Shutdown race condition with presshell resize timer

RESOLVED INCOMPLETE

Status

()

Core
Layout: Misc Code
P2
critical
RESOLVED INCOMPLETE
16 years ago
9 years ago

People

(Reporter: jesup, Unassigned)

Tracking

({crash})

Trunk
Future
x86
All
crash
Points:
---

Firefox Tracking Flags

(Not tracked)

Details

(Whiteboard: [wgate])

Attachments

(1 attachment)

(Reporter)

Description

16 years ago
Note: this is with our variant of mozilla, though we haven't played with any of
the presshell resize code.

There appears to be a race condition in shutdown with the presshell resize
timer.  See attached backtrace where we crash in ConvertToLowercase() because
gCaseConv is null.  We should probably kill the timer (all timers?) early in
shutdown.

See also bug 159953 for another (different) shutdown race condition.  I don't
believe any of our modifications are relevant; they are all resolved before
ShutdownXPCOM is called.
(Reporter)

Comment 1

16 years ago
Marc doesn't work at netscape anymore...  Reassigning
Assignee: attinasi → karnaze
Keywords: crash
Whiteboard: [wgate]
(Reporter)

Comment 2

16 years ago
Created attachment 92917 [details]
GDB backtrace

Updated

16 years ago
QA Contact: petersen → moied

Updated

16 years ago
Priority: -- → P2

Updated

16 years ago
Target Milestone: --- → Future

Comment 3

15 years ago
->layout misc
Assignee: karnaze → misc
Component: Layout → Layout: Misc Code
QA Contact: moied → nobody

Comment 4

15 years ago
By the definitions on <http://bugzilla.mozilla.org/bug_status.html#severity> and
<http://bugzilla.mozilla.org/enter_bug.cgi?format=guided>, crashing and dataloss
bugs are of critical or possibly higher severity.  Only changing open bugs to
minimize unnecessary spam.  Keywords to trigger this would be crash, topcrash,
topcrash+, zt4newcrash, dataloss.
Severity: normal → critical

Comment 5

10 years ago
with no testcase, is this worth keeping?
reporter appears to be gone (pinged 12/2006)
Assignee: layout.misc-code → nobody
QA Contact: nobody → layout.misc-code

Comment 6

9 years ago
The stack trace has almost only methods which have been removed from the
code ages ago.

And the resize runnables are "revoked" in ::destroy.
Status: NEW → RESOLVED
Last Resolved: 9 years ago
Resolution: --- → INCOMPLETE
You need to log in before you can comment on or make changes to this bug.