Open Bug 1703437 Opened 4 years ago Updated 2 years ago

Frequent mostly win /navigation-timing/nav2_test_redirect_server.html | Navigation Timing 2 WPT - uncaught exception: Error: assert_true: Expected startTime to be no greater than redirectStart. expected true got false

Categories

(Core :: DOM: Performance, defect, P5)

Product:
Core
Component:
DOM: Performance
Type:
defect

Tracking

()

Tracking Flags:
Tracking Status
firefox-esr78 --- unaffected
firefox87 --- unaffected
firefox88 --- unaffected
firefox89 --- affected

People

(Reporter: intermittent-bug-filer, Unassigned)

References

Details

(Keywords: intermittent-failure, leave-open, Whiteboard: [retriggered][stockwell unknown])

Attachments

(1 file)

Bug 1703437 - updated expectancies in nav2_test_redirect_server.html for frequent failures on win. r=#intermittent-reviewers
(deleted), text/x-phabricator-request
Details

Filed by: csabou [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer?job_id=335659834&repo=autoland
Full log: https://firefox-ci-tc.services.mozilla.com/api/queue/v1/task/U_byr93qTS21qmV8Nnv9hg/runs/0/artifacts/public/logs/live_backing.log

[task 2021-04-07T04:06:37.119Z] 04:06:37     INFO - TEST-OK | /navigation-timing/nav2_test_redirect_none.html | took 66ms
[task 2021-04-07T04:06:37.120Z] 04:06:37     INFO - TEST-START | /navigation-timing/nav2_test_redirect_server.html
[task 2021-04-07T04:06:37.121Z] 04:06:37     INFO - Closing window 8589934639
[task 2021-04-07T04:06:37.195Z] 04:06:37     INFO - PID 7780 | JavaScript error: http://web-platform.test:8000/resources/testharness.js, line 1211: uncaught exception: Error: assert_true: Expected startTime to be no greater than redirectStart. expected true got false
[task 2021-04-07T04:06:37.201Z] 04:06:37     INFO - 
[task 2021-04-07T04:06:37.201Z] 04:06:37     INFO - TEST-UNEXPECTED-FAIL | /navigation-timing/nav2_test_redirect_server.html | Navigation Timing 2 WPT - uncaught exception: Error: assert_true: Expected startTime to be no greater than redirectStart. expected true got false
[task 2021-04-07T04:06:37.201Z] 04:06:37     INFO - verifyTimingEventOrder@http://web-platform.test:8000/navigation-timing/nav2_test_redirect_server.html:16:32
[task 2021-04-07T04:06:37.202Z] 04:06:37     INFO - onload_test@http://web-platform.test:8000/navigation-timing/nav2_test_redirect_server.html:36:39
[task 2021-04-07T04:06:37.202Z] 04:06:37     INFO - onload@http://web-platform.test:8000/navigation-timing/nav2_test_redirect_server.html:1:1
[task 2021-04-07T04:06:37.202Z] 04:06:37     INFO - TEST-OK | /navigation-timing/nav2_test_redirect_server.html | took 82ms```

occurs on non-fission too: https://treeherder.mozilla.org/logviewer?job_id=335688287&repo=mozilla-central&lineNumber=8492

Summary: Frequent win wr fission /navigation-timing/nav2_test_redirect_server.html | Navigation Timing 2 WPT - uncaught exception: Error: assert_true: Expected startTime to be no greater than redirectStart. expected true got false → Frequent win /navigation-timing/nav2_test_redirect_server.html | Navigation Timing 2 WPT - uncaught exception: Error: assert_true: Expected startTime to be no greater than redirectStart. expected true got false

Set release status flags based on info from the regressing bug 1701668

status-firefox87: --- → unaffected
status-firefox88: --- → unaffected
status-firefox89: --- → affected
status-firefox-esr78: --- → unaffected

The change I made in bug 1701668 doesn't have anything to do with this test, so it is much more likely to be something wrong with the test, or the dom/performance code. From a brief look at what the test code does, it seems to assume that the 'start time' is less than the 'redirect time' but will fail if they are equal.

There some code for the properties of the performance objects that the test is comparing that reduces the precision of the return values based on some randomly generated numbers. I found it difficult to reproduce, but if I simply moved the nav2_test_redirect_server.html test to a different directory, this randomness caused the rounding of redirectStart to result in zero, and failed the test.

I don't know if that is the problem here, but someone more familiar with the performance timing code might want to take a look.

Flags: needinfo?(enndeakin)

Hi Kim, can you assign someone to take a look at this?

Flags: needinfo?(kmoir)
No longer regressed by: 1701668

Sean is this something in your area of expertise? If not, let me know and I can redirect.

Flags: needinfo?(kmoir) → needinfo?(sefeng)

This is not my area of expertise, I think Andrew knows necko stuff quite well?

Flags: needinfo?(sefeng) → needinfo?(acreskey)

Are there instructions for running these locally?

You can run the test locally with './mach wpt navigation-timing/nav2_test_redirect_server.html', but it doesn't fail for me locally.

Summary: Frequent win /navigation-timing/nav2_test_redirect_server.html | Navigation Timing 2 WPT - uncaught exception: Error: assert_true: Expected startTime to be no greater than redirectStart. expected true got false → Frequent mostly win /navigation-timing/nav2_test_redirect_server.html | Navigation Timing 2 WPT - uncaught exception: Error: assert_true: Expected startTime to be no greater than redirectStart. expected true got false
Assignee: nobody → smolnar
Status: NEW → ASSIGNED
Assignee: smolnar → nobody
Status: ASSIGNED → NEW
Keywords: leave-open
Whiteboard: [retriggered][stockwell needswork:owner] → [retriggered][stockwell needswork:owner][stockwell disabled]
Attachment #9217055 - Attachment description: Bug 1703437 - updated expectancies in navigation-timing/nav2_test_redirect_server.html for frequent failures on win. r=#intermittent-reviewers → Bug 1703437 - updated expectancies in nav2_test_redirect_server.html for frequent failures on win. r=#intermittent-reviewers
Pushed by smolnar@mozilla.com: https://hg.mozilla.org/integration/autoland/rev/9281a1d9cabf updated expectancies in nav2_test_redirect_server.html for frequent failures on win. r=intermittent-reviewers,jmaher
Whiteboard: [retriggered][stockwell disable-recommended] → [retriggered][stockwell needswork:owner]

Sorry, didn't get to this for a bit.

What I believe is happening is that we're reporting 0 for both startTime and redirectStart but the test will fail unless startTime is strictly less than redirectStart:
https://searchfox.org/mozilla-central/rev/98a9257ca2847fad9a19631ac76199474516b31e/testing/web-platform/tests/navigation-timing/nav2_test_redirect_server.html#16

This seems like the test is incorrect as I think timing of 0 for each entry is valid.

I found it easiest to reproduce using the the w3c-test:
http://w3c-test.org/navigation-timing/nav2_test_redirect_server.html

Not sure what the next best step is here.

Flags: needinfo?(acreskey)

If we agree on my hypothesis from Comment 21, can we push for a change to the w3c test?

Component: Performance → DOM: Performance
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: