[05:10] elopio, balloons: the ADT_AUTOPILOT_MODULE hack is fine for locally running just a single test; I'm just saying we shouldn't use that in official infrastructure [05:11] there --override-control should be used, and is the official way [05:11] that also works locally, of course === chihchun_afk is now known as chihchun === _salem is now known as salem_ === salem_ is now known as _salem [13:53] pitti, ahh yes I couldn't remember the override-control switch, ty === chihchun is now known as chihchun_afk === roadmr is now known as roadmr_afk === roadmr_afk is now known as roadmr [18:23] hi, anyone can tell me what's wrong on here, and how i might be able to fix it? http://pastebin.ubuntu.com/9385488/ [18:57] that looks unfun [18:58] yeah, not nice [18:58] i just want to run the tests to make sure they work :( [19:05] dobey, I would ping jibel or pitti about it if you don't figure it out [19:06] i'll build a utopic .img and try with it instead i guess [19:06] that would be my first guess.. blame the archive / devel nature of vivid [19:12] well, this has to work in vivid too. :) [19:12] and i don't think anything has actually chagned with autopkgtest in vivid [19:12] dobey, well actually it has [19:12] what version are you running/ [19:13] balloons: i'm running trusty [19:13] those errors seem to be from inside the vm though [19:14] dobey, sure, but you sure grab the latest autopkgtest actually [19:14] hmm, and i get same error with utopic vm image [19:15] dobey, this should install on trusty just fine: http://mirrors.kernel.org/ubuntu/pool/main/a/autopkgtest/autopkgtest_3.8.1_all.deb [19:15] see if that doesn't solve the problem [19:15] then of course, ask to get a more modern version in trusty [19:17] well, the problem is updating service stuff during configuration of packages inside the vm [19:17] so i don't see how a new autopilot on the host will help :) [19:17] but trying anyway [19:18] yea, I agree but . . . [19:18] and indeed, it didn't help. same error [19:18] err, s/autopilot/autopkgtest/ [19:18] too many auto* [19:28] meh :-/ [19:34] are the scripts in lp:auto-package-testing no longer used? [19:39] hmm [19:40] balloons: so i guess proposed is broken. removing the --apt-pocket=proposed lets it continue on [19:41] dobey, bah, didn't really think that one through eh? that should have been obvious [19:41] whoops [19:41] typically you would pull specifically what you need out [19:42] balloons: the problem is that autopkgtests get run whenever anything is uploaded in the dependency tree [19:42] balloons: so, proposed is what is used in live deployment [19:43] so i was trying to run with that. i'm surprised it is breaking at that point though with proposed enabled [19:43] seems like nothing would be able to land in that case [19:44] so there must be something else going on there [19:45] https://jenkins.qa.ubuntu.com/view/Vivid/view/AutoPkgTest/ [19:45] notice specifically; https://jenkins.qa.ubuntu.com/view/Vivid/view/AutoPkgTest/job/vivid-adt-apport/ [19:53] balloons: hrmm, but that's not failing due to the same issue [19:54] and i am getting the same insserv issue still without proposed, only it happens much later in the setup [19:54] yea, not quite so easy [19:55] :( [20:00] * dobey wonders what to do there === roadmr is now known as roadmr_afk [20:37] elopio ping, when you have a min, I'm seeing that the sim pin test I need to write is already in u-s-s trunk, however the test may need some brushing up, would like to get your input on what's good to fix http://bazaar.launchpad.net/~ken-vandine/ubuntu-system-settings/less_flaky/view/head:/tests/autopilot/ubuntu_system_settings/tests/test_security.py [20:38] hello Letozaf_ [20:38] balloons, hello :-) [20:38] balloons, I made the changes to the mp [20:40] alesage: first, I would make sure that each test is as small as possible, testing only one workflow and ideally with only one set of assertions at the end. [20:41] alesage: next, keep all the user actions out of the test case class. Put then in a page custom proxy object. [20:41] Letozaf_, ohh cool. I see you added test_cancel_extract_zip_file. can you assert something in that test. the expectations would be the file and folder count stays the same, no extracted folder [20:41] balloons, ok I will [20:41] elopio, ok that's a start thanks, will try to get a review for you in a bit [20:42] alesage: third, and I'm not sure about this part, we might have