=== chihchun is now known as chihchun_afk [00:44] veebers: freshly flashed and the test passed. [00:44] I will try some more after the gym. === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk === chihchun_afk is now known as chihchun === chihchun is now known as chihchun_afk [07:45] halp I can't log in into iso.qa.ubuntu.com [07:45] I tried my launchpad openid login and I'm getting a weird error, "The name mgedmin is already taken" [07:45] I'm not trying to sign up, I'm trying to log in, of course the name is taken, *I took it* [07:46] I had logged in successfully a few months ago, when I was testing utopic images [07:46] also, there's a 404 link for the orgotten password bit: https://bugs.launchpad.net/ubuntu-qa-website/+bug/1418381 [07:46] Launchpad bug 1418381 in Ubuntu QA Website ""Have you forgotten your password?" is a 404 link" [Undecided,New] === chihchun_afk is now known as chihchun [07:49] curiously, I have a 'qa.ubuntu.com' username + password pair saved in my password safe, but it's not accepted by http://iso.qa.ubuntu.com/user/login === chihchun is now known as chihchun_afk [07:50] mgedmin: well there's not anything wrong with the login itself - just had to do the same here [07:51] do what? [07:51] * mgedmin files https://bugs.launchpad.net/ubuntu-qa-website/+bug/1418383 [07:51] Launchpad bug 1418383 in Ubuntu QA Website "Log In tries to register a new user account for an existing user" [Undecided,New] [07:53] manually attempting openid login at /user/login with http://launchpad.net/~mgedmin as my openid url results in the same error: I'm at /user/register with "username already taken" :/ [07:54] can't log in, can't report test test results :( === phillip is now known as Guest45840 === chihchun_afk is now known as chihchun [15:47] brendand: before you EOD, please make sure you review all the cards. I'm taking a look at yours. [15:48] elopio, i think i have [15:49] elopio, i didn't reply to alesage's because i don't have anything to add. i could just leave an approve i guess [15:50] brendand: for the cards that are in the review column? [15:50] elopio, yes [15:51] brendand: I don't see it. Like https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/add-verbose-flag/+merge/248700 [15:52] https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/copy_data/+merge/248679 [16:18] * alesage is doing reviews [16:23] balloons: https://wiki.ubuntu.com/QATeam/DevelopmentSetup/SyncImage [16:47] * balloons loks [16:47] elfy, nice interview btw.. Nice to know a little more history. I loved the opening line of selling from a van :-) [16:47] :) [16:47] where is this linked from elfy ? [16:48] balloons: it's linked from https://wiki.ubuntu.com/QATeam/DevelopmentSetup now [16:48] and :) [16:49] perfect [16:49] I cleaned up the unity8 stuff a bit more and sent round some mail about testing it [16:49] when you say loks I assume you lokked properly at it [16:50] I glanced.. soon enough I will read it [16:51] okey doke === roadmr is now known as roadmr_afk === chihchun is now known as chihchun_afk [18:31] erm [18:31] http://pastebin.ubuntu.com/10077296/ [18:31] elopio, alesage: ^^ any idea why i'm getting an unknown option error for an option that's listed as an option, there? or do i need to bug uitk people? [18:33] dobey: yes, I don't know about that. I think that's the launcher kalikiana made for some tests. [18:34] elopio: yeah, get_toolkit_launcher_command() uses it. i changed to that, because get_qml_launch_command() throws a deprecation warning [18:34] i guess i should just leave it as deprecated for now huh? [18:35] dobey: yes, I'm not sure why using qmlscene is deprecated. [18:36] originally, the launcher was to reproduce some weird scenarios. I think it's ok for you to use qmlscene, which is what your app will be doing in the real world. [18:40] ok [18:50] elopio: http://pastebin.ubuntu.com/10077617/ <- any idea what's going on here exactly? seems the process starts and then immediately exits? [18:52] dobey: it is finding the proxy object for the process, then it's failing to select 'payui' from it. It doesn't seem like the process stops, that shows a different error. [18:52] dobey: if you push it I can take a look. [18:53] elopio: look at lines 3 and 4 in that log. what process is it waiting on there if not qmlscene? [18:54] dobey: right, but I think that it's between lines 2 and 3 that the test is running. [18:54] the traceback doesn't have a timestamp, so that's hard to confirm with the log. [18:55] so autopilot is bad at logging :) [18:55] dobey: try running with -v. That prints more things. [18:55] elopio: it's lp:~dobey/pay-ui/adt-run-ap [18:55] i have no idea how one would pass -v to autopilot when running it via adt-run like this [18:55] dobey: yes, we suck at logging. [18:56] adt-run --click-source . --click ../build-adt-run-ap-UbuntuSDK_for_amd64_GCC_ubuntu_sdk_14_10_utopic-Default/com.canonical.payui_15.01.last_amd64.click -U --apt-pocket proposed --setup-commands ubuntu-touch-session --- qemu ~/adt-vivid-amd64-cloud.img [18:56] dobey: ah, there's a trick for that. It's unsupported and pitti said not to use it, so I always forget about it. [18:56] that is how i'm running the tests [18:56] let me try to find it. [18:57] * balloons wonders if he should give up the goods [18:57] balloons: and become a monk? [18:57] balloons: I was trying to find it myself, as this is like the 10th time I ask you :) [18:57] and walk the earth [18:58] balloons: but go ahead. [18:58] alesage: i'm much better at salting it [18:58] dobey, just don't scorch it [18:59] dobey, you need to pass autopilot args into ADT_AUTOPILOT_MODULE [18:59] so something like ADT_AUTOPILOT_MODULE="-v testsuite" adt-run [18:59] but yes it's secret, bad stuff, undocumented, etc. I believe there may be a better way now to do it [19:00] I see on https://people.debian.org/~mpitt/autopkgtest/README.click-tests.html that you can define the command for running the autopilot tests. [19:00] maybe you would like to put -v there, so it's always verbose. [19:01] yes, you can change it inside the manifest, but you aren't likely to want to do it that way [19:01] that is of course the supported way I guess [19:02] I thought the default output got changed to subunit though? [19:03] eh, yeah, that's still very not nice [19:04] I've finished my test \o/ [19:05] dobey: the click scope tests are still not running on jenkins, right? [19:05] elopio: the scope tests are running, yes [19:06] elopio: pay-ui autopilot tests aren't yet i think [19:06] dobey: great. I've just finished a test to uninstall a click package, for the sanity suite. [19:06] I will me moving some code upstream. [19:06] cool [19:08] alesage: please review this one as soon as you can, so I have time to make any changes you like before EOD [19:08] https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/uninstall_click/+merge/248710 [19:09] elopio: i think the click scope autopilot test runs are still a bit unreliable in jenkins though [19:10] elopio, shall do [19:10] dobey: last thing I knew was the problem with adding the u1 account. But this test doesn't require a u1 account, it install the package with pkcon. [19:11] elopio: right. we aren't running the tests that actually require a u1 account though, are we? [19:11] I need first to add vertical swiping for the scope, and that needs some code in the toolkit. So it will take some time. [19:11] dobey: ah, let me see. [19:12] right, the test with credentials is skipped. [19:13] anyway, i'm more interested in getting a working run of payui autopilot tests under adt-run [19:14] but i guess i'll have to push it off further as well, because i also need to fix this other bug that actually affects users right now [19:14] dobey: yes, adt-run and payui, that will open many things. I'll look at your branch now. [19:17] elopio: oh i guess it's failing because it's probably trying to open online-accounts and that will fail === roadmr_afk is now known as roadmr [19:25] elopio, maybe I missed something about this in stand-up, where do I get dash.DashApps? [19:27] alesage: that comes from unity8, but adt-run should take care of the deps. Are you getting an error? [19:27] elopio, let me paste hang on [19:28] elopio, http://pastebin.ubuntu.com/10078188/ [19:29] I guess we're needing that on import? [19:30] alesage: you have an older version. [19:31] is your host vivid? [19:31] elopio, utopic [19:31] there's something weird there, because the host should never import ubuntu_sanity_tests.tests [19:32] elopio, I think I encountered this at some point going over a veebers branch, [19:32] alesage: yes, we made a mistake somewhere. [19:32] there's a tests import to check something (foggy details) [19:33] I delayed the import but that bit might not have landed b/c yours landed first? [19:33] alesage: I have no idea what you are talking about. [19:34] elopio, ok then :) , I'll investigate a bit [19:34] where do you have the delayed import? [19:35] elopio, let me find [19:35] hum, actually on register_test_options, we are importing tests. So this will always happen. [19:36] alesage, elopio - have you guys hit this issue i'm seeing yet? [19:36] I think we need to remove that feature. There's no way to make it work with an older host. [19:36] good evening [19:37] elopio, I was mistaken--I remember now (and see) that you're delaying that same import [19:37] balloons: which issue? [19:37] elfy, can you confirm: "after booting into xubuntu daily x64 live image, when opening firefox, the start page appears in turkish" ? [19:38] evening to you as well dkessel [19:38] evening balloons :) [19:38] I'm a bit confused as to why we're doing these imports on the test-runner-side at all [19:38] the page it opens is http://start.ubuntu.com/15.04/Google/?sourceid=hp [19:39] or just tests I suppose [19:39] elopio, you can manage to type br right ;) ? [19:40] brendand: not while hungry. [19:40] dkessel: vivid? [19:40] elopio, we had to make fginther leave our channel because he shares *3* initial characters with fgiminez [19:40] elopio, lol [19:40] elopio, it hangs for me [19:40] elfy: yup [19:40] :) [19:40] brendand: ah, right. Not yet. I'll flash after lunch and give it a try. [19:41] came across it while reinstalling my development partition so i can test vivid again ;) [19:41] dkessel: just booting it - been looking at trusty the last few days [19:41] I guess I hope we'll support utopic, else I'll have to do a risky upgrade [19:41] i doubt it is something about the iso really... i guess it would have to be fixed on that web server the page points to [19:42] dkessel: yep - can confirm that [19:42] elopio, this feels blocked for a bit? I'll resume my flight-mode work, or should I dig in here? [19:43] alesage: I'm removing all the things for this config feature that we have never used. [19:43] I'll ping you soon. [19:43] elopio, ok [19:43] dkessel: I would check ubuntu - but that's unbootable [19:43] balloons: what do you think about it? ^ report a bug, or tell someone at canonical to fix that web server? ;) [19:43] elfy: good to know - i won't even try i then [19:48] is it a geoip issue? [19:48] I'm slightly struggling to follow this. [19:50] grab xubuntu image - open web-browser - get turkish [19:50] post or pre install? [19:50] live session [19:52] http://i.imgur.com/9jYPEtL.png [19:52] balloons: ^^ [19:53] * balloons boots [19:55] does it stay like that after boot? ohh.. that's actually different than I thought [19:57] heh [20:00] alesage: can you please try again merging this? [20:00] https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/remove_tests_import/+merge/248817 [20:01] elopio, ok [20:02] mhh maybe that's part of the default firefox profile somehow? though the settings said my preferred language was english... [20:03] pretty sure it's not that [20:07] brendand: getting stuck, you are right. [20:07] I hate this. Every single week, it breaks when we are about to finish. [20:08] this looks like unity not being able to start. [20:30] elopio, it's weird - i've gone back to old revision of vivid, gone back many branches [20:30] elopio, semi-glad it's not just me, but it's really frustrating [20:30] elopio, it's getting stuck in adt-run right? [20:31] brendand: I'm investigating. I think it's a change they did on the lightdm mock. [20:31] elopio, last thing i see is : Loading tests from: /tmp/adt-run.mKS8sN/build.Xxo/real-tree [20:31] elopio, ahhh. so dev teams can break our s**t by changing their packages [20:31] elopio, no fair :P [20:31] brendand: totally. [20:31] that's why they have to run this test suite as part of their release process. [20:32] and we need to be smart about the code we keep in this branch. [20:32] elopio, well seems you have a better idea than i do. if you find a workaround please send a mail to the list [20:32] elopio, i don't want to be still stuck tomorrow [20:32] elopio, even if you don't, send your findings, any info will be useful [20:32] if I'm right, is in good part my fault for using self._lightdm_mock_type = None in our branch. It's clearly a private thing that we shouldn't be touching. [20:33] brendand: yes, I will. [20:35] if the symptom is hanging and a black screen I may be affected also [20:35] elopio: do you know how we can quickly get autopilot tests running in jenkins for payui MPs? [20:36] dobey: quickly, ask ci to add payui to their jobs. But that will not use autopkg, it will do something like phablet-test-run. [20:38] elopio: right, we don't have autopkgtest set up in trunk yet anyway. i want it for another MP i'm about to make. i made a small change to the mock test server to better reflect reality, but i can't verify that the tests still pass [20:40] dobey: yes, you'll need a config like the one you have for click scope tests. Ask the ci vanguard and he'll make it. It will be easier if you have a make task like the one in the scope too. [20:41] elopio: yeah, there's a make autopilot [20:59] elopio, this flash is taking a while, will have your review by EOD though [21:00] alesage: if you flashed, you will need my new branch too. [21:00] too many fires today. [21:01] elopio, meaning to install on device? [21:01] alesage: newest image has newest unity which breaks our tests. [21:03] elopio, ok then--what's your new branch? [21:03] alesage: in progress. I still haven't found a good workaround. [21:05] elopio ok that's good info, I'll stop flashing and resume flight-mode just to get that MP in (current flash is write-mode--do you know if there's a way to 'undo' write mode?) [21:06] alesage: you have to umount the partition. I don't know the path of the partition though. [21:06] elopio, k [22:23] elopio, any news about the issue? [22:25] brendand: still debugging. [22:25] it just seems unity doesn't want to start when called through adt run. [22:25] but it works just ok when done via adb shell [22:25] might be that we are doing it too fast. [22:26] I don't really know now. I will just keep trying things. [22:26] elopio, did you notice that neither mir socket is present? [22:26] elopio, either in /tmp or /run/user [22:27] brendand: yes, the log says that mir was killed. [22:27] elopio, that's not supposed to happen is it? [22:27] brendand: no. And the weird thing is that the whole unity suite succeeds. [22:28] that's why I suspect of stopping and starting too fast. The only difference I can see is that we stop during test set up. [22:28] will do some more runs removing parts, I still have some hours today. [22:29] hmm, no, it shouldn't be trying to open online-accounts here [22:55] oh well, time to go. later all