[13:52] <kalikiana> ubuntu-qa, ping, I need help with autopilot failing to reliably hit buttons on screen https://code.launchpad.net/~ubuntu-sdk-team/ubuntu-ui-toolkit/orientationManual/+merge/254753 -> https://jenkins.qa.ubuntu.com/job/generic-deb-autopilot-runner-vivid-mako/2718/
[13:55] <brendand> kalikiana, let me see
[13:55] <rhuddie> kalikiana, can you reproduce the same issue when you run the tests locally?
[13:55] <brendand> kalikiana, are all those failures relevant?
[13:57] <kalikiana> rhuddie: not all of them. but locally the button_huge reliably hits the button_small instead and fails in the same way as in CI
[13:57] <kalikiana> brendand: yes
[13:57] <kalikiana> brendand: except for the 2 about writing in a textfield
[13:57] <rhuddie> kalikiana, is this when you're doing shell rotation?
[13:58] <brendand> rhuddie, do you want to look at it?
[13:58] <rhuddie> brendand, yeah, i can take a look
[13:59] <kalikiana> rhuddie: no shell involved
[13:59] <kalikiana> only the window contents
[14:00] <rhuddie> kalikiana, we've seen strange issues like this when you get the screen co-ordinates whilst an object is still moving on the screen, like an animation or something, and then when the press is done, the object is in a different place
[14:08] <kalikiana> rhuddie: even adding a sleep however doesn't fix it.
[14:10] <rhuddie> kalikiana, seems very strange. I would need to try and run it to see what is going on.
[14:11] <rhuddie> kalikiana, I seem to remember there is an animating property, could be worth trying. like: object.animating.wait_for(False)
[14:12] <kalikiana> rhuddie: it won't fix anything that a 10 second delay didn't fix, though...
[14:12] <rhuddie> kalikiana, nope.
[14:14] <rhuddie> kalikiana, regarding my OSK mp, I made those review updates. I've been running app tests, they seem fine. https://code.launchpad.net/~canonical-platform-qa/ubuntu-ui-toolkit/textfield-use-osk/+merge/261098
[14:21] <kalikiana> rhuddie: you removed the language setting.. how do you know there will be a latin-based keyboard?
[14:25] <rhuddie> kalikiana, I was not sure about forcing it to change language, if it had been set to some other. So if a key was not available it would fail when trying to type it.
[14:27] <kalikiana> rhuddie: yes. which will potentially was a lot of time trying to debug. for the same reason context menu tests run in english
[14:27] <rhuddie> kalikiana, ok, if there's a good reason to keep it in english, I'll set it back
[14:28] <kalikiana> with the slight difference that the app locale can be overridden without changing any setting
[14:29] <kalikiana> rhuddie: changing permanent settings is always ugly, but still better than surprising false negatives
[15:41] <rhuddie> kalikiana, i just tried running that test. what i see is that when the popup page rotates from portrait to landscape, the buttons are not shown on the screen any more. did you get that?
[16:01] <kalikiana> rhuddie: the buttons are shown here - as I said above it actually passes except it hits the wrong button in one case