239966 - Hang in nsAppShell::Run

Status: RESOLVED WORKSFORME

(SeaMonkey :: General, defect)

Description

[Michael Moy]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7b) Gecko/20040323
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7b) Gecko/20040323

I use a pay site called stockcharts.com and every once in a while, it will hang
Mozilla, Phoenix, Firebird, FireFox (I've seen this problem since late 2002).
I can kill the tab that's open but then I can't do anything with the remaining
tabs in the window. I can see the pages but can't click on links or navigate. If
I type something in the URL Window, the type appears to the left instead of the right so I'm typing backwards. If I close the browser, I get a crash.

I managed to get a debu build up and running and the stack at the time of the
hang looks like:

 	user32.dll!77d43c53() 	
>	firefox.exe!nsAppShell::Run()  + 0x17c	C++
 	firefox.exe!nsAppShellService::Run()  + 0x1e	C++
 	firefox.exe!nsCOMPtr<nsIClipboard>::~nsCOMPtr<nsIClipboard>()  + 0x896	C++
 	firefox.exe!xre_main()  + 0x31c	C++
 	firefox.exe!_main()  + 0x53	C++
 	firefox.exe!_WinMain@16()  + 0x1a	C++
 	firefox.exe!WinMainCRTStartup()  Line 390 + 0x1b	C
 	kernel32.dll!77e814c7() 	
 	SHLWAPI.DLL!70a9f1ab() 	

The assembly code of the caller to user32.dll is:

00603E57  push        eax  
00603E58  call        dword ptr [nsToolkit::mPeekMessage (11376C4h)] 
00603E5E  test        eax,eax 
00603E60  je          nsAppShell::Run+114h (603E14h) 
00603E62  jmp         nsAppShell::Run+17Ch (603E7Ch) 
00603E64  cmp         dword ptr [nsWidgetModule_gModuleInfo+30h (11136DCh)],0 
00603E6B  jne         nsAppShell::Run+176h (603E76h) 
00603E6D  mov         dword ptr [ebp-24h],0 
00603E74  jmp         nsAppShell::Run+17Ch (603E7Ch) 
00603E76  call        dword ptr [__imp__WaitMessage@0 (11424CCh)] 
00603E7C  cmp         dword ptr [ebp-24h],0 
00603E80  jne         nsAppShell::Run+84h (603D84h) 
00603E86  mov         ecx,dword ptr [ebp+8] 
00603E89  mov         edx,dword ptr [ecx] 
00603E8B  mov         eax,dword ptr [ebp+8] 
00603E8E  push        eax  

which is why I think that it's hanging on WaitMessage. I couldn't get source
level debugging to work so I had to look for matches in the assembler code and
the source code in an editor.

If I attempt to close the browser in the debugger, it crashes in >	firefox.exe!XPCCallContext::GetJSContext()  + 0xa	C++

Which is this code:

/* readonly attribute JSContextPtr JSContext; */
NS_IMETHODIMP
XPCCallContext::GetJSContext(JSContext * *aJSContext)
{
    *aJSContext = mJSContext;
    return NS_OK;
}

But this just looks like an innocent bystander.

Reproducible: Always
Steps to Reproduce:
I can reproduce this problem by going into www.stockcharts.com, logging into my
account and then playing around with chart settings. Sometimes it takes a few
seconds to reproduce and sometimes it takes 15 minutes to reproduce. It doesn't
occur when I'm not logged in (there is some free content for non-members available).


Expected Results:  
It would be nice if there was a more graceful exit possible.

I'll keep the debug environment around for a while in case there's stuff you
want me to try or things that you'd like dumped out. I don't have local variables
available in my hacked debug build. At the moment I can't do a proper debug build
due to two other bugs where FireFox crashes in the startup with debug builds so
I compiled everything debug and then did a debug link by hand which gave me the
symbols in the stack for the modules, routines and offsets.

Comment 1

[Michael Moy]

Note: The cut and paste of the assembly didn't show that the arrow was pointing to WaitMessage or the line after it.

Comment 2

[Michael Moy]

*** Bug 230958 has been marked as a duplicate of this bug. ***

Comment 3

[Michael Lefevre]

from my limited understanding of nsAppShell::Run, and the info in your other bug
about the page and that the fact it has a java applet, I guess this is something
to do with java, or the interface between Firefox/Mozilla and the java.

what version of the jvm are you running now?

Comment 4

[Michael Moy]

(In reply to comment #3)
> from my limited understanding of nsAppShell::Run, and the info in your other bug
> about the page and that the fact it has a java applet, I guess this is something
> to do with java, or the interface between Firefox/Mozilla and the java.
> 
> what version of the jvm are you running now?

It doesn't really matter. I've had this problem with a few versions of Java.

Comment 5

[Michael Moy]

This website consistently crashes FF on Linux. It's intermittant on Windows but
crashes just about everytime I go to that site on Linux. I don't have a debug
environment on my Linux box though nor do I have a build environment.

