423957 - Firefox crashes when I open The Sun website.

Status: RESOLVED WORKSFORME

(Core :: XUL, defect)

Description

[Paul]

User-Agent:       Mozilla/4.0 (compatible; MSIE 7.0; Windows NT 5.1; .NET CLR 2.0.50727; .NET CLR 3.0.04506.30; .NET CLR 3.0.04506.648)
Build Identifier: Firefox 2.0.0.12

When I go to the Sun website (www.thesun.co.uk) Firefox crashes.  I tried to report this using the "Report Broken Website" feature in Firefox, however the application crashes when I access the site.

Reproducible: Always

Steps to Reproduce:
1. Open Firefox
2. In the address bar, type in http://www.thesun.co.uk
3. Press enter
Actual Results:  
Firefox Crashes after a few seconds while the page is loading.

Expected Results:  
Firefox should have displayed the webpage correctly.  I'm pretty sure there must be some sort of a Flash control or add-in used on the website that is causing this behaviour to occur.  Currently I have to use Internet Explorer if I want to view this site as Firefox crashes every time.

I can repro this bug every time, on two different machines.  One of the machines runs Windows XP SP2, the other is running Windows Vista.

I don't know which module this crash occurs in.  The message that comes up is "An unhandled win32 exception occurred in firefox.exe [3740]"

When I use the Visual Studio debugger, the following information is shown.

Unhandled exception at 0x77ea1611 in firefox.exe: 0xC0000005: Access violation reading location 0x00000004.

The Visual Studio debugger seems to highlight rpcrt4.dll however I don't know if this is where the crash is coming from.

Comment 1

[Carlos Roberto]

It seems to be another Flash Player bug. Are you using 9,0,115,0 version?

Comment 2

[Matthias Versen [:Matti]]

Please either provide a brakpad ID from about ctahses and if breakpad doesn't work create a stack trace with this instructions :
http://developer.mozilla.org/en/docs/How_to_get_a_stacktrace_with_WinDbg

Comment 3

[Paul]

I have tried to capture a stack trace using the instructions in the link but it didn't seem to work.  Instead I have got the information above which has come from the Visual Studio debugger.  When the crash happens I am given the option to debug in Visual Studio.

 	firefox.exe!nsXULTreeBuilder::CompareMatches(nsTemplateMatch * aLeft=0x006e0069, nsTemplateMatch * aRight=0x006f0064)  Line 1916	C++
 	firefox.exe!nsXULTreeBuilder::CompareMatches(nsTemplateMatch * aLeft=0x006e0069, nsTemplateMatch * aRight=0x006f0064)  Line 1916	C++
 	firefox.exe!nsXULTreeBuilder::CompareMatches(nsTemplateMatch * aLeft=0x006e0069, nsTemplateMatch * aRight=0x006f0064)  Line 1916	C++
 	firefox.exe!nsXULTreeBuilder::CompareMatches(nsTemplateMatch * aLeft=0x006e0069, nsTemplateMatch * aRight=0x006f0064)  Line 1916	C++
 	firefox.exe!nsXULTreeBuilder::CompareMatches(nsTemplateMatch * aLeft=0x00000000, nsTemplateMatch * aRight=0x00000000)  Line 1916	C++


>	rpcrt4.dll!SWMRWaiter::FreeWaiterData() 	
 	rpcrt4.dll!SWMRLock::FreeWaiter()  + 0xd bytes	
 	rpcrt4.dll!SWMRLock::FreeWaiterCache()  + 0x