Open Bug 1400368 Opened 7 years ago Updated 2 years ago

Permafail toolkit/browser/components/tooltiptext/browser_input_file_tooltips.js | Test timed out on Win10

Categories

(Core :: Widget: Win32, defect, P3)

x86_64
Windows 10
defect

Tracking

()

Tracking Status
firefox57 --- fix-optional

People

(Reporter: ahal, Unassigned)

References

Details

https://treeherder.mozilla.org/logviewer.html#?job_id=131094283&repo=try&lineNumber=5406

10:28:26     INFO -  1043 INFO TEST-START | toolkit/components/tooltiptext/tests/browser_input_file_tooltips.js
10:29:11     INFO -  TEST-INFO | started process screenshot
10:29:11     INFO -  TEST-INFO | screenshot: exit 0
10:29:11     INFO -  Buffered messages logged at 10:28:26
10:29:11     INFO -  1044 INFO Entering test bound setup
10:29:11     INFO -  1045 INFO Leaving test bound setup
10:29:11     INFO -  1046 INFO Entering test bound test_singlefile_selected
10:29:11     INFO -  1047 INFO starting test {"value":true,"result":"testfile_bug1251809"}
10:29:11     INFO -  1048 INFO Moving mouse out of the way.
10:29:11     INFO -  Buffered messages finished
10:29:11    ERROR -  1049 INFO TEST-UNEXPECTED-FAIL | toolkit/components/tooltiptext/tests/browser_input_file_tooltips.js | Test timed out -
10:29:11     INFO -  GECKO(6216) | MEMORY STAT | vsize 1738MB | vsizeMaxContiguous 130086749MB | residentFast 202MB | heapAllocated 76MB
10:29:11     INFO -  1050 INFO TEST-OK | toolkit/components/tooltiptext/tests/browser_input_file_tooltips.js | took 45028ms

This was discovered while migrating tests from Windows 8 -> Windows 10 in bug 1397229. This looks similar to bug 1295497, though the timeout is happening earlier on in the test.

If there's no quick fix for this, we'll need to disable it on Windows 10.
The log suggests this is an issue with native mouse event synthesis.
Component: XUL Widgets → Widget: Win32
Priority: -- → P3
Product: Toolkit → Core
Thanks Neil. Do you (or anyone else) have any objections to disabling this test on Windows 10 so we can proceed with the migration in bug 1397229? Otherwise we'll need a fix fairly soon (i.e by around mid-next week).
Flags: needinfo?(enndeakin)
(In reply to Andrew Halberstadt [:ahal] from comment #2)
> Thanks Neil. Do you (or anyone else) have any objections to disabling this
> test on Windows 10 so we can proceed with the migration in bug 1397229?
> Otherwise we'll need a fix fairly soon (i.e by around mid-next week).

No, you can disable it on windows 10 for now, there are already other tests that test the underlying APIs here. I agree with the diagnosis in comment #1, but it's really not clear to me why anything would be different here on Windows 10...

Looking at the screenshot though, why is there an "open file" dialog up? Does this happen on all these failures? (I don't know how to find other trypushes / runs / instances of this failure)

It's also hard to say whether the mouse synthesis did anything or not, because the mouse pointer isn't visible on the screenshot (or at least, I don't see it). Does this happen on Windows 8? The screenshots from the intermittent on windows 8 show the start menu, so it's hard to say if the mouse pointer not making it onto screenshots is a regression caused by the switch to win10 or not.
Flags: needinfo?(ahalberstadt)
No, it doesn't seem to fail on Windows 8.

The switch to Windows 10 hasn't been made yet, so there aren't any other try pushes with this. On my push, it was perma failing. To do your own pushes, you'll need:
https://hg.mozilla.org/try/rev/38bdc14a559ba51030c34a64c179250859ce5901

Though this should be landing today or tomorrow (along with disabling this test on Windows 10).
Flags: needinfo?(ahalberstadt)
Flags: needinfo?(enndeakin)
Severity: normal → S3
You need to log in before you can comment on or make changes to this bug.