1446293 - Intermittent testing\marionette\harness\marionette_harness\tests\unit\test_navigation.py TestBackForwardNavigation.test_non_remote_about_pages | JavascriptException: TypeError: win is null

Status: RESOLVED INCOMPLETE

(Testing :: Marionette Client and Harness, defect, P5)

Description

[Treeherder Bug Filer]

Filed by: rgurzau [at] mozilla.com

https://treeherder.mozilla.org/logviewer.html#?job_id=168413111&repo=mozilla-inbound

https://queue.taskcluster.net/v1/task/bj7kT0fnTWie8BtrQulctA/runs/0/artifacts/public/logs/live_backing.log

06:50:26    ERROR -  TEST-UNEXPECTED-ERROR | testing\marionette\harness\marionette_harness\tests\unit\test_navigation.py TestBackForwardNavigation.test_non_remote_about_pages | JavascriptException: TypeError: win is null
06:50:26     INFO -  stacktrace:
06:50:26     INFO -  	execute_script @test_navigation.py, line 111
06:50:26     INFO -  	inline javascript, line 16
06:50:26     INFO -  	src: "              if (win.BrowserApp) {"
06:50:26     INFO -  	Stack:
06:50:26     INFO -  	@test_navigation.py:16:1
06:50:26     INFO -  	@test_navigation.py:0:2
06:50:26     INFO -  	evaluate.sandbox/promise<@chrome://marionette/content/evaluate.js:163:13
06:50:26     INFO -  	evaluate.sandbox@chrome://marionette/content/evaluate.js:114:17
06:50:26     INFO -  	GeckoDriver.prototype.execute_@chrome://marionette/content/driver.js:1024:19
06:50:26     INFO -  	async*GeckoDriver.prototype.executeScript@chrome://marionette/content/driver.js:887:27
06:50:26     INFO -  	Async*despatch@chrome://marionette/content/server.js:288:20
06:50:26     INFO -  	async*execute@chrome://marionette/content/server.js:262:11
06:50:26     INFO -  	async*onPacket/<@chrome://marionette/content/server.js:237:15
06:50:26     INFO -  	async*onPacket@chrome://marionette/content/server.js:236:8
06:50:26     INFO -  	_onJSONObjectReady/<@chrome://marionette/content/transport.js:500:9

Comment 1

[Intermittent Failures Robot]

3 failures in 814 pushes (0.004 failures/push) were associated with this bug in the last 7 days.    

Repository breakdown:
* mozilla-inbound: 2
* mozilla-central: 1

Platform breakdown:
* windows10-64: 3

For more details, see:
https://brasstacks.mozilla.com/orangefactor/?display=Bug&bugid=1446293&startday=2018-03-12&endday=2018-03-18&tree=all

Comment 2

[Intermittent Failures Robot]

22 failures in 822 pushes (0.027 failures/push) were associated with this bug in the last 7 days.    

Repository breakdown:
* autoland: 8
* try: 7
* mozilla-inbound: 4
* mozilla-central: 3

Platform breakdown:
* windows10-64: 19
* windows10-64-nightly: 2
* windows10-64-devedition: 1

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1446293&startday=2018-03-20&endday=2018-03-26&tree=trunk

Comment 3

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Attachment: screenshot

All the failures are happening on Windows 10 only, and what fails here is that `RecentWindow.getMostRecentBrowserWindow()` returns `null`. It means we are talking about that code:

https://dxr.mozilla.org/mozilla-central/rev/b906009d875d1f5d29b0d1252cdb43a9b1a5889c/browser/modules/RecentWindow.jsm#56-62

This can only happen when `getZOrderDOMWindowEnumerator()` returns no elements. But the screenshot shows that there is a window with 2 tabs. 

I have a hard time to find the definition of that method. Neil, do you know if something changed in that code lately?

Comment 4

[Neil Deakin]

Did this just start failing recently? The only bug I can think of is 1439875 and related.

Comment 5

[Henrik Skupin [:whimboo][⌚️UTC+2]]

(In reply to Neil Deakin from comment #4)
> Did this just start failing recently? The only bug I can think of is 1439875
> and related.

The first time this failure occurred was 13 days ago. Then we had only 3 failures in a week because it started Friday. But then last week it increased to 22 failures. So yes, it's around the time when the patches of bug 1439875 landed. Thank you for the hint!

Emilio, can you please have a look at comment 3? Could this be a side-effect from your patch?

Comment 6

[Emilio Cobos Álvarez (:emilio)]

Perhaps, yeah... But it's not clear to me how to go about figuring this out. What do I need to do to run the test?

My idea is to try to repro under rr chaos mode and figure out what's going on, but lacking that I don't have a good sense of how this may be happening. Chances are that the test / the code it's testing is somewhat racy, this test is testing navigation stuff, and I suspect that across navigations getting the DOM window corresponding to a docshell can fail.

Comment 7

[Intermittent Failures Robot]

2 failures in 718 pushes (0.003 failures/push) were associated with this bug in the last 7 days.    

Repository breakdown:
* mozilla-inbound: 1
* autoland: 1

Platform breakdown:
* windows10-64: 2

For more details, see:
https://treeherder.mozilla.org/intermittent-failures.html#/bugdetails?bug=1446293&startday=2018-03-26&endday=2018-04-01&tree=trunk

Comment 8

[Henrik Skupin [:whimboo][⌚️UTC+2]]

Emilio, as usual you can use `mach` to run the tests. It means that the following basic command will run the test for you:

> mach test testing/marionette/harness/marionette_harness/tests/unit/test_navigation.py

If you want to see all the log output add `-vv` for trace logs, and `--gecko-log -` to print it all to the command line. If you miss the latter a file called `gecko.log` will be created. You don't necessarily have to build Firefox, but you can use `--binary %path%` to run a specific binary. Please note that this failure so far only occurs for opt/pgo builds.

Comment 9

[Firefox Bug Husbandry Bot]

https://wiki.mozilla.org/Bug_Triage#Intermittent_Test_Failure_Cleanup

Comment 10

[BMO Automation]

Moving bug to Testing::Marionette Client and Harness component per bug 1815831.