Fix increased test failure after turning off browser.theme.native-theme
Categories
(Core :: Widget: Cocoa, defect)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox139 | --- | fixed |
People
(Reporter: julianwels, Assigned: julianwels)
References
Details
Attachments
(2 files)
The subtest helper_click_interrupt_animation.html of gfx/layers/apz/test/mochitest/test_group_zoom-2.html shows increased failure rates after landing bug 1955113.
| Assignee | ||
Comment 1•7 months ago
|
||
The subtest helper_click_interrupt_animation.html of gfx/layers/apz/test/mochitest/test_group_zoom-2.html shows increased failure rates after landing bug 1955113, which this patch attempts to address.
| Assignee | ||
Comment 2•7 months ago
|
||
Did a try run with this patch yesterday[1] that came back okay, since it's a test that fails intermittently, I started a second one[2] just to be sure.
[1] https://treeherder.mozilla.org/jobs?repo=try&revision=e9c25efb1324e3aade54cecdf9068c68e9bf2e6a
[2] https://treeherder.mozilla.org/jobs?repo=try&revision=0b95d151d2422bdcc1ee1b67ed538343f32c09c7
Updated•7 months ago
|
Comment 4•7 months ago
|
||
| bugherder | ||
Comment 5•7 months ago
•
|
||
Julian, this needs further attention as the failure rate looks unchanged, treeherder link.
Failure log looks to be the same: https://treeherder.mozilla.org/logviewer?job_id=503427662&repo=autoland
Comment 6•7 months ago
•
|
||
I think it's clear that the issue is not with the browser UI style changes themselves, but rather there is an issue either in the test or in the scrolling implementation which those changes tickle.
As such, rather than backing out the style changes, I think it would make sense to disable the test on the affected platform (Mac) until the underlying issue can be investigated in more detail in bug 1959553.
I will post a patch to do this.
Comment 7•7 months ago
|
||
Comment 8•7 months ago
|
||
Fwiw I did comment with this elsewhere, but the difference is that with those changes the browser chrome is one device pixel smaller, so it's not crazy
Comment 9•7 months ago
|
||
I checked screenshots taken at the time the test failed in the runs in comment 5. All of them are 1920x1080. On our CI there is another mac instance whose resolution is 4k, at least there was, so maybe this failure is not intermittent, it might be rather perma on 1920x1080 resolution macs.
Comment 10•7 months ago
•
|
||
(In reply to Emilio Cobos Álvarez (:emilio) from comment #8)
Fwiw I did comment with this elsewhere, but the difference is that with those changes the browser chrome is one device pixel smaller, so it's not crazy
I figure that's a potentially legitimate consequence of the style changes. If it's unintended, then maybe there is additional investigation to do on the front-end side. The impact of a 1px browser chrome height change on this scrolling test is worth understanding/investigating either way.
Comment 11•7 months ago
|
||
Comment 12•7 months ago
|
||
| bugherder | ||
| Assignee | ||
Updated•7 months ago
|
Updated•7 months ago
|
Description
•