osk-autopilot/+merge/186042? it fixes the current test failures [16:06] thomi: I think I'm almost done on the sdk. So, yes, noted on the top of my queue. [16:07] balloons, I approve it on trust ;) [16:08] balloons, feel like confirming a few bugs ? [16:09] om26er, I'll be confirming everything later this afternoon [16:09] * balloons notes his phone is stuck in boot after flashing :-( [16:10] balloons, adb reboot [16:10] do that multiple times if neede [16:10] needed [16:10] om26er, sure, I just wanted to see if there was something to have a look at [16:11] i can get a busybox shell, which is coo [16:11] elopio: thanks [16:11] balloons, busy box means you need to reflash (from my experience) [16:12] cool, so the update failed all over the place [16:12] i might be wrong, who knows [16:17] thomi: this one is ready to land: https://code.launchpad.net/~elopio/autopilot/loggging/+merge/185870 [16:17] I had to use pprint to get a consistent order of kwargs on python 2 and 3. [16:19] pitti: how did you solve that problem in the python 3 port? I remeber you had a more elegant solution, but I can't remember what it was? [16:20] thomi: https://code.launchpad.net/~chris.gagnon/unity/fix-endless-while-loop-in-launcher-tests/+merge/186090 [16:20] thomi: when you have time please look ^ [16:21] sure, will review now [16:21] robotfuel: should be "raise AssertionError(...)", not "raise(AssertionError(...))" [16:22] robotfuel: otherwise LGTM [16:22] robotfuel: fix that and I'll approve [16:22] * thomi -> lunch === wxl_ is now known as wxl [17:06] thomi: https://code.launchpad.net/~chris.gagnon/unity/fix-endless-while-loop-in-launcher-tests/+merge/186090 is updated [17:07] robotfuel: approved [17:13] veebers: are you working today? [17:17] elopio: ping? [17:18] thomi: pong. [17:18] elopio: I fixed the first issue you found on this MP: https://code.launchpad.net/~thomir/autopilot/forward-port-changes/+merge/186082 - the second issue is arelady fixed - we depend on qmlscene which pulls in the correct deps [17:19] thomi: veebers said he will be here tomorrow. [17:19] elopio: if you could re-review and approve when you're happy, that would be great [17:19] elopio: thanks [17:19] thomi: done. [17:20] elopio: thanks [17:20] np. [17:20] and now I'm leaving for lunch. [17:20] bbl. [17:24] fginther: got a second? [17:40] thomi, what's up? [17:49] thomi: ugh the issue is also in mouse.move in autopilot. the stagnant loop_detector.check_state doesn't detect the mouse is stagnant. this is probably also an issue between xmir and autopilot not seeing the cursor position. [17:53] jsut discovered that headphones don't work for phone calls [17:54] balloons, can I grab you for one test ? [17:54] on the phone [17:57] elopio: let me know when you're back form lunch please? [18:15] fginther: hey, is autolanding broken? I looked at the jenkins server, and the build queue seems small, and yet these are taking *ages* to land: https://code.launchpad.net/autopilot/+activereviews [18:15] 2 hours so far === chihchun is now known as chihchun_afk [18:19] they are building, but you're right things are slow today. I know the trigger job was waiting as well, that creates a delay to start new jobs [18:19] ugh, cool - thanks [18:45] thomi: something like "if not _PY3: [ do the unicode() ] thing"? [18:45] pitti: uhh.. wha? [18:46] thomi: or what do you mean in particular? [18:46] pitti: sorry, I have no idea what you're talking about. my brain is obviously dead [18:47] thomi: that's what I said in https://code.launchpad.net/~thomir/autopilot/fix-state-not-found/+merge/185675 [18:47] thomi: that test which asserts that unicode(StateNotFoundError instance) works [18:47] that can't work in py3 [18:48] thomi: but there were a few other str vs. bytes confusions pointed out in the tests [18:48] * thomi swears. I'm trying to do too many things at once. [18:48] pitti: you're saying that my forward-port branch failed? [18:48] 'I hadn't seen that yet [18:48] thomi: yes [18:49] thomi: should be reproducible with python3 bin/autopilot t.a.unit [18:49] of course [18:49] i'll fix it up [18:51] pitti: any chance you could review this please? https://code.launchpad.net/~thomir/autopilot/1.3-fix-scan-speed/+merge/186113 [18:53] thomi: that's a big one, will take me a bit (post-lunch keynote starting in 8 mins) [18:53] thomi: is that a backport from trunk, or new code? [18:53] pitti: it's new code, but it's pretty simple [18:55] thomi: forward-port warning, psutil doesn't exist in py3 [18:56] pitti: I checked