Closed Bug 31934 Opened 25 years ago Closed 24 years ago

rapid clicking on <img> leads to crash

Categories

(SeaMonkey :: General, defect, P3)

x86
Linux
defect

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: moshev, Assigned: pavlov)

References

()

Details

(Keywords: crash, qawanted)

Attachments

(2 files)

When several pages are linked with navigation bar made of <IMG> like following: <BODY > <!--Navigation Panel--> <A NAME="tex2html41" HREF="node2.html"> <IMG WIDTH="37" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="next" SRC="next.gif"></A> <A NAME="tex2html37" HREF="dhtml.html"> <IMG WIDTH="26" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="up" SRC="up.gif"></A> <A NAME="tex2html31" HREF="dhtml.html"> <IMG WIDTH="63" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="previous" SRC="prev.gif"></A> <A NAME="tex2html39" HREF="node2.html"> <IMG WIDTH="43" HEIGHT="24" ALIGN="BOTTOM" BORDER="0" ALT="index" SRC="index.gif"></A> <BR> <B> Next:</B> <A NAME="tex2html42" HREF="node2.html">Index</A> <B> Up:</B> <A NAME="tex2html38" HREF="dhtml.html">iBASE DHTML Client Architecture</A> <B> Previous:</B> <A NAME="tex2html32" HREF="dhtml.html">iBASE DHTML Client Architecture</A> &nbsp <B> <A NAME="tex2html40" HREF="node2.html">Index</A></B> <BR> <BR> <!--End of Navigation Panel--> (Which is produced by latex2html) and one is rapidly clicking on navigation buttons, the browser will crash and leave the following on the console: Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' .//run-mozilla.sh: line 29: 32509 Segmentation fault $prog ${1+"$@"} or sometimes: Gtk-WARNING **: invalid class type `GtkSignal' in cast to `GtkObject' Gtk-WARNING **: invalid class type `GtkSignal' in cast to `GtkObject' .//run-mozilla.sh: line 29: 32606 Segmentation fault $prog ${1+"$@"} It may be affected by the fact that the web server is on local machine, and the speed of page loading is fast enough to be done before all mouse events are processed. The second condition seems to be that there must be similar navigation bar on the next linked page, etc, in order for the mouse events to go through to another object of the same type, but of different instance. This is on Linux, build 2000031418
moshev@easybase.com - we aren't going to be able to do anything with this unless you can provide a URL of a page where you see this problem. Where did you find this HTML snippet? Gerv
It's on my pc, I really can not give you access to this page, since it has internal company's info, and i also think that this will only happen with files residing on a local machine, since when working through even LAN connection, the synchronization problems might be gone. I'll try to create a small test case with two html files that will reproduce this.
moshev@easybase.com if you can create that small test case use the "Create a new attachment" link (located under the SUmmary and Keywords fields of this bug report) to attach the test files. Thanks.
Whiteboard: Awaiting Test Case from reporter
OK, It will take me a day or two, i am a bit busy. (Trying to figure out how macintoshes work :)
asadotzler@netscape.net (Asa Dotzler) taking ownership of Browser General bugs. New QA Contact is jelwell@singleclick.com (Joseph Elwell). Sorry for the spam.
Assignee: cbegle → asadotzler
QA Contact: asadotzler → jelwell
moshev@easybase.com - any progress on a test case for this bug? :-) Gerv
I'll try to produce one today, sorry it takes so long.
I have created an attachment, however, today's build is rendering so slow for me, that i can't see the crash. (i believe that the speed with which the pages change is crucial to see the crash). I'd still leave this open for further testing.
clearing whiteboard. moshev@easybase.com, can you report back here if you can repeat this crash. Thanks
Whiteboard: Awaiting Test Case from reporter
I'll try again in couple of days.
Ok, great, it happens on M15 with talkback, so i just sent the talkback in. Talkback incident id is: TB8983341H
moshev@easybase.com is it happening with current M16 nightly builds?
Yes, but not with all of them, It does fail with today's nightly (2000042208). The last lines I see on the command line are: Error loading URL http://mvl/dhtml/node2.html ->>>>>>>>>>>>>> Write Clipboard to memory ->>>>>>>>>>>>>> Write Clipboard to memory Gtk-WARNING **: invalid unclassed pointer in cast to `GtkObject' Gtk-WARNING **: invalid unclassed pointer in cast to `GtkObject' .//run-mozilla.sh: line 29: 10109 Segmentation fault $prog ${1+"$@"}
leger, elig, I'm not sure where to go with this one. Any help would be appreciated. Can someone from netscape get a look at the talkback report posted by moshev@easybase.com. Thanks.
