dom/tests/mochitest/pointerlock/test_pointerlock-api.html fails intermittently with the stencil navigation enabled
Categories
(Core :: DOM: Core & HTML, task, P1)
Tracking
()
| Tracking | Status | |
|---|---|---|
| firefox151 | --- | fixed |
People
(Reporter: arai, Assigned: arai)
References
(Blocks 1 open bug)
Details
Attachments
(1 file)
test_pointerlock-api.html fails intermittently at random timing.
The following fails after opening file_locksvgelement.html file, but other runs fail at some other points.
[task 2026-04-01T11:23:08.872+00:00] 11:23:08 INFO - Testing file_locksvgelement.html
[task 2026-04-01T11:23:08.873+00:00] 11:23:08 INFO - file_locksvgelement.html: Resetting fullscreen enter count.
[task 2026-04-01T11:23:08.873+00:00] 11:23:08 INFO - Buffered messages finished
[task 2026-04-01T11:23:08.874+00:00] 11:23:08 INFO - TEST-UNEXPECTED-FAIL | dom/tests/mochitest/pointerlock/test_pointerlock-api.html | Test timed out. - expected PASS
| Assignee | ||
Comment 1•5 months ago
|
||
The test is known to fail on many configuration.
["test_pointerlock-api.html"]
...
skip-if = [
"os == 'win' && os_version == '11.26200'", # Bug 2019745
"opt", # Bug 1833142
"os == 'android'", # Bug 1612553
"os == 'linux' && os_version == '22.04' && arch == 'x86_64' && display == 'wayland'", # Bug 1857057
"os == 'mac' && os_version == '14.70' && arch == 'x86_64' && debug", # Bug 1886190
]
This time it starts failing on debug config, after enabling a pref to make things faster,
and there's already a condition to skip it on "opt".
| Assignee | ||
Updated•5 months ago
|
| Assignee | ||
Comment 2•5 months ago
|
||
This test is known to intermittently fail on various configuration,
and enabling the stencil navigation in bug 2027803 increases the frequency.
The test has already been marked as skip on opt build, which means this test
fails if things work quickly, and indeed the test has several timing dependent
things, such as setTimeout etc, for stabilizing purpose.
The test should be rewritten to use more stable and reliable way.
| Assignee | ||
Comment 3•5 months ago
|
||
The testcase contains timing dependent things, such as the following:
// Try to stabilize the state before running the next test.
SimpleTest.waitForFocus(
() => requestAnimationFrame(() => setTimeout(runNextTest)));
And also some indirection to override the testing functions, in pointerlock_utils.js etc.
The test should be rewritten, to use non-timing dependent thing with clearer architecture.
| Assignee | ||
Comment 4•5 months ago
|
||
there's bug 1886190 for the intermittent failure.
and there's indeed failures on debug, tsan, asan, and ccov.
added with "opt" above, all configuration are known to fail intermittently or permanently.
Comment 7•4 months ago
|
||
Authored by https://github.com/arai-a
https://github.com/mozilla/enterprise-firefox/commit/77755db878aafc8120147e136b1d7551c5bd7d77
[enterprise-main] Bug 2028442 - Skip test_pointerlock-api.html on all configurations. r=smaug
Updated•4 months ago
|
Description
•