Created attachment 652526 [details] Screenshot STR (1) Get otoro build from inbound ~early this morning (11/16) (2) Launch settings app See attached screenshot --- we don't fully repaint the screen. Just got an otoro again so I'm not sure if this is new or not. Vivien, do you know?
Probably goes without saying that scrolling up and down a little makes us repaint properly.
Trying from before bug 539356, candidate 1.
https://hg.mozilla.org/integration/mozilla-inbound/rev/ed614ea130c0 doesn't exhibit the bug.
https://hg.mozilla.org/mozilla-central/rev/22f6adf8685b is clean. Not bug 539356, yay.
https://hg.mozilla.org/mozilla-central/rev/4db58ba7848f exhibits the bug So initial regression range is http://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=22f6adf8685b&tochange=4db58ba7848f
Sorry, not everything is merged http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=22f6adf8685b&tochange=4db58ba7848f
Awesome! $ hg bisect -g Due to skipped revisions, the first bad revision could be any of: changeset: 102444:a0bddf5fcb91 parent: 102443:b9b3b3373c08 parent: 102428:3b46b03dff5c user: Kyle Huey <firstname.lastname@example.org> date: Wed Aug 15 06:25:20 2012 -0700 summary: Merge m-c to b-s. changeset: 102445:d67c02074ced user: Kyle Huey <email@example.com> date: Wed Aug 15 06:52:06 2012 -0700 summary: Bug 780618: Follow up to switch to nsError.h in merged patches. r=me
khuey, (if the bisect isn't bogus,) was this a nontrivial merge by any chance?
The merge was straightforward ... I think the bisect took a wrong turn somewhere.
Narrower regression range http://hg.mozilla.org/integration/mozilla-inbound/pushloghtml?fromchange=dad98b444955&tochange=62bb2e74fe67 Suspect bug 685516.
That's what I would suspect.
The first bad revision is: changeset: 102435:5c730c1f2274 user: Kyle Huey <firstname.lastname@example.org> date: Mon Aug 13 15:04:19 2012 -0700 summary: Bug 697230: Part 1 - Centralize style image observers. r=bz khuey, can you take a look?
Note: testing with gaia rev b3153b347c7b9357f26c0584e676fee3ce5d8aa0
If your bisection is accurate, I suspect this is a dupe of bug 783379 (and unrelated to the tests I disabled).
Possibly. The sequence of events on opening the "settings app" is 1. create iframe 2. set iframe.src = 'app://settings' (approximately) 3. add to enclosing document 4. [do some unrelated animation] 5. call iframe.setVisible(), which ends up doing docShell.active = true 6. call iframe.focus() So in between (3) and (5) the iframe looks like an inactive "tab".
Ok. I'll take this and look into it after I fix bug 783162.
Fixed by bug 783379.