Closed Bug 1563697 Opened 5 years ago Closed 5 years ago

Intermittent dom/media/tests/crashtests/1429507_1.html | assertion count 1 is more than expected 0 assertions

Categories

(Core :: WebRTC, defect, P5)

Product:
Core
Component:
WebRTC
Platform:
ARM
Android
Type:
defect

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla70
Tracking Flags:
Tracking Status
firefox-esr60 --- unaffected
firefox-esr68 --- unaffected
firefox68 --- unaffected
firefox69 --- unaffected
firefox70 --- fixed

People

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

References

(Regression)

Details

(Keywords: intermittent-failure, regression, Whiteboard: [stockwell unknown])

Attachments

(2 files)

Bug 1563697 - Move script tags from <head> to <body> to avoid window.open asserts. r?jib
(deleted), text/x-phabricator-request
Details
Bug 1563697 - Allow one intermittent assert on Android for bug 1429507 crashtests. r?jib
(deleted), text/x-phabricator-request
Details

Filed by: aciure [at] mozilla.com
Parsed log: https://treeherder.mozilla.org/logviewer.html#?job_id=254911368&repo=mozilla-central
Full log: https://queue.taskcluster.net/v1/task/P8Za7N7TQgCWjj_O_WqddQ/runs/0/artifacts/public/logs/live_backing.log
Reftest URL: https://hg.mozilla.org/mozilla-central/raw-file/tip/layout/tools/reftest/reftest-analyzer.xhtml#logurl=https://queue.taskcluster.net/v1/task/P8Za7N7TQgCWjj_O_WqddQ/runs/0/artifacts/public/logs/live_backing.log&only_show_unexpected=1

09:23:26 INFO - REFTEST TEST-PASS | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_1.html | (LOAD ONLY)
09:23:26 INFO - REFTEST TEST-END | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_1.html
09:23:26 INFO - REFTEST TEST-UNEXPECTED-FAIL | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_1.html | assertion count 1 is more than expected 0 assertions
09:23:26 INFO - REFTEST TEST-START | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_2.html
09:23:26 INFO - REFTEST INFO | RESTORE PREFERENCE pref(browser.link.open_newwindow,3)
09:23:26 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.getusermedia.insecure.enabled,false)
09:23:26 INFO - REFTEST INFO | RESTORE PREFERENCE pref(dom.disable_open_during_load,true)
09:23:26 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.navigator.permission.disabled,false)
09:23:26 INFO - REFTEST INFO | SET PREFERENCE pref(media.navigator.permission.disabled,true)
09:23:26 INFO - REFTEST INFO | SET PREFERENCE pref(dom.disable_open_during_load,false)
09:23:26 INFO - REFTEST INFO | SET PREFERENCE pref(media.getusermedia.insecure.enabled,true)
09:23:26 INFO - REFTEST INFO | SET PREFERENCE pref(browser.link.open_newwindow,2)
09:23:26 INFO - REFTEST TEST-LOAD | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_2.html | 682 / 3775 (18%)
09:23:26 INFO - REFTEST TEST-PASS | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_2.html | (LOAD ONLY)
09:23:26 INFO - REFTEST TEST-END | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_2.html
09:23:26 INFO - REFTEST TEST-UNEXPECTED-FAIL | http://10.7.205.216:8854/tests/dom/media/tests/crashtests/1429507_2.html | assertion count 1 is more than expected 0 assertions

OS: Unspecified → Android
Regressed by: 1525314
Hardware: Unspecified → ARM

Any ideas?

Flags: needinfo?(apehrson)
Type: -- → defect

I'd guess since it's android only that setting mWindowCreator can happen after we try to open the new window. Since it's intermittent it must be racy.

I could try to move the script into <body> instead of <head> to see if that helps, but I'd have to do some testing to see that the crashtests still exercise the original crash.

Assignee: nobody → apehrson
Flags: needinfo?(apehrson)

In the last 7 days, there have been 30 failures on android-em-7-0-x86_64, android-hw-p2-8-0-arm7-api-16-qr platforms, debug build type.

An example of a recent log file:
https://treeherder.mozilla.org/logviewer.html#/jobs?job_id=257679646&repo=autoland&lineNumber=4447

And the relevant part of the log:
task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST TEST-PASS | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html | (LOAD ONLY)
[task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST TEST-END | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html
[task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST TEST-UNEXPECTED-FAIL | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html | assertion count 1 is more than expected 0 assertions
[task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST TEST-START | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_2.html
[task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST INFO | RESTORE PREFERENCE pref(browser.link.open_newwindow,3)
[task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.getusermedia.insecure.enabled,false)
[task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.devices.insecure.enabled,false)
[task 2019-07-22T12:56:43.541Z] 12:56:43 INFO - REFTEST INFO | RESTORE PREFERENCE pref(dom.disable_open_during_load,true)
[task 2019-07-22T12:56:43.542Z] 12:56:43 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.navigator.permission.disabled,false)
[task 2019-07-22T12:56:43.542Z] 12:56:43 INFO - REFTEST INFO | SET PREFERENCE pref(media.navigator.permission.disabled,true)
[task 2019-07-22T12:56:43.542Z] 12:56:43 INFO - REFTEST INFO | SET PREFERENCE pref(dom.disable_open_during_load,false)
[task 2019-07-22T12:56:43.542Z] 12:56:43 INFO - REFTEST INFO | SET PREFERENCE pref(media.devices.insecure.enabled,true)
[task 2019-07-22T12:56:43.542Z] 12:56:43 INFO - REFTEST INFO | SET PREFERENCE pref(media.getusermedia.insecure.enabled,true)
[task 2019-07-22T12:56:43.542Z] 12:56:43 INFO - REFTEST INFO | SET PREFERENCE pref(browser.link.open_newwindow,2)

Nils, as you are the triage owner of this component, could you please take a look at this?
Thank you!

