[05:05] vila: it worked the second time I ran it. [05:05] the first time it gave an error about timing out waiting for login. Not sure what that's about. === chihchun_afk is now known as chihchun [05:39] elopio: O_o you got a log or something ? [05:39] * vila yawns === willcooke is now known as francooke === francooke is now known as JupiterWhistles === JupiterWhistles is now known as willcooke === qwebirc916967 is now known as slickymasterWork === qwebirc422687 is now known as slickymasterWork === pgraner-afk is now known as pgraner [13:19] elopio: did I mention that with the latest changes the same dep8 tests runs on the phone, the archive and -proposed ? [13:20] vila: \o/ [13:20] ^_^ [13:20] so we are ready to get the MP reviewed by the dev, right? [13:20] that's what I asked you in my last comment yes [13:21] elopio: keep in mind that since those are new dep8 tests, they won't block promotion if they fail at first [13:21] elopio: they will need to succeed at least once and *thne* they'll block ;) [13:22] oh, that's handy. [13:22] kudos to jibel&pitti [13:22] vila: lets get a review from oSoMoN and pitti. [13:22] https://code.launchpad.net/~canonical-platform-qa/webbrowser-app/autopkgtest/+merge/256858 [13:22] wut? [13:23] also, anybody else from ubuntu-qa who wants to chime in ^ [13:23] oh yes, ISTR that it was mentioned on the ML, but after two sprint weeks I'm hopelessly behind in mails [13:23] pitti: it's a branch to run the browser autopilot tests as dep8. [13:23] right, I'm not sure all the tricks I used are really... well, I'd be happy to have pitti's feedback on them indeed ;) [13:23] elopio: nice work! [13:26] pitti: well, lots of people involved to get this to work, with yourself first in the list :) [13:27] * vila nods . o O (ubuntu-touch-sessioin) [13:29] elopio, bing [13:29] elopio, i chimed. nothing major [13:29] balloons: bong. [13:29] oh boy balloons is going to be deeply confused [13:30] brendand_: agree with your comment. [13:30] * balloons scratches head [13:30] and btw, now we are in a better shape to document this in a central place instead of the README. [13:30] do you have any recommendations for a good place in the wiki? [13:31] pitti, did alesage ping you about showing off adt as a demo session @ UOS? [13:31] elopio, a good place for ? [13:31] elopio: by the way, you mentioned a login issue on first run ? Details ? Did you get logs for that ? [13:34] * balloons suddenly feels the bong was for brendand_ [13:34] * balloons feels left out [13:36] balloons, hey whoah man - i don't do drugs okay! [13:38] elopio, vila: reviewed the MP [13:38] balloons: no, not yet [13:39] pitti, we're soliciting folks to give cool demos, and alesage mentioned perhaps showing off adt would make a nice demo. It would be cool to see the different / novel ways it can be used [13:40] and/or if you have some other pitti created goodness to show off, feel free :-) [13:40] pitti: ghaaaa, I tried set +a instead of set -a.... [13:40] just boring boot stuff last cycle [13:41] so if anything there, I could demo some ways how to do boot speed and debug stuff now [13:41] balloons: good place to write the steps to run the dep8 tests and analyze the results. [13:42] pitti, ohh right.. I imagine the bootspeed chart and hackery has changed eh [13:42] pitti: thanks for the review ;-D I stumble across many things and didn't mention them all (but some hints in the history), you found them all and then some \o/ [13:43] pitti: about needs-recommends, I wondered if python3-evdev should move to Depends in autopilot-touch ? [13:44] elopio, hmm.. we have a page on autpkgtest already, let me dig it up. It might make sense there [13:44] https://wiki.ubuntu.com/Touch/Testing#Running_tests_with_autopkgtest [13:45] http://packaging.ubuntu.com/html/auto-pkg-test.html [13:45] those are the two I'm aware of, aside from the docs that autopkgtest ships itself [13:46] pitti: i.e. I thought using needs-recommends may install too much or something (will use it for now) [13:50] pitti: ahh, right. That's actually a good place to document about trv. [13:50] we should probably clean it up a little and push it to the archive. [13:52] elopio, yea pitti linked the page I was thinking of [13:52] and yea, upstream as well: http://anonscm.debian.org/cgit/autopkgtest/autopkgtest.git/tree/doc/README.package-tests.rst [13:54] jibel: http://d-jenkins.ubuntu-ci:8080/view/Vivid/view/AutoPkgTest/job/vivid-adt-libanyevent-cachedns-perl/ [13:56] pitti: ADT_NORMAL_USER is only set for setup-commands, not for tests (IIRC my late-night readings). Should it ? [14:09] vila: err yes, it absolutely should; it's not supposed to be set for setup-commands [14:09] (it might be, but not specified) [14:10] vila: oh sorry, you are right [14:10] I totally misremembered what this was for [14:20] pitti: so, the next best choice after ADT_NORMAL_USER is USER or LOGNAME ? [14:21] vila: yes, the solution in the branch seems fine [14:21] ack === chihchun is now known as chihchun_afk [15:02] vila: you might prefer the command line tool corey wrote. [15:03] https://github.com/cgoldberg/subunitdetails [15:03] hmm [15:04] I'm torn between my personal preferences and the greater good ;) Sounds like enhancing trv may benefit more people no ? [15:06] vila: yes. And trv should probably be a UI for that library. [15:06] vila: do you have a trace for that Mouse error? [15:06] I want to see where it decides it should create a Mouse instead of a Touch. [15:06] elopio: yup, but in standup right now ;) [15:07] elopio: ack, I'm re-running without needs-recommends to isolate [15:40] elopio: so, one traceback is https://pastebin.canonical.com/130219/ but I've seen the message for tests that succeeded O_o [15:41] elopio: I've push my latest change, if you use revision 978 you should be able to reproduce on a phone [15:41] elopio: reverting needs-recommends fixes the issue [15:42] pitti: ^ needs-recommends is too greedy, not fully diagnosed yet but it seems to install too much which probably confuse autopilot [15:49] vila: hm, that might be a problem in itself, but I guess you instaed just explicitly add the deps you need? [15:50] pitti: yup, it's only python3-evdev at that point (you may have missed my previous question, let me re-paste) [15:50] pitti: about needs-recommends, I wondered if python3-evdev should move to Depends in autopilot-touch ? [15:50] vila: for -touch, I think yes [15:51] vila: recommends on such meta-packages are a bit pointless [15:51] i.e. I'm unclear about what the package does but ... right will file a MP explaining my issue and see what the reviewers say ;) [15:51] ha [15:52] I'm also unclear about whether this meta-package is mentioned somewhere to get installed... [15:52] pitti: why is this pointless ? (That may wait for next week if you don't have time to reply now ;) [15:53] vila: well, metapackages sohld install a group of packages that you need for this case, and make up its mind if it wants a recommended one or not [15:53] vila: but yeah, I suppose it's perfect bikeshedding material :) [15:53] vila: I think the more interesting issue is why ap3 breaks when you install all recommends [15:54] pitti: yup, that's the one elopio wants to look at more closely [15:54] elopio: correct ? [15:55] vila: yes. Now I'm wondering why on test end it's moving the mouse. [15:55] it's in the way ? [15:56] ;-D [15:56] I suppose this comes from earlier, it somehow instantiates the Mouse on setup and tries to clean it up. [15:56] I'll get a phone and debug. [15:59] elopio: https://bugs.launchpad.net/autopilot/+bug/1447688 [15:59] Ubuntu bug 1447688 in Autopilot "python3-evdev should be a Depends for autopilot-touch" [Undecided,New] [16:01] vila: I'm not yet sure how to define a dependency on autopilot-touch [16:01] we need to install autopilot-touch only if we are on the phone, bug debs can't depend on that, only on architecture, right? [16:02] how would we know if we have a touchscreen or not? [16:04] yeah, it's unclear to me too, we may want to install it explicitly when we're not using ubuntu-touch-session (still hackish) [16:05] It's a different axis, we use ubuntu-touch-session to setup a screen, not an input device [16:06] I don't think it matters much for now but it's something to keep in mind === alecu is now known as alecu-lunch === bladernr-lex is now known as bladernr === bladernr is now known as bladernr_ [17:24] ubuntu-qa: meeting with balloons in 30 minutes to talk about docs for the test helpers. [17:42] elopio: thx for highlighting that, I think everyone who participates in developing those should attend, likely everyone in projects-team no? [17:43] where is the meeting taking place ? [17:43] jfunk: also ops should be interested, as they can quickly write a high level test to reproduce an issue. === pgraner is now known as pgraner-afk [17:44] ki7mt: google hangouts. [17:44] ki7mt: I saw your email about wanting to write tests. Welcome, and please ping if you have questions. [17:45] elopio, thanks, I'm sure there will be many questions, but have a fare amount of reading to do first :-) [17:48] ops-team ^ [17:49] ops-team is really really busy with silos [17:53] not a good week for meetings could be why jibel cancelled them all === alecu-lunch is now known as alecu [17:59] balloons: do you have a hangout link? [18:01] nuclearbob: https://plus.google.com/hangouts/_/canonical.com/leo-nicholas [18:04] isn't it now? [18:04] balloons: we are waiting for you. [18:04] yes, google hates m [18:11] now internet hates me, the hangout seems fine but I'm talking to myself [18:11] no worries, we'll wait [18:19] elopio, that's ok m8, I talk to myself all the time :-) === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [18:34] ki7mt, nice to see you here as well. You're going to enjoy test writing :-) [19:23] balloons, Hi, yes, I think so, will be a steep learning curve, but I like those sort of challenges.