Closed Bug 453969 Opened 16 years ago Closed 9 years ago

Intermittent test_bug382113.html | Child got load event - got false, expected true

Categories

(Core :: DOM: Events, defect)

Product:
Core
Component:
DOM: Events
Platform:
x86
All
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED FIXED
Milestone:
mozilla39
Tracking Flags:
Tracking Status
firefox37 --- fixed
firefox38 --- fixed
firefox39 --- fixed
firefox-esr31 --- fixed
b2g-v1.4 --- unaffected
b2g-v2.0 --- fixed
b2g-v2.1 --- fixed
b2g-v2.1S --- fixed
b2g-v2.2 --- fixed
b2g-master --- fixed

People

(Reporter: sgautherie, Assigned: bzbarsky)

References

(
URL
)

Details

(Keywords: intermittent-failure)

Attachments

(1 file)

Fix the race in test_bug382113.html so we don't set our child-onload-fired boolean to false _after_ the child onload has already fired
2.55 KB, patch
froydnj
: review+
Details | Diff | Splinter Review 
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1220691024.1220695552.1067.gz#err0
(SeaMonkey) Win2k3 comm-central dep unit test on 2008/09/06 01:50:24
[
*** 3136 INFO Running /tests/content/base/test/test_bug382113.html...
NEXT ERROR *** 3137 ERROR TEST-UNEXPECTED-FAIL | 
/tests/content/base/test/test_bug382113.html | Child got load event - got false, expected true
*** 3138 ERROR TEST-UNEXPECTED-FAIL | /tests/content/base/test/test_bug382113.html | Object got load event - got false, expected true
*** 3140 INFO Running /tests/content/base/test/test_bug383430.html...
]

NB: 1st time I see this, so filing as Unconfirmed for now.
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1223155511.1223159473.29335.gz
Win2k3 comm-central dep unit test on 2008/10/04 14:25:11
Blocks: 438871
Status: UNCONFIRMED → NEW
Ever confirmed: true
Summary: "ERROR TEST-UNEXPECTED-FAIL | .../test_bug382113.html | Child got load event - got false, expected true" → [SeaMonkey, Windows] "ERROR TEST-UNEXPECTED-FAIL | .../test_bug382113.html | Child got load event - got false, expected true"
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1225632626.1225637359.17878.gz
Win2k3 comm-central dep unit test on 2008/11/02 05:30:26
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1227386398.1227390660.27581.gz
Win2k3 comm-central dep unit test on 2008/11/22 12:39:58
Firefox 3.1 too:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox3.1/1230398438.1230402573.21602.gz
WINNT 5.2 mozilla-1.9.1 unit test on 2008/12/27 09:20:38
Flags: wanted1.9.1?
Summary: [SeaMonkey, Windows] "ERROR TEST-UNEXPECTED-FAIL | .../test_bug382113.html | Child got load event - got false, expected true" → [Windows] "ERROR TEST-UNEXPECTED-FAIL | .../test_bug382113.html | Child got load event - got false, expected true"
Whiteboard: [orange]
http://tinderbox.mozilla.org/showlog.cgi?log=SeaMonkey/1238877579.1238882604.26256.gz
WINNT 5.2 comm-central unit test on 2009/04/04 13:39:39
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1255718450.1255720922.13556.gz

Linux mozilla-central test mochitests  [testfailed] Started 11:40, finished 12:23
OS: Windows Server 2003 → All
Summary: [Windows] "ERROR TEST-UNEXPECTED-FAIL | .../test_bug382113.html | Child got load event - got false, expected true" → mochitest-plain: intermittent "test_bug382113.html | Child got load event - got false, expected true"
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1256655436.1256655972.1395.gz
WINNT 5.2 mozilla-central test opt mochitests-1/5 on 2009/10/27 07:57:16
WINNT 5.2 mozilla-central opt test mochitests-1/5 [testfailed] Started 12:33, finished 12:41
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox-Unittest/1259872413.1259872841.29053.gz&fulltext=1
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1263039407.1263039954.15922.gz

Linux mozilla-central opt test mochitests-1/5 on 2010/01/09 04:16:47
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1267370165.1267370942.17069.gz
Linux mozilla-central opt test mochitests-1/5 on 2010/02/28 07:16:05
s: moz2-linux-slave26
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1270136228.1270136687.17688.gz
Linux mozilla-central opt test mochitests-1/5 on 2010/04/01 08:37:08
s: moz2-linux-slave28
Flags: wanted1.9.1?
Whiteboard: [orange]
Summary: mochitest-plain: intermittent "test_bug382113.html | Child got load event - got false, expected true" → Intermittent test_bug382113.html | Child got load event - got false, expected true
Closing inactive keywords:intermittent-failure bugs where the TBPLbot has previously commented and the test isn't marked as disabled; filter on orange-cleanup-201401.
Status: NEW → RESOLVED
Closed: 10 years ago
Resolution: --- → WORKSFORME
Status: RESOLVED → REOPENED
Resolution: WORKSFORME → ---
(After 6 years,) I suggest to disable this test, until someone can try and diagnose/fix it.
Status: REOPENED → NEW
Flags: needinfo?(bzbarsky)
Uh...  This should so not be failing.  All this test is testing is that we get onload in the child frame before we get it in the parent!

Why did it start failing so much in the last week?
Flags: needinfo?(bzbarsky) → needinfo?(bugzillamozillaorg_serge_20140323)
(In reply to Boris Zbarsky [:bz] from comment #213)
> Why did it start failing so much in the last week?

This is indeed the currently interesting clue/question.
(Though the answer might be the same as why it was failing 6 years ago...)

I don't know: that would be the diagnose (todo) part.
Application regression? (badly) Failing/Unclean previous test? Test harness improvement (at detecting)? ...
Flags: needinfo?(bugzillamozillaorg_serge_20140323)
Presumably the first of those.... can we get something like a regression range?
URL: http://mxr.mozilla.org/mozilla-centra...
Keywords: qawanted
Whiteboard: [qawanted: do comment 215]
clearing outdated keywords - please re-add / update if needed
Keywords: qawanted
(In reply to Joshua Mitchell [:Joshua_M] from comment #329)
> clearing outdated keywords

My 2014-08-26 request was not outdated, I think.
Keywords: qawanted
(In reply to Serge Gautherie (:sgautherie) from comment #352)
> My 2014-08-26 request was not outdated, I think.

Not that it didn't any good anyway.

I'm going to attempt to bisect this on Try, but it'll be a PITA given that OSX 10.8 is chronically backlogged and Win8 x32 doesn't even run in production anymore.
I just ran a set of 60 OSX 10.8 debug mochitest-1 runs reverted to revision dac8b4a0bd7c, which per comment 204 should have been at the front end of the spike. The result? No failures.

I'm afraid that at this point we have little to no chance of accurately pinpointing the cause of the spike via Try bisection. Boris, I'm afraid that given the frequency of this failure and the lack of activity for the past 2 months, I'm going to have little choice but to just disable the test at this point.
Flags: needinfo?(bzbarsky)
So I don't get this.  How is it possible that this fails so often on tbpl in general but not on try?

Is this failing on particular types of slaves?  Something else?

This test failing is REALLY REALLY BAD.  Like web compat fail, breaks sites bad.  The fact that we're not getting reports of that in the real world and can't reproduce this failure on try makes me very suspicious....
Flags: needinfo?(bzbarsky) → needinfo?(ryanvm)
Comment 377 references reverting to a changeset from August when the spike started, not the current situation in the ~200 reports in this bug since. My point is that if you're waiting on QA to pinpoint what caused this to spike, this bug is going to continue going nowhere as it has for months at this point.
Flags: needinfo?(ryanvm)
Oh, I guess this is only about <object>, not <iframe>.  That would reduce the web compat impact significantly...

John, are you able/willing to take a look here?  Did something change in object loading that would cause this to happen?  Something happening off an event that doesn't block onload in the parent document?
Flags: needinfo?(jschoenick)
Disabled on OSX and Windows debug.
https://hg.mozilla.org/integration/mozilla-inbound/rev/b1f8b6f4541b
Whiteboard: [qawanted: do comment 215] → [test disabled on OSX and Windows debug][leave open]
Flags: needinfo?(jst)
I think it's time to admit defeat and disable the test outright.
Flags: needinfo?(bzbarsky)
This is ridiculous.  I'm going to see if I can figure out what's up here.
I'm going to clear qawanted since bz is looking into it.
Keywords: qawanted
Continuing the ridiculousness, I spent a while doing logging and retriggering and stuff to see why the onload event is not firing on the child, but it clearly _is_ firing on the child based on the logs.  So then I actually _read_ the test.  I am a moron.  Patch coming up.
Assignee: nobody → bzbarsky
Flags: needinfo?(jst)
Flags: needinfo?(john)
Flags: needinfo?(bzbarsky)
Whiteboard: [test disabled on OSX and Windows debug][leave open] → [test disabled on OSX and Windows debug]
Comment on attachment 8574971 [details] [diff] [review]
Fix the race in test_bug382113.html so we don't set our child-onload-fired boolean to false _after_ the child onload has already fired

Review of attachment 8574971 [details] [diff] [review]:
-----------------------------------------------------------------

bz asked me over IRC whether I could review this.
Attachment #8574971 - Flags: review?(bugs) → review+
https://hg.mozilla.org/mozilla-central/rev/88634c2aafcb
Status: NEW → RESOLVED
Closed: 10 years ago → 9 years ago
status-firefox39: --- → fixed
Resolution: --- → FIXED
Target Milestone: --- → mozilla39
Fwiw, note that treeherder comments 613-617 and 619-620 seem unrelated to this bug.
URL: http://mxr.mozilla.org/mozilla-centra... → http://mxr.mozilla.org/mozilla-centra...
Whiteboard: [test disabled on OSX and Windows debug]
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: