1861827 - Perma OSX 11 Shippable TEST-UNEXPECTED-FAIL | browser/base/content/test/fullscreen/browser_fullscreen_window_open.js | Test timed out - when Gecko 121 merges to Beta on 2023-11-20

Status: RESOLVED FIXED

(Core :: DOM: Core & HTML, defect)

Description

[Iulian Moraru]

[Tracking Requested - why for this release]:

• Central-as-beta simulation.
• How to run these simulations.

Failure log:

[task 2023-10-28T13:26:22.064Z] 13:26:22     INFO - TEST-PASS | browser/base/content/test/fullscreen/browser_fullscreen_window_open.js | Should leave full-screen - 
[task 2023-10-28T13:26:22.064Z] 13:26:22     INFO - Leaving test bound test_parentWindowClickLinkOpenExistWindow
[task 2023-10-28T13:26:22.064Z] 13:26:22     INFO - Entering test bound test_iframeWindowClickLinkOpenExistWindow
[task 2023-10-28T13:26:22.064Z] 13:26:22     INFO - Console message: [JavaScript Error: "The Components object is deprecated. It will soon be removed." {file: "https://example.com/tests/SimpleTest/EventUtils.js" line: 22}]
[task 2023-10-28T13:26:22.064Z] 13:26:22     INFO - Buffered messages logged at 13:25:43
[task 2023-10-28T13:26:22.064Z] 13:26:22     INFO - Console message: [JavaScript Error: "The Components object is deprecated. It will soon be removed." {file: "https://example.com/tests/SimpleTest/EventUtils.js" line: 22}]
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - re-focusing main window
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - Entering full-screen
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - open existing popup window
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - Clicking link with same target name should reuse the existing window
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - Console message: [JavaScript Error: "The Components object is deprecated. It will soon be removed." {file: "https://example.org/tests/SimpleTest/EventUtils.js" line: 22}]
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - Buffered messages finished
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/fullscreen/browser_fullscreen_window_open.js | Test timed out - 
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - GECKO(3211) | MEMORY STAT | vsize 402441MB | residentFast 400MB | heapAllocated 141MB
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - TEST-OK | browser/base/content/test/fullscreen/browser_fullscreen_window_open.js | took 45004ms
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - Not taking screenshot here: see the one that was previously logged
[task 2023-10-28T13:26:22.065Z] 13:26:22     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/fullscreen/browser_fullscreen_window_open.js | Found a tab after previous test timed out: https://example.com/browser/browser/base/content/test/fullscreen/open_and_focus_helper.html - 
[task 2023-10-28T13:26:22.066Z] 13:26:22     INFO - GECKO(3211) | console.error: "not able to get browsingContext for chrome window."
[task 2023-10-28T13:26:22.066Z] 13:26:22     INFO - checking window state
[task 2023-10-28T13:26:22.066Z] 13:26:22     INFO - Not taking screenshot here: see the one that was previously logged
[task 2023-10-28T13:26:22.066Z] 13:26:22     INFO - TEST-UNEXPECTED-FAIL | browser/base/content/test/fullscreen/browser_fullscreen_window_open.js | Found a browser window after previous test timed out - 
[task 2023-10-28T13:26:22.066Z] 13:26:22     INFO - Console message: [JavaScript Warning: "Exited fullscreen because fullscreen element was removed from document." {file: "chrome://browser/content/tabbrowser.js" line: 4313}]
[task 2023-10-28T13:26:23.086Z] 13:26:23     INFO - GECKO(3211) | Completed ShutdownLeaks collections in process 3288
[task 2023-10-28T13:26:23.089Z] 13:26:23     INFO - GECKO(3211) | Completed ShutdownLeaks collections in process 3216
[task 2023-10-28T13:26:23.090Z] 13:26:23     INFO - GECKO(3211) | Completed ShutdownLeaks collections in process 3212
[task 2023-10-28T13:26:23.122Z] 13:26:23     INFO - GECKO(3211) | Completed ShutdownLeaks collections in process 3314
[task 2023-10-28T13:26:23.123Z] 13:26:23     INFO - GECKO(3211) | Completed ShutdownLeaks collections in process 3214
[task 2023-10-28T13:26:23.272Z] 13:26:23     INFO - GECKO(3211) | Completed ShutdownLeaks collections in process 3211
[task 2023-10-28T13:26:23.272Z] 13:26:23     INFO - TEST-START | Shutdown
[task 2023-10-28T13:26:23.272Z] 13:26:23     INFO - Browser Chrome Test Summary
[task 2023-10-28T13:26:23.273Z] 13:26:23     INFO - Passed:  118
[task 2023-10-28T13:26:23.273Z] 13:26:23     INFO - Failed:  3
[task 2023-10-28T13:26:23.273Z] 13:26:23     INFO - Todo:    0
[task 2023-10-28T13:26:23.273Z] 13:26:23     INFO - Mode:    e10s
[task 2023-10-28T13:26:23.274Z] 13:26:23     INFO - *** End BrowserChrome Test Results ***

Comment 1

[Iulian Moraru]

Hi Dave! I have tried to figure out what could have caused this, but I didn't manage to do so. Can you please take a look at this? Maybe you can spot something that I've missed.
This seems to have started with today's central as early beta simulation, in which these two merges from central were tested(I mean the new code from central that was not tested in yesterday's beta-sim)
Thank you!

