473281 - [win32] ###!!! ASSERTION: Failed to create the object: 'Error' in dom/src/base/nsDOMClassInfo.cpp:5011

Status: RESOLVED WORKSFORME

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

Description

[Nick Thomas [:nthomas] (UTC+13)]

Attachment: Partial build log with assertion and stack trace.

Over the last 5 days we've hit 3 of these in mozilla-central leaktest builds, and 4 in mozilla-1.9.1 leaktest builds. It's older than that though, I reported something similar in bug 471950.

This is occurring on 5 separate slaves [1] so either there's something screwy in the build system that's leading them all into a confused state or a genuine bug here. Can't find this on linux or mac over the same period.

[1] moz2-win32-slave15,07,05 for m-c; 15,06,09 for m-1.9.1

Comment 1

[Peter Van der Beken [:peterv]]

There's also

WARNING: NS_ENSURE_TRUE(doc) failed: file e:/builds/moz2_slave/mozilla-central-win32-debug/build/content/html/content/src/nsHTMLImageElement.cpp, line 168

which means nsContentUtils::GetDocumentFromCaller() failed to return a document.

Comment 2

[(not currently active) Ted Mielczarek]

You can run the same leaktest those machines do by running:
python $OBJDIR/_leaktest/leaktest.py

Comment 3

[Nick Thomas [:nthomas] (UTC+13)]

There were 19 instances of this assertion over the last 5 days, spread over these commands
  7 counts: python leaktest.py -- -CreateProfile default
  7 counts: python leaktest.py -- -P default
  7 counts: python leaktest.py -l bloat.log
and the mozilla-central and mozilla-1.9.1 trees. Each one of those is a false red on the tree, so could we get someone's time to figure out what is going here ?

Comment 4

[Blake Kaplan (:mrbkap) (inactive, use needinfo)]

Do we have a time window when this started happening?

Comment 5

[Nick Thomas [:nthomas] (UTC+13)]

Have asked IT for data from the tinderbox server to attempt to figure that out. Logs there go back 61 days.

Comment 6

[(not currently active) Ted Mielczarek]

Just happened on Linux:
http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1232638440.1232639897.26228.gz

Comment 7

[Nick Thomas [:nthomas] (UTC+13)]

The assertion goes back to at least November 22, the oldest logs tinderbox has for the Firefox tree. I've asked for the same check on Firefox3.1 in the hope the cause landed there later.

Comment 8

[Nick Thomas [:nthomas] (UTC+13)]

(In reply to comment #7)
> The assertion goes back to at least November 22, the oldest logs tinderbox has
> for the Firefox tree. I've asked for the same check on Firefox3.1 in the hope
> the cause landed there later.

Same for Firefox3.1.

Comment 9

[Johnny Stenback (:jst)]

While this is obviously really bad over all, we're not going to hold the release for it, as it's not something that we've seen in the wild as of yet, as far as we know.

Comment 10

[(not currently active) Ted Mielczarek]

It does make the leak test boxes go red on occasion though, so it's quite annoying.

Comment 11

[David Baron :dbaron: 🏴󠁵󠁳󠁣󠁡󠁿 ⌚UTC-8]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1238780273.1238780889.30946.gz&fulltext=1
(mozilla-central win32)

Comment 12

[Robert O'Callahan (:roc) (email my personal email if necessary)]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1241634822.1241635249.31925.gz
(mozilla-central Linux leak test)

Comment 13

[Jeff Muizelaar [:jrmuizel]]

http://tinderbox.mozilla.org/showlog.cgi?log=Firefox/1274806256.1274807945.6055.gz