When posting a reply to a discussion board in WebCT 4.1, the browser crashes.




JavaScript Engine
10 years ago
10 years ago


(Reporter: Brian, Unassigned)



1.8 Branch
Windows XP

Firefox Tracking Flags

(Not tracked)



(3 attachments)



10 years ago
User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv: Gecko/20080311 Firefox/
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv: Gecko/20080311 Firefox/

Using Webct 4.1 through firefox, when you post a reply to a discussion post or an mail message the browser will crash after to post has been submitted.  The post seems to be saved on the server.

Reproducible: Always

Steps to Reproduce:
1. Log into WebCT
2. Open a course with a discussion board.
3. Open a post.
4. Click reply.
5. Type something in the message box.
6. Click Post.
7.  The popup window will close and then the browser will crash.
Actual Results:  
Browser crashes.

Expected Results:  
Page should refresh and show you the new post.

Have reproduced on both Windows Xp and Windows Vista.

Comment 1

10 years ago
Link to someone else with same issue.

Version: unspecified → 2.0 Branch
Do you get the crash also in the safemode ?
- http://kb.mozillazine.org/Safe_Mode

If yes: Please provide a Talkback id from that crash :
- http://kb.mozillazine.org/Talkback
Keywords: crash

Comment 3

10 years ago
TB43373889Q Talkback ID

Yes it also crashes in Safemode.

js_MarkGCThing  [mozilla/js/src/jsgc.c, line 2447]
JS_GC  [mozilla/js/src/jsapi.c, line 1879]
nsAppStartup::Run  [mozilla/toolkit/components/startup/src/nsAppStartup.cpp, line 152]
main  [mozilla/browser/app/nsBrowserApp.cpp, line 61]
kernel32.dll + 0x16fd7 (0x7c816fd7)
could be related to bug 425576
Assignee: nobody → general
Component: General → JavaScript Engine
Product: Firefox → Core
QA Contact: general → general
Version: 2.0 Branch → 1.8 Branch
Duplicate of this bug: 426816

Comment 7

10 years ago
Crash also occurs w/ posting in WebCT CE 4 Mail tool. Prob. same mechanism invoked on post to close popup compose window. Crash does not occur on Mac OS X or Linux FF

Comment 8

10 years ago
I believe the issue is triggered in the following manner.

1 - pop up window opens
2 - pop up calls a CGI action
3 - CGI action sends back a close()

I don't have access to the CGI source, but here's what I was able to snag traffic-wise.  The CGI sends back:

Content-Type: text/html; charset=utf-8

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.01 Frameset//EN">
<html dir="ltr">
<link rel="STYLESHEET" href="/web-ct/en8/style/styles_all4x.css" type="text/cs
<style type="text/css">
@import "/web-ct/en8/style/styles_nonn4.css";
A.action IMG { vertical-align:0% !important; }

Comment 9

10 years ago
Created attachment 313818 [details]
show compose has the pop up to open compose.html

Comment 10

10 years ago
Created attachment 313819 [details]
compose.html has the post button on it w/c calls the cgi

Comment 11

10 years ago
Created attachment 313820 [details]
close_compose is a PERL CGI that spits back what I saw traffic wise

Comment 12

10 years ago
In my close_compose, I've stripped out link and stylesheet lines as that the same outcome was still generated.

Unhandled exception at 0x600f0290 in firefox.exe: 0xC0000005: Access violation reading location ...

Comment 13

10 years ago
test case. put show_compose.html and compose.html on a server, put close_compose in the server's cgi-bin.  call show_compose.html from the ff browser, click on the 'Compose message' link.  in the compose.html pop up, click on the post.


10 years ago
Ever confirmed: true

Comment 14

10 years ago
May have been resolved with the release of FF  So far testing has not crashed.
we created because of bug 425576 (see comment #5).
If this works for everynone else with i will mark this as dupe of bug 425576

Comment 16

10 years ago
The similarity to bug 425576 is that its a crash in the js3250.dll.  However, the excite.jp case also crashed on FF on OS X, which we have not seen.  Perhaps the fix implement for the excite.jp case fixed a broader issue, and we were a subset?  Anyways, I'm happy FF was released!
You can't compare crashes based on the crashing ddl like "js3250.dll", but I can compare based in the stack trace.

I'm sure that that this is the same garbage collector crash, marking as dupe
Last Resolved: 10 years ago
Resolution: --- → DUPLICATE
Duplicate of bug: 425576
You need to log in before you can comment on or make changes to this bug.