Comment 2

[BugBot [:suhaib / :marco/ :calixte]]

The Bugbug bot thinks this bug should belong to the 'Core::DOM: Core & HTML' component, and is moving the bug to that component. Please correct in case you think the bot is wrong.

Comment 3

[Cosmin Sabou [:CosminS]]

Redirecting this to Sean as per Bugbug.
Pushlog between last good - first bad revision: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=4024630f5ffa1704a32f70c816f1003b0af9f105&tochange=0be08aa0812f81d5eb9f2235165d8478ebaf825b

Comment 4

[Sebastian Hengst [:aryx] (needinfo me if it's about an intermittent or backout)]

Might be a regression from bug 1851992 which added code controlled by prefs to only run in Nightly and run other code for beta and release.

Comment 5

[Sean Feng [:sefeng]]

I also can't figure out anything obvious from the pushlog. Not sure about bug 1851992 because that one should only change necko related stuff.

Redirecting to edgar since edgar knows fullscreen.

Comment 6

[Edgar Chen [:edgar]]

(In reply to Cosmin Sabou [:CosminS] from comment #3)

Redirecting this to Sean as per Bugbug.
Pushlog between last good - first bad revision: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=4024630f5ffa1704a32f70c816f1003b0af9f105&tochange=0be08aa0812f81d5eb9f2235165d8478ebaf825b

I cannot find some clue from pushlog, didn't see anything obvious.
And I tried to find which commit might cause this, but I still see the test fails on the last good revision on my central-as-beta simulation try push, https://treeherder.mozilla.org/jobs?repo=try&revision=de84f0eacb4d7753fd906ad34cc4f06d3eecd608&selectedTaskRun=VpT9wUujQ7iFDERINmDNcA.0.
So I wonder if this push log accurate. Can you share the last try that pass the test and also the first try link that you see this failure from your side, so I could try to find some clue from them? Thanks!

Comment 7

[Edgar Chen [:edgar]]

Hmm, I tried to narrow down what might cause the existing intermittent timeout becomes worse, but have no luck. Probably just because of timing change...

Comment 8

[Iulian Moraru]

(In reply to Edgar Chen [:edgar] from comment #6)

(In reply to Cosmin Sabou [:CosminS] from comment #3)

Redirecting this to Sean as per Bugbug.
Pushlog between last good - first bad revision: https://hg.mozilla.org/mozilla-central/pushloghtml?fromchange=4024630f5ffa1704a32f70c816f1003b0af9f105&tochange=0be08aa0812f81d5eb9f2235165d8478ebaf825b

I cannot find some clue from pushlog, didn't see anything obvious.
And I tried to find which commit might cause this, but I still see the test fails on the last good revision on my central-as-beta simulation try push, https://treeherder.mozilla.org/jobs?repo=try&revision=de84f0eacb4d7753fd906ad34cc4f06d3eecd608&selectedTaskRun=VpT9wUujQ7iFDERINmDNcA.0.
So I wonder if this push log accurate. Can you share the last try that pass the test and also the first try link that you see this failure from your side, so I could try to find some clue from them? Thanks!

This is a link where I found this.
Here is the beta simulation run from the day before I found this. I've retriggered the job to see if it was perma failing on that run, and as you can see, it was not. The failed jobs from the link above, failed because of a high frequency failure that was happening that day, and if you check the logs for the failed jobs, you can see that browser_fullscreen_window_open.js was passing.
I've also checked the link you provided for your run, it is the same situation as a described above, the job failed but with another failure. Here is a link to a failure log from your run, the test is passing.

Comment 9

[Edgar Chen [:edgar]]

Thanks for the links! I tried the same revision, I can see test pass, but only once.
So it seems that the test is in a very high frequency failure.
It hard to debug this as I can not reproduce the failure locally.
So lets disable it on mac opt build first.

Comment 10

[Edgar Chen [:edgar]]

Attachment: Bug 1861827 - Disable browser_fullscreen_window_open.js test on mac opt;

Comment 11

[Iulian Moraru]

The patch seems to be working in today's central as early beta simulation.

Comment 12

[Pulsebot]

Pushed by [e-mail]:
https://hg.mozilla.org/integration/autoland/rev/40eb04d50217
Disable browser_fullscreen_window_open.js test on mac opt; r=CosminS DONTBUILD

Comment 13

[Sandor Molnar[:smolnar]]

https://hg.mozilla.org/mozilla-central/rev/40eb04d50217

Comment 14

[Norisz Fay [:noriszfay]]

Verified fixed in latest central as beta sim.

Edit: Reopening it as the issue itself is not fixed, only the test being disabled for now.

Comment 15

[Cosmin Sabou [:CosminS]]

Edgar, okay to close this one and track the disabled test in the single tracking bug of it - Bug 1795491? This being disabled it's not gonna show up anymore in beta-sims so we would just drag along this bug with us.

Comment 16

[Edgar Chen [:edgar]]

There is a bug 1795491 tracking this, so I am fine with closing this.