to make the ofono_phonesim part optional. [20:42] so we can run the tests with a real sim in the real phone. [20:43] elopio, ok, considering [20:43] that's where the config values come into play. Maybe if there is no config value for sim_pin, run the tests with the fake [20:43] if there is one, run them with the real sim. Something like that. [20:43] elopio, ok [20:44] elopio, I'm consciously trying to avoid the config stuff at the moment :) as it seems to be in progress, right? [20:45] Letozaf_, something. A test isn't a test if it doesn't have an assert ;-) [20:46] balloons, you are absolutely right, :( wonder how I forgot that :P [20:46] alesage: in progress, but you can assume that it will be in autopilot configuration [20:46] elopio, ok very well [20:47] from autopilot import get_test_configuration [20:47] sim_pin = get_test_configuration['sim_pin'] [20:47] elopio, thanks for that [20:48] alesage: with veebers branch, it is actually runnable [20:48] https://code.launchpad.net/~canonical-platform-qa/ubuntu-sanity-tests/config/+merge/243830 [20:48] I'm now trying to solve how to get the ppas into the phone. [20:48] elopio, noted [20:48] but it's almost there. [20:49] elfy, does zsync work for you on vivid? [20:51] balloons: checking === roadmr_afk is now known as roadmr [20:52] balloons: yep [20:52] I wonder if it's the image itself... hmm [20:52] the unity8-desktop-next image refuses to zsync [20:52] I end up re-downloading it everytime [20:52] I'd have no idea about that image I'm afraid [20:53] right right.. thanks for the sanity check [20:53] * balloons tries another image [20:53] ubuntu is zsyncing ok [20:54] hmm.. yea ubuntu fails for me too [20:54] argh [20:54] do you have a semi-complete file [20:54] ? [20:54] how do you mean - one that needs to be updated? [20:54] yes [20:54] if so - both were [20:55] well, glad to know it only hates me [20:55] can't even find the download info for the u8 image [20:55] http://cdimage.ubuntu.com/ubuntu-desktop-next/daily-live/current/ [20:56] working - downloading the whole file for me [20:57] balloons: is that right - nothing on the image tracker for unity 8 ? [20:57] definitely local I'd say [21:07] balloons, I pushed a while ago, hasn't built yet [21:08] opps got a filure [21:08] * Letozaf_ is checking [21:08] elopio, does a convenience function exist to scroll to something from within a page? [21:09] alesage, if it is a flickable, jes. [21:10] elopio, more info pls [21:10] looking for a link [21:11] http://bazaar.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/trunk/view/head:/tests/autopilot/ubuntuuitoolkit/_custom_proxy_objects/_flickable.py [21:11] alesage: swipe_child_into_view, swipe_to_top, swipe_to_bottom [21:11] elopio, thanks [21:11] also if you have a reference to the child, and one of it's parents is a flickable [21:11] * alesage bookmarks [21:12] then you can call item.swipe_into_view(), and it will look for the parent where it needs to swipe. [21:13] elfy, right, nothing on image tracker for unity8 image [21:13] k - just making sure :) [21:19] elopio, ok maybe it's not a flickable :) --if it's not, then what is it? [21:21] alesage: maybe a qml type that extends flickable. [21:21] alesage: find in autopilot3 vis the component that you want to swipe. [21:22] it should have atYstart and atYEnd properties [21:22] s/atYstart/atYBeginning [21:23] also a flicking attribute. [21:23] elopio, ok [21:24] after that, find the qml that corresponds to that object. That will tell you how to select it. If it's a type that extends flickable, then you do a python class that extends the ubuntuuitoolkit.QQuickFlickable. [21:25] elopio, it's already ubuntuuitoolkit.UbuntuUIToolkitCustomProxyObjectBase FWIW [21:25] alesage: yes, if it's a cpo for a class that extends flickable, it should be ubuntuuitoolkit.QQuickFlickable instead. [21:26] that way it will inherit all the things from the common cpo base, and the ones from the flickable. [21:26] elopio hmm, so in the QML I'll discover if it's flickable and then change the class if so [21:26] alesage: yup. [21:27] or you will find that the flickable is a child of the page, or you will find that they are doing some crazy nonsense. [21:28] elopio works thanks [21:39] balloons, what about the extract tar feature of filemanager app, is it going to be merged ? [21:42] Letozaf_, I don't see why not. Might be worth posting into the MP [21:44] * Letozaf_ posted into the MP