Closed Bug 1394228 Opened 4 years ago Closed 4 years ago

Intermittent dom/events/test/pointerevents/test_touch_action.html | Test timed out.

Categories

(Core :: DOM: Events, defect, P2)

defect

Tracking

()

RESOLVED FIXED
mozilla59
Tracking Status
firefox59 --- fixed

People

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

References

Details

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

Attachments

(1 file)

This bug has failed 38 times in the last 7 days, Fails on Windows 10-64, build types are pgo and opt.

Here is a recent log:
https://treeherder.mozilla.org/logviewer.html#?repo=mozilla-inbound&job_id=149232626&lineNumber=5221

And here is a part of the log:

01:18:12     INFO -  Buffered messages logged at 01:13:12
01:18:12     INFO -  2855 INFO TEST-PASS | dom/events/test/pointerevents/test_touch_action.html | pointerevent_touch-action-inherit_parent-none_touch-manual.html | Synthesized vertical drag
01:18:12     INFO -  Buffered messages finished
01:18:12    ERROR -  2856 INFO TEST-UNEXPECTED-FAIL | dom/events/test/pointerevents/test_touch_action.html | Test timed out.
01:18:12     INFO -      reportError@SimpleTest/TestRunner.js:121:7
01:18:12     INFO -      TestRunner._checkForHangs@SimpleTest/TestRunner.js:142:7
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      setTimeout handler*TestRunner._checkForHangs@SimpleTest/TestRunner.js:163:5
01:18:12     INFO -      TestRunner.runTests@SimpleTest/TestRunner.js:380:5
01:18:12     INFO -      RunSet.runtests@SimpleTest/setup.js:194:3
01:18:12     INFO -      RunSet.runall@SimpleTest/setup.js:173:5
01:18:12     INFO -      hookupTests@SimpleTest/setup.js:266:5
01:18:12     INFO -  parseTestManifest@http://mochi.test:8888/manifestLibrary.js:36:5

:overholt Can you please take a look at this bug?
Flags: needinfo?(overholt)
Stone knows pointer events best so hopefully he can find time to take a look.
Flags: needinfo?(overholt) → needinfo?(sshih)
Priority: -- → P2
The fail rate increased a lot recently. I'll check this in these days.
There are two problems with this intermittent failure. The failures are timeout when running pointerevent_touch-action-button-test_touch-manual.html and pointerevent_touch-action-inherit_highest-parent-none_touch-manual.html
Flags: needinfo?(sshih)
Regarding the case about pointerevent_touch-action-button-test_touch-manual.html, found something [1] interesting. Here are partial logs.
11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 0
11:32:40     INFO -  GECKO(2832) | InitTouchInjection ok
11:32:40     INFO -  GECKO(2832) | InjectTouchPoint
11:32:40     INFO -  GECKO(2832) | OnTouch
11:32:40     INFO -  GECKO(2832) | OnTouch touch move
11:32:40     INFO -  GECKO(2832) | OnTouch touch start

On Windows, we push preference 'apz.test.fails_with_native_injection' to true. However, the value is changed while synthesizing touch events for the test. This happened when the test failed. Synthesizing touch events by sending fake touch events can reduce the failure rate a lot.

[1] https://public-artifacts.taskcluster.net/E3otp40QSvKA-Rz7ThTRdw/0/public/logs/live_backing.log
(In reply to Ming-Chou Shih [:stone] from comment #25)
> Regarding the case about
> pointerevent_touch-action-button-test_touch-manual.html, found something [1]
> interesting. Here are partial logs.
> 11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
> 11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
> 11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
> 11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
> 11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
> 11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 1
> 11:32:40     INFO -  GECKO(2832) | SynthesizeNativeTouchPoint 0
> 11:32:40     INFO -  GECKO(2832) | InitTouchInjection ok
> 11:32:40     INFO -  GECKO(2832) | InjectTouchPoint
> 11:32:40     INFO -  GECKO(2832) | OnTouch
> 11:32:40     INFO -  GECKO(2832) | OnTouch touch move
> 11:32:40     INFO -  GECKO(2832) | OnTouch touch start
> 
> On Windows, we push preference 'apz.test.fails_with_native_injection' to
> true. However, the value is changed while synthesizing touch events for the
> test. This happened when the test failed. Synthesizing touch events by
> sending fake touch events can reduce the failure rate a lot.
> 
> [1]
> https://public-artifacts.taskcluster.net/E3otp40QSvKA-Rz7ThTRdw/0/public/
> logs/live_backing.log

It seems the test case got scroll event, called async_test.done, and close the window before the chrome process handled all requests to synthesize touch events.
looking forward to this landing soon!
Attachment #8936541 - Flags: review?(bugmail)
Attachment #8936541 - Flags: review?(bugmail) → review+
Duplicate of this bug: 1424202
Pushed by sshih@mozilla.com:
https://hg.mozilla.org/integration/mozilla-inbound/rev/d55477dfbd5a
Intermittent test_touch_action.html test timed out. r=kats.
https://hg.mozilla.org/mozilla-central/rev/d55477dfbd5a
Status: NEW → RESOLVED
Closed: 4 years ago
Resolution: --- → FIXED
Target Milestone: --- → mozilla59
Whiteboard: [stockwell disable-recommended] → [stockwell fixed:timing]
Blocks: 1394229
Blocks: 1371671
See Also: → 1512032
See Also: → 1672141
You need to log in before you can comment on or make changes to this bug.