that it does this morning [18:56] pitti: but maybe I messed that up as well [18:56] pitti: it does exist [18:56] no, it doesn't, I'm afraid [18:57] python3-psutil package exists in ubuntu [18:57] python3-psutil - module providing convenience functions for managing processes (Python3) [18:57] ooh, ok; separate dep [18:57] :) [18:57] sorry, I don't have that installed, nevermind [18:57] thomi: so you are going to need a new dependency at least [18:58] pitti: for trunk, yes, but this is going into 1.3 [18:59] thomi: I don't understan -- how is the branch relevant for the missing dep? [19:00] apt-cache rdepends python-psutil -> nothing APish [19:00] pitti: 1.3 is python 2 only. I'm saying we don't need to add python3-psutil to this MP, since this is merging into lp:autopilot/1.3 [19:00] ah, it already depends on that [19:00] pitti: yeah, I already have the python2 version in there [19:00] thomi: ack; sorry, we meant different things [19:01] nw :) [19:04] thomi: btw, do you know about "else" in a for loop? looks a bit nicer than the "if i == 9" (no biggie nor rejection reason of course) [19:04] pitti: I do know about it, I tend not to use it much, although maybe I should use it more often [19:04] pitti: I find that it's often not very expressive, ut.. *shrugs* [19:05] thomi: communicate()'s documentation says that it waits for the process to terminate; are you doing this in a loop because that somehow fails? [19:05] pitti: I believe that python2 and python3 are different in that regard [19:06] pitti: and (from memory - this is actually old code) this is a bit of premature optimisation [19:06] om26er, sure, my phone is feeling better now :-) [19:06] thomi: interesting; so far I relied on that in my projects (also for py2) [19:06] pitti: my memory could be wrong :-/ [19:07] thomi: what does test_content() do? [19:07] but I wrote that code *ages* ago - the fix in this branch is a total ugly hack, and I have plans to fix it properly in the 1.4 [19:07] pitti: attaches the content to the test [19:07] ack [19:07] it turns a str into some kind of magical object which you use with addDetail()? [19:08] just wondering how it deals with unicode objects [19:09] pitti: pretty much [19:09] + with maximum_runtime(5.0): [19:10] I hope that works on arm as well [19:10] thomi: how long does that take now on your box? [19:10] pitti: 3.4 seconds [19:10] so there's a bit to spare [19:10] thomi: arm boxes might be quite a bit slower [19:10] thomi: but let's see how it holds up in the CI instances [19:11] yeah [19:13] thomi: ok, reviewed; let's wait for CI results [19:17] balloons, bug 1226739 but it seems Pat already verified it [19:17] bug 1226739 in telephony-service (Ubuntu Saucy) "Ubuntu phone calls back automatically" [Critical,Confirmed] https://launchpad.net/bugs/1226739 [19:17] om26er, gotcha [19:53] thomi: I'm back. [19:54] elopio: eh, sorry - I figured it out [19:54] that's good. [19:55] should we put the avengers tag only to the new bugs we open, or also to the ones we confirm? [19:56] jfunk: ^ [19:57] elopio, if it is missing the avengers tag and you confirm, add the tag [19:58] elopio, also the qa-touch tag [19:58] ack. [20:06] balloons, bug 1226739 is what you need to confirm :) [20:07] bug 1226739 in telephony-service (Ubuntu Saucy) "Ubuntu phone calls back automatically" [Critical,Confirmed] https://launchpad.net/bugs/1226739 [20:07] om26er, it is confimreD? [20:08] balloons, I mean you need to verify on your phone. Tony says its only me who faced the issue [20:09] om26er, ok sure not a problem.. just finishing up some tests [20:09] ok thanks [20:18] pitti: bug #1226810 was the crash I got [20:18] Error: Launchpad bug 1226810 could not be found [20:18] hmm [20:18] cjohnston: ah, -EPERM for me; did you just report it? [20:18] yes [20:19] ah, so apport still needs to deliberate about it [20:19] its private [20:19] and yay for talking to your roommate on IRC :) [20:19] :-) [20:51] cgoldberg: isn't tox supposed to install depends before running the tests? it misses testscenarios and mock, at least [21:16] thomi, normally you have it install the dependencies explicitly (it uses a virtualenv). since we are using system packages, I set that explicitly in tox.ini... so you have to have deps installed before running [21:17] cgoldberg: awesome, thanks for setting that up [21:17] np [21:54] * elopio goes for a walk. === salem_ is now known as _salem [22:24] * slickymaster goes for some fresh coffee [22:34] * slickymaster has returned with a fresh cup of coffee