Flags: needinfo?(drno)
Whiteboard: [stockwell needswork]

(In reply to Andreas Pehrson [:pehrsons] from comment #4)

I'd guess since it's android only that setting mWindowCreator can happen after we try to open the new window. Since it's intermittent it must be racy.

I could try to move the script into <body> instead of <head> to see if that helps, but I'd have to do some testing to see that the crashtests still exercise the original crash.

I wasn't able to build the original build myself, so I downloaded it from the archives and was able to repro 1429507_1.html with and without the proposed patch. However, 1429507_2.html wouldn't repro with the patch, nor without. Could be a timing difference in loading the file normally as opposed to through the crashtest harness.

I'll push the patch up, hoping it'll help.

Status: NEW → ASSIGNED
Flags: needinfo?(drno)
Pushed by pehrsons@gmail.com: https://hg.mozilla.org/integration/autoland/rev/828428225701 Move script tags from <head> to <body> to avoid window.open asserts. r=jib

https://hg.mozilla.org/mozilla-central/rev/828428225701

Status: ASSIGNED → RESOLVED
Closed: 5 years ago
status-firefox70: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla70

Recent failure https://treeherder.mozilla.org/#/jobs?repo=mozilla-inbound&searchStr=android%2C7.0%2Cx86-64%2Cdebug%2Creftests%2Ctest-android-em-7.0-x86_64%2Fdebug-geckoview-crashtest-e10s%2Cr%28c%29&revision=047e16b38566f319ee1a9d566dc9b7e7c82bdd7d

Log failure https://treeherder.mozilla.org/logviewer.html#?job_id=260731250&repo=mozilla-inbound

Part of that log:

[task 2019-08-09T04:25:45.512Z] 04:25:45 INFO - REFTEST TEST-START | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html
[task 2019-08-09T04:25:45.512Z] 04:25:45 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.getusermedia.insecure.enabled,false)
[task 2019-08-09T04:25:45.512Z] 04:25:45 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.devices.insecure.enabled,false)
[task 2019-08-09T04:25:45.513Z] 04:25:45 INFO - REFTEST INFO | RESTORE PREFERENCE pref(dom.disable_open_during_load,true)
[task 2019-08-09T04:25:45.513Z] 04:25:45 INFO - REFTEST INFO | RESTORE PREFERENCE pref(media.navigator.permission.disabled,false)
[task 2019-08-09T04:25:45.513Z] 04:25:45 INFO - REFTEST INFO | SET PREFERENCE pref(media.navigator.permission.disabled,true)
[task 2019-08-09T04:25:45.514Z] 04:25:45 INFO - REFTEST INFO | SET PREFERENCE pref(dom.disable_open_during_load,false)
[task 2019-08-09T04:25:45.514Z] 04:25:45 INFO - REFTEST INFO | SET PREFERENCE pref(media.devices.insecure.enabled,true)
[task 2019-08-09T04:25:45.514Z] 04:25:45 INFO - REFTEST INFO | SET PREFERENCE pref(media.getusermedia.insecure.enabled,true)
[task 2019-08-09T04:25:45.514Z] 04:25:45 INFO - REFTEST INFO | SET PREFERENCE pref(browser.link.open_newwindow,2)
[task 2019-08-09T04:25:45.515Z] 04:25:45 INFO - REFTEST TEST-LOAD | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html | 685 / 3786 (18%)
[task 2019-08-09T04:25:45.515Z] 04:25:45 INFO - REFTEST TEST-PASS | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html | (LOAD ONLY)
[task 2019-08-09T04:25:45.516Z] 04:25:45 INFO - REFTEST TEST-END | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html
[task 2019-08-09T04:25:45.516Z] 04:25:45 INFO - REFTEST TEST-UNEXPECTED-FAIL | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_1.html | assertion count 1 is more than expected 0 assertions
[task 2019-08-09T04:25:45.517Z] 04:25:45 INFO - REFTEST TEST-START | http://10.0.2.2:8854/tests/dom/media/tests/crashtests/1429507_2.html

Status: RESOLVED → REOPENED
Resolution: FIXED → ---

No idea what's up here. I think I'll just allow an assertion to fail on Android. It's unrelated to what the crashtests are testing FWIW, and they still test the original crash they were designed for.

Pushed by pehrsons@gmail.com: https://hg.mozilla.org/integration/autoland/rev/149f140f46f7 Allow one intermittent assert on Android for bug 1429507 crashtests. r=jib

https://hg.mozilla.org/mozilla-central/rev/149f140f46f7

Status: REOPENED → RESOLVED
Closed: 5 years ago → 5 years ago
status-firefox70: affected → fixed
Resolution: --- → FIXED

(In reply to Byron Campen [:bwc] from comment #3)

Any ideas?

See https://bugzilla.mozilla.org/show_bug.cgi?id=679841#c2

I think we need to set |mWindowCreator| by calling |nsWindowWatcher::SetWindowCreator|
https://bugzilla.mozilla.org/show_bug.cgi?id=1239896#c3

Has Regression Range: --- → yes
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: