14715 - Assertion failure: me == CurrentThreadId(), at jslock.c:699

Status: RESOLVED WORKSFORME

(Core :: DOM: UI Events & Focus Handling, defect, P3)

Description

[milindc]

hi!
I have CC'ed  Christopher Blizzard for Part 2


PART 1 of 2
-----------
Apprunner crashes...  Toolkits: Gtk and Xlib (others not tested)

steps for reproduction...
1. run apprunner
2. load http://www.corel.com
3. click on   Product Index   text link on left hand side
4. (new page loads)  scroll to bottom to see a combo-box
5. select product CorelDRAW 3.5 for UNIX (any choice would do I guess)
6. click on  Go
7. *crash*


PART 2 of 2
-----------
I have created a sample application  webshell/embed/xlib/qt/embed_xlib_into_qt
for Qt 1.44
This is similar to the other embed_xlib_into_xxx applications.

clicking on a particular link crashes embed_xlib_into_qt and not
embed_xlib_into_gtk

Steps:
1. build  embed_xlib_into_qt  by setting  MOZ_QT_CFLAGS and MOZ_QT_LDFLAGS
   in config/autoconf.mk
   I recommend this so that you are not required to build with  --with-qt
2. run embed_xlib_into_qt
3. load http://www.corel.com
4. click on the graphical menu link labelled as "SHOP" just below top banner
5. application crashes

IMPORTANT:
1. However, if the Gtk sample, embed_xlib_into_gtk is used for this test, it
   doesn't crash.
2. Also, clicking on some other link, particularly non-graphic, the application
   mostly does not crash.
3. The timing or instant of the crash is not consistent.
   Sometimes the application crashes before rendering the next page....
   sometimes it crashes after the page is rendered completely.

Eventually I want to use the Qt based webshell in my application and hopefully
I won't face these bugs.


Thanks.

Comment 1

[joki (gone)]

Any chance of a stack trace on this?  I'd just like to try to figure out a
potential owner more closely before I try to tackle this.

Comment 2

[milindc]

Part 1 of 2
-----------
I hate this... I have 44 stack frames but I get a sigsuspend() from
libc.so.6
from pthread_create@@GLIBC_2.1   even before I see apprunner
I doubt this dump is of any use.

My system configuration:
Debian Linux i386
kernel 2.0.36    i686
g++ 2.95.1, GNU gdb 4.18
Qt 1.44
libX11.so.6.1
icewm 0.9.48-2
Mozilla pull: cvs co -D "09/22/1999 14:00:00 PST" SeaMonkeyAll

Hope this helps somewhat.


Part 2 of 2
-----------
Working for a stack dump on this...
(stupid Qt build)

Comment 3

[joki (gone)]

Waqar?  Could you take a look at this one too?  Thanks.

Comment 4

[waqar]

User is running on Debian Linux. I will have to get a machine with Debian
installed so I can test it. Redhat does not show the problem. I will try to a
machine with Debian linux.

Comment 5

[waqar]

At this time we only have RedHat Linux setup, I will resolve this bug as later.

Comment 6

[ckritzer (gone)]

Mass update:  changing qacontact to [e-mail]

Comment 7

[Loco]

Updating QA Contact.

Comment 8

[Blake Ross]

reopening and marking Future...

Comment 9

[waqar]

Needs a test case, Corel has changed their web site.

Comment 10

[Hixie (not reading bugmail)]

Reassigning QA Contact for all open and unverified bugs previously under Lorca's
care to Gerardo as per phone conversation this morning.

Comment 11

[Madhur Bhatia]

QA contact updated

Comment 12

[Kevin McCluskey (gone)]

bulk reassigning Waqar's bugs to Don.

Comment 13

[dcone (gone)]

I can not repro this.  Marking WFM.