Stack crawl:      libwidget_gtk.so + 0x2940c (0x4048040c)        libgdk-1.2.so.0 + 0x15b80 (0x405a5b80)        libglib-1.2.so.0 + 0x10736 (0x405d1736)        libglib-1.2.so.0 + 0x10d43 (0x405d1d43)        libglib-1.2.so.0 + 0x10eec (0x405d1eec)        libgtk-1.2.so.0 + 0x785d9 (0x405275d9)        libwidget_gtk.so + 0x22e1a (0x40479e1a)        libnsappshell.so + 0x1463a (0x4041d63a)        mozilla-bin + 0x320c (0x0804b20c)        mozilla-bin + 0x3523 (0x0804b523)        libc.so.6 + 0x189ee (0x4030d9ee)
Pavlov, might you have an idea who this should go to? The top of the stack trace references Gimp Toolkit widgets, which are your specialty. Thanks!
Assignee: asadotzler → pavlov
Adding crash keyword.
Keywords: crash
Could this be the same event reentrancy issue discussed in bug 37338?
Added qawanted keyword; can someone on Linux have a look at this and (probably) close it? Gerv
Keywords: qawanted
I'm not able to reproduce this bug with Mozilla Build 2000053108 on a PowerMac G3 with 160 MB RAM. NB. I loaded the file with file:///, not using a local web server. Just to be sure if I get it right, the procedure I did was: 1. Downloaded the attachment. 2. Opened it with file:///... 3. Click on the Next button as fast as I can.
May be it's linux specific. I will test it again and resend the results for latest nightly.
Well, that's strange... When running today's build in latest KDE2 build, it didn't segfault. But when i tried it in gnome, it failed as always, although it took me some more time, and i had to adjust the speed of my clicking a bit to hit the bug. Anyway, when falling down the console output looked as follows: ->>>>>>>>>>>>>> Write Clipboard to memory ->>>>>>>>>>>>>> Write Clipboard to memory Document: Done (0.723 secs) *** check number of frames in content area Document http://mvl/dhtml/node4.html loaded successfully ->>>>>>>>>>>>>> Write Clipboard to memory ->>>>>>>>>>>>>> Write Clipboard to memory Document: Done (0.492 secs) *** check number of frames in content area Document http://mvl/dhtml/node5.html loaded successfully Document: Done (0.128 secs) Error loading URL http://mvl/dhtml/node6.html ->>>>>>>>>>>>>> Write Clipboard to memory ->>>>>>>>>>>>>> Write Clipboard to memory Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' Gtk-WARNING **: invalid cast from `(unknown)' to `GtkObject' .//run-mozilla.sh: line 29: 25212 Segmentation fault $prog ${1+"$@"} Hope this helps...
And here is the backtrace with gdb (this time it failed almost immediately) Document: Done (0.144 secs) Error loading URL http://mvl/dhtml/node2.html ->>>>>>>>>>>>>> Write Clipboard to memory ->>>>>>>>>>>>>> Write Clipboard to memory Program received signal SIGSEGV, Segmentation fault. 0x405a9a10 in ?? () from /home/dwnl/mozilla/package/components/libwidget_gtk.so (gdb) bt #0 0x405a9a10 in ?? () from /home/dwnl/mozilla/package/components/libwidget_gtk.so #1 0x406d1b80 in ?? () from /usr/lib/libgdk-1.2.so.0 #2 0x406fa736 in ?? () from /usr/lib/libglib-1.2.so.0 #3 0x406fad43 in ?? () from /usr/lib/libglib-1.2.so.0 #4 0x406faeec in ?? () from /usr/lib/libglib-1.2.so.0 #5 0x406535d9 in ?? () from /usr/lib/libgtk-1.2.so.0 #6 0x405a289c in ?? () from /home/dwnl/mozilla/package/components/libwidget_gtk.so #7 0x4038bc7a in ?? () from /home/dwnl/mozilla/package/components/libnsappshell.so #8 0x804d117 in JS_PushArguments () #9 0x804d539 in JS_PushArguments () #10 0x40257cae in ?? () from /lib/libc.so.6
I have created an attachment with gzipped strace of sigserv, only this time i clicked on hrefs, not on buttons. It's the same html, but with strace, buttons were not quick enough to produce failure.
Some similarities to bug 30583: - a stack trace that shows libwidget_gtk.so at the top - Gtk-WARNING **: invalid cast from (NULL) pointer to `GtkObject'
Sorry for the spam. New QA Contact for Browser General. Thanks for your help Joseph (good luck with the new job) and welcome aboard Doron Rosenberg
QA Contact: jelwell → doronr
->future. If you disagree, please nominate for nsbeta3, giving a good reason why we have to fix this rather than our other bugs.
Target Milestone: --- → Future
Updating QA Contact
QA Contact: doronr → paw
Updated QA Contact
QA Contact: paw → lchiang
Target Milestone: Future → ---
reporter: are you still able to reproduce this?
I'll try later today or perhaps in a next millenium :-)
taking over qa
QA Contact: lchiang → doronr
wfm
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → WORKSFORME
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.

Attachment

General

Creator:
Created:
Updated:
Size: