547070 - TM: xpcshell/tests/test_places/unit/test_empty_tags.js | 2 == undefined

Status: RESOLVED DUPLICATE of bug 545980

(Core :: JavaScript Engine, defect)

Description

[Nicholas Nethercote [inactive]]

We've been getting this xpcshell failure on TM for over a week now:

TEST-UNEXPECTED-FAIL | /builds/slave/tracemonkey-linux-debug-unittest-xpcshell/build/xpcshell/tests/test_places/unit/test_empty_tags.js | test failed (with xpcshell return code: 0), see following log:
TEST-UNEXPECTED-FAIL | /builds/slave/tracemonkey-linux-debug-unittest-xpcshell/build/xpcshell/tests/test_places/unit/test_empty_tags.js | 2 == undefined - See following stack:

I did some hunting on the tinderbox, AFAICT this commit is to blame:

  changeset:   38038:a353e155804e
  user:        Brendan Eich <[e-mail]>
  date:        Wed Feb 10 15:17:52 2010 -0800
  summary:     Bug 542002 - Optimize to flat closures even if some upvars can't be copied (r=jorendorff).

This is the log of the first failing run:

  http://tinderbox.mozilla.org/showlog.cgi?log=TraceMonkey/1265848478.1265850261.18528.gz

Comment 1

[Nicholas Nethercote [inactive]]

here's tinderbox for the relevant 24 hour period:

http://tinderbox.mozilla.org/showbuilds.cgi?tree=TraceMonkey&maxdate=1265882194&hours=24&legend=0&norules=1

OS X and Window columns corroborate that r38038 is the cause.

r38041 was a follow-up to "Fix orange in wake of patch for bug 542002".  Either it didn't work or it fixed a different orange.

Comment 2

[Brendan Eich [:brendan]]

The rNNNNN links in comment 2 don't work for me.

This should be fixed now, the earlier fix was for different orange. Sorry for the delay in getting the second fix in.

/be

Comment 3

[Nicholas Nethercote [inactive]]

(In reply to comment #2)
> The rNNNNN links in comment 2 don't work for me.

I didn't realise rNNNNN strings became links.  Looks like they link to m-c.

> *** This bug has been marked as a duplicate of bug 545980 ***

I searched for "2 == undefined" before filing the bug.  Is it standard practice to put the failing assertion in the title of a bug?  Seems like it makes it easier to find and thus avoid duplicates.

Comment 4

[Brendan Eich [:brendan]]

(In reply to comment #3)
> (In reply to comment #2)
> I searched for "2 == undefined" before filing the bug.  Is it standard practice
> to put the failing assertion in the title of a bug?  Seems like it makes it
> easier to find and thus avoid duplicates.

It does -- I shoulda done that. Thanks for the reminder.

/be