Closed Bug 1383056 Opened 3 years ago Closed 3 years ago

Intermittent TEST-UNEXPECTED-PASS | /cssom-view/elementFromPoint.html | Fieldsets - expected FAIL

Categories

(Testing :: web-platform-tests, defect)

Version 3
defect
Not set
normal

Tracking

(firefox56 fixed)

RESOLVED FIXED
mozilla56
Tracking Status
firefox56 --- fixed

People

(Reporter: intermittent-bug-filer, Assigned: jgraham)

Details

(Keywords: intermittent-failure, Whiteboard: [stockwell disabled])

Attachments

(1 file)

These were apparently only running against m-c until a patch currently on autoland turned them on elsewhere. So expect more failures to flood in now.
james: can you take a look, this is happing frequently
Flags: needinfo?(james)
Comment on attachment 8889309 [details]
Bug 1383056 - Disable unstable elementFromPoint test on win 10,

https://reviewboard.mozilla.org/r/160376/#review165616

thanks james!
Attachment #8889309 - Flags: review?(cbook) → review+
Pushed by cbook@mozilla.com:
https://hg.mozilla.org/integration/autoland/rev/00eedf2ff919
Disable unstable elementFromPoint test on win 10, r=Tomcat
Flags: needinfo?(james)
Whiteboard: [stockwell disabled]
https://hg.mozilla.org/mozilla-central/rev/00eedf2ff919
Status: NEW → RESOLVED
Closed: 3 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla56
Assignee: nobody → james
So there is definitely an issue here, and I suspect it's the test, not Firefox.  If you look  at https://wpt.fyi/results/css/cssom-view/elementFromPoint.html?aligned&label=stable every single browser fails the "Fieldsets" test over there.  But when I test locally by just loading http://w3c-test.org/css/cssom-view/elementFromPoint.html I get a pass in Firefox, Chrome, and Safari (didn't test Edge).
Flags: needinfo?(james)
I think looking at stable releases here is misleading, for prerelease browsers there are a lot of passes: https://wpt.fyi/results/css/cssom-view/elementFromPoint.html?label=experimental

On linux using nightly I can't make the test fail, and it looks like the issue is almost entirely on Windows 10. My guess is that it's either that the test is making inaccurate assumptions about what the control looks like on Windows, or that it's somehow focus related.
I can't make the test fail on Windows 10 either. Given that this was disabled a year ago and the test has different behaviour in nightly and stable, I'm guessing it's OK to reenable the test on Windows now.

https://treeherder.mozilla.org/#/jobs?repo=try&revision=1f191696db618f776e1d95a062c19b4348170227
Flags: needinfo?(james)
> Given that this was disabled a year ago and the test has different behaviour in nightly and stable

Hmm.  I was getting passes in stable too, but I do see that in wpt automation at least it passes in nightlies.

Filed bug 1502948 to re-enable.
You need to log in before you can comment on or make changes to this bug.