Comment 6

[Michael Moy]

I finally got around to building another Debug image and found that there's an
assert just before the crash. Will look at getting more info the next time I run
the debug image.

Comment 7

[Michael Moy]

I tried various versions of Netscape and Netscape 6.2 doesn't show this problem
so that's my workaround for now. Anyone know when Netscape 6.2 forked off the
Mozilla code line?

Comment 8

[Michael Lefevre]

heh - rather a big window to be useful. Netscape 6.2 was based on Mozilla 0.9.4,
which branched on 2001-09-04. See http://www.mozilla.org/releases/cvstags.html
and/or http://www.holgermetzger.de/Netscape_History.html

Comment 9

[Michael Moy]

I was just going to check the module that it crashes/hangs in to see if they'res
anything really obvious. This bug was really getting to me and I was thinking of
going with Opera or IE for this application.

Nice to know that there's at least one other person out there.

Comment 10

[Michael Moy]

Another person reported this problem in a Yahoo Group. I will ask him to post
his comments here.

I do have a workaround though most people will not find it useful. It seems that
this problem goes away on my latest Firefox build running on Windows-64.

Comment 11

[Kevin Brosnan]

*** Bug 276386 has been marked as a duplicate of this bug. ***

Comment 12

[Michael Moy]

One point on this bug: I haven't run into it in quite some time since making
various performance modifications to Firefox. I don't see it on Windows-64
either, which does some emulation for 32-bit software.

I think that the problem is still there but faster software or hardware seems to
mask it.

Comment 13

[Gervase Markham [:gerv]]

This is an automated message, with ID "auto-resolve01".

This bug has had no comments for a long time. Statistically, we have found that
bug reports that have not been confirmed by a second user after three months are
highly unlikely to be the source of a fix to the code.

While your input is very important to us, our resources are limited and so we
are asking for your help in focussing our efforts. If you can still reproduce
this problem in the latest version of the product (see below for how to obtain a
copy) or, for feature requests, if it's not present in the latest version and
you still believe we should implement it, please visit the URL of this bug
(given at the top of this mail) and add a comment to that effect, giving more
reproduction information if you have it.

If it is not a problem any longer, you need take no action. If this bug is not
changed in any way in the next two weeks, it will be automatically resolved.
Thank you for your help in this matter.

The latest beta releases can be obtained from:
Firefox:     http://www.mozilla.org/projects/firefox/
Thunderbird: http://www.mozilla.org/products/thunderbird/releases/1.5beta1.html
Seamonkey:   http://www.mozilla.org/projects/seamonkey/

Comment 14

[Michael Moy]

I personally haven't seen this problem in a long time but then I'm using 64-bit
builds now on a faster machine. The performance of the browser/hardware was a
factor in seeing the problem.

But if noone else considers it a problem and I continue to not see it, then
maybe it was fixed.

Comment 15

[Stockman222]

I'm still finding that Firefox hangs up when I close the "Add Annotations"
window while in stockcharts.com.  I am running Firefox 1.0.7.

To get to the "Add Annotations" window, go to: 
http://stockcharts.com/def/servlet/SC.web?c=$SPX,uu[w,a]daclyyay[de][pb20!b40][vc60][iLb14!La12,26,9]&pref=G

and click on the "Add Annotations" box at the lower part of the page.  A new
window will pop up.  When exiting this window, Firefox hangs up every time.

The link I gave above refers to the non-members area.

Comment 16

[Eyal Rozenberg]

(In reply to comment #15)

(using sm 2005-11-09 nightly) A new window doesn't popup; instead, I get the following JS error:

Error: annotate is not defined
Source File: http://stockcharts.com/def/servlet/SC.web?c=$SPX,uu[w,a]daclyyay[de][pb20!b40][vc60][iLb14!La12,26,9]&pref=G
Line: 1

Can someone reproduce the bug with recent nightlies?

Comment 17

[Michael Moy]

This may or may not be related but when I build with Profile Guided Optimization, the stockcharts annotate screen always crashes on the first build.

I stopped seeing the original problem with my unofficial builds quite some time ago and I suspect that there's a timing window and that the faster unofficial builds close the timing window so that the problem doesn't occur.

Comment 18

[Michael Moy]

Could we mark this as confirmed as it was marked a duplicate of Bugzilla Bug 276386 which was marked as confirmed?

Further comment on Profile-Guided-Optmization: If Mozilla wants to use PGO
in the future, I think that this bug will have to be fixed. I have a manual
procedure that I can use to get PGO to work but it is non-deterministic.

Comment 19

[lenochod]

Michael, Do this bug still appear?

SeaMonkey or Firefox? 

Comment 20

[Michael Moy]

I have not seen this problem in quite some time and I still use stockcharts regularly. I'd be happy to close it out if there are no other complaints on it.

Comment 21

[lenochod]

OK

Mark as WFM