134361 - crash when save source view in JSD

Status: VERIFIED DUPLICATE of bug 134070

(Core :: XPCOM, defect)

Description

[Andrew Schultz]

with Linux build 20020328, I crash when saving source view with attachment
67608 [details].  It seems to be dependent on the Javascript being active while saving, as
I don't crash on Bugzilla pages which have Javascript that is inactive during
the save.  The source view is properly saved to disk before it crashes.  Saving
profile data has the same effect.

talkback ID TB4627039K and TB4626590W
generated with http://www.globalserve.net/~brent/tech/beam/

regression occurred between 2002032006 and 2002032108

Comment 1

[Andrew Schultz]

Attachment: stacktrace

Comment 2

[Andrew Schultz]

based on the stacktrace, this looks like a dupe of topcrash bug 119693.

cc'ing pavlov

Comment 3

[Robert Ginda]

*** Bug 134377 has been marked as a duplicate of this bug. ***

Comment 4

[Andrew Schultz]

current CVS now usually throws an assertion before crashing:

Assertion failure: 0 == rv, at ptsynch.c:191

also, the top of the stacktrace is slightly different:

#0  0x405a0ae1 in __kill () at __kill:-1
#1  0x402a676b in raise (sig=6) at signals.c:65
#2  0x405a2062 in abort () at ../sysdeps/generic/abort.c:88
#3  0x40262112 in PR_Assert () at prlog.c:510
#4  0x40273cd1 in PR_Lock (lock=0x89e773c) at ptsynch.c:191
#5  0x40274794 in PR_EnterMonitor (mon=0x89e7738) at ptsynch.c:514
#6  0x401bc944 in PL_ProcessPendingEvents (self=0x8996528) at plevent.c:530

over to NSPR (not XPCOM)... 

Comment 5

[Andrew Schultz]

if the real problem is farther up the stack, then this isn't NSPR... feel free
to kick it back to JSD or over to XPCOM if you think NSPR is doing its job.

Comment 6

[Andrew Schultz]

back to XPCOM

if I back out this change:
http://bonsai.mozilla.org/cvsquery.cgi?treeid=default&module=SeaMonkeyAll&branch=HEAD&branchtype=match&dir=mozilla%2Fxpcom&file=nsEventQueueService.cpp&filetype=match&who=&whotype=match&sortby=Date&hours=2&date=explicit&mindate=03%2F20%2F2002+05%3A00%3A00&maxdate=03%2F21%2F2002+09%3A00%3A00&cvsroot=%2Fcvsroot

everything is happy.  bonsai lists bug 128508 and [e-mail] for that
checkin, but bug 128508 seems unrelated...

Comment 7

[Andrew Schultz]

oops.  and reassign.  also marking NEW.

Comment 8

[Wan-Teh Chang]

Can you print the value of rv in the debugger?

This bug may be a duplicate of bug 134070 that Brian
Ryner reported.

Comment 9

[Doug Turner (:dougt)]

bryner and I believe that this indeed is the dup wtc mentioned.

*** This bug has been marked as a duplicate of 134070 ***

Comment 10

[Andrew Schultz]

the patch in bug 134070 does fix the crash.
any comment on comment 6?  it seems to be the change that caused the regression
here, and maybe in bug 134070 also.

Comment 11

[Andrew Schultz]

this is working now (with patch from 134070)