[01:30] <dazza5000> anyone able to help me troubleshoot a calendar autopilot test?
[01:30] <dazza5000> the test fails because it thinks the day on my computer is yesterday
[01:31] <dazza5000> but the calendar app selects the correct day for today
[01:31] <dazza5000> and when I run the commands in python outside of the test they give the correct date
[01:39] <dazza5000> this test
[01:39] <dazza5000> autopilot3 run calendar_app.tests.test_monthview.TestMonthView.test_monthview_go_to_today_prev_year
[01:39] <dazza5000> gives this
[01:39] <dazza5000> testtools.matchers._impl.MismatchError: After 10.0 seconds test failed: 22 != 21
[01:40] <dazza5000> datetime.now() gives 2015-03-22 20:40:24.686199
[01:41] <dazza5000> but the test is getting 21 as the result somehow
[16:08] <elopio> I had to join using the webchat.
[16:08] <elopio> for some reason, I can't connect to freenode with quassel.
[16:09] <elopio> if I don't reply quickly, it's because I have no notifications.
[16:16] <paulliu> elopio: hi.. So I'm wondering why fake-apps doesn't start on the AP test for shell-rotation.
[16:18] <paulliu> elopio: and why we use fixtures to start unity8? Can we use ApplicationLifeCycleTestCase with environments set for the fake-sensor?
[16:19] <paulliu> elopio: Currently we copy the code from ApplicationLifeCycleTestCase, but it doesn't work as we expect.
[16:26] <elopio> hello paulliu. brendand, you are taking a look at the fake apps, right?
[16:28] <brendand> elopio, i am
[16:28] <brendand> elopio, right now unity8 doesn't start for some reason
[16:29] <elopio> brendand: that's nice, because you and paulliu have more working hours together (I think). It would be great if you two pair to investigate.
[16:30] <elopio> paulliu: we prefer fixtures over inheritance, because when you inherit test cases things get messy.
[16:30] <paulliu> brendand: why it doesn't start?
[16:30] <elopio> composition is clearer.
[16:30] <paulliu> elopio: ok..
[16:30] <paulliu> brendand: I mean, why unity8 doesn't start?
[16:31] <elopio> paulliu: if there's something useful from a test case setup, you can wrap that in a fixture and share it in both tests.
[16:31] <brendand> paulliu, i don't know yet. it seems to work ok in other tests
[16:31] <paulliu> elopio: ok.. just curious why we don't re-use the code.
[16:32] <elopio> paulliu: I wonder, do your life cycle tests work? is it only on the rotation tests that you get errors?
[16:32] <paulliu> brendand: hmm. you should adb reboot and it might work again.
[16:32] <paulliu> elopio: yes. Life cycle tests works find. It did start the apps.
[16:33] <paulliu> elopio: So I'm wondering if we make some weirdness by start unity8 in different ways.
[16:33] <paulliu> elopio: Just as brendand said, the unity8 might stop launching after you run the test several times. but adb reboot fixes it.
[16:34] <paulliu> elopio: also, the launched unity8 acts weird. If I touch the screen anywhere, it pops the on-screen-keyboard.
[16:35] <paulliu> elopio: In other tests, I'm able to interact with unity8 by touching the screen.
[16:36] <elopio> paulliu: brendand: yes, take a look at the launch_unity in the base UnityTestCase. That's different from the LaunchUnityWithSensors fixture.
[16:37] <elopio> ideally, we should turn the launch_unity in UnityTestCase into a fixture LaunchUnity. And use that fixture from LaunchUnityWithSensors.
[16:40] <elopio> paulliu: the thing with the osk is weird. I can't imagine what might be causing it, but it needs to be fixed too. Unity should be launched as close as possible to a simple initctl start unity8.
[16:42] <elopio> brendand: let me know if you need an extra hand. For now I'll leave you two to it and take a look at the errors I introduced in the new autopilot version :(
[16:56] <dazza5000> does anyone know how to get an ubuntu phone in the US?
[16:58] <teward> dazza5000: i don't think you can - not offered yet last I heard
[16:58] <dazza5000> how about unofficially though?
[16:59] <elopio> dazza5000: there are some places still selling nexus 4.
[16:59] <dazza5000> where?
[16:59] <elopio> http://www.amazon.com/LG-Google-Unlocked-International-Version/dp/B00D94OAPU/ref=sr_1_2?s=wireless&ie=UTF8&qid=1427129945&sr=1-2&keywords=nexus+4
[17:00] <dazza5000> are most people using the 4 to test with?
[17:01] <elopio> dazza5000: we test with it and with the bq phone. I'm not sure which one has more testers.
[17:09] <brendand> elopio, the rotation test seems to use process_helpers.restart_unity and other tests use launch_unity from the base class
[17:10] <brendand> elopio, the latter does something with the mir socket, which might make a difference
[17:19] <dobey> sigh
[17:20] <elfy> good afternoon quality channel peoples
[18:01] <elopio> jibel: the phonedations team will work on nitz to get the clock hour right from your phone carrier. That sounds like a good addition at the end of the wizard test, check that the clock is right.
[18:01] <elopio> jibel: what should I do to propose the addition?
[18:30] <dazza5000> good afternoon elfy
[19:30] <balloons> dazza50002, howdy
[19:30] <elfy> balloons: !!!
[19:30] <elfy> :)
[19:30] <balloons> hey elfy! final beta week, you excitied?
[19:31] <elfy> well ...
[19:31] <elfy> it might be a milestone for you - but down here in the mines - we've already done some :p
[19:31] <dazza50002> hi all
[19:31] <balloons> elfy, I added you to the mail since you spoke up :-)
[19:31] <dazza50002> Need to figure out how to login using yaaic
[19:32] <elfy> mail?
[19:32] <elopio> barry: meeting again.
[19:32] <elfy> balloons: oh - see it now :p
[19:33] <elfy> I tend to only post that stuff on xub-dev list specifically for our peeps
[19:33] <elfy> balloons: but ... http://iso.qa.ubuntu.com/qatracker/milestones/336/builds for the general page :)
[19:34] <balloons> mmm.. indeed
[20:06] <Letozaf_> balloons, hi
[20:10] <balloons> Letozaf_, hello hello
[20:10] <Letozaf_> balloons, howzit ?
[20:10] <balloons> excellent, yourself?
[20:10] <Letozaf_> balloons, fine, thanks :)
[20:10] <Letozaf_> balloons, I have an issue : http://91.189.93.70:8080/job/generic-mediumtests-utopic/2393/testReport/junit/calendar_app.tests.test_dayview/TestDayView/test_default_view/
[20:10] <Letozaf_> balloons, cannot find how to solve it
[20:11] <Letozaf_> balloons, it's about week day
[20:11] <Letozaf_> balloons, on calendar app, can you help or should I ask someone else?
[20:13] <dobey> is there a way in autopilot to do a "wait until object is available, *then* tap it" sort of operation?
[20:14] <dobey> because adding time.sleep() everywhere kind of sucks
[20:14] <veebers> dobey: you can use wait_select_single
[20:15] <veebers> dobey: fyi: http://developer.ubuntu.com/api/devel/ubuntu-14.10/python/autopilot/api/autopilot.introspection.html#autopilot.introspection.ProxyBase.wait_select_single
[20:19] <balloons> Letozaf_, looking
[20:20] <Letozaf_> balloons, thanks
[20:20] <dobey> veebers: hmm, trying that
[20:27] <balloons> Letozaf_, dare I ask, does this work locally?
[20:29] <Letozaf_> balloons, no :( but this happens only on Sundays today the test passes :)
[20:29] <balloons> Letozaf_, right that was my first thought
[20:29] <Letozaf_> balloons, suppose it has to do with the firs day of the week stuff
[20:29] <balloons> it's a locale thing
[20:29] <balloons> yes I agree
[20:29] <balloons> which we should have accounted for already
[20:30] <balloons> perhaps check the logic in the test? Or it's a bug in the app itself.
[20:30] <Letozaf_> balloons, to me looks like an app bug
[20:30] <balloons> This is definitely week 12, so week 11 would be for the week to start on monday
[20:31] <Letozaf_> balloons, If yesterday I ran now.strftime("%U") in python shell I got 12
[20:33] <Letozaf_> balloons, In Italy the first day of the week is sunday
[20:34] <dobey> Today is Boomtime, the 9th day of Discord in the YOLD 3181
[20:36] <balloons> Letozaf_, right. I would file a bug against calendar app and see what mihir says
[20:36] <Letozaf_> balloons, ok thanks I will
[20:36] <dobey> Letozaf_: i'm guessing the locale is not set correctly in the tests perhaps?
[20:37] <Letozaf_> dobey, no Calendar app should display 12 as week number instead it displays 11
[20:38] <Letozaf_> dobey, the test is right, it's Calendar app that in Italy has got the wrong week number
[20:38] <balloons> it's the ultimate revenge, blame the app and be correct!
[20:38] <elfy> LOL
[20:39] <dobey> if the week start is on monday, and it's the sunday prior, the 11 is correct, in that locale, no?
[20:39] <Letozaf_> dobey, in Italy the first week day is Sunday :)
[20:40] <dobey> in most of the world it is
[20:40] <Letozaf_> dobey, :)
[20:42] <dobey> so, if i have a dialog in qml, and in the ap test i'm doing dialog.wait_until_destroyed() for that object, and then take a screen shot after it's been destroyed, shouldn't the screen shot *not* have that dialog in it?
[20:48] <veebers> doeby, assuming you're waiting for the right thing to be destroyed it shouldn't be in the screenshot
[20:48] <veebers> dobey: did wait_select_single help you out btw?
[20:49] <veebers> dobey: also, feel free to ask these autopilot questions in #ubuntu-autopilot too :-)
[20:49] <dobey> yes i think so
[20:50] <dobey> too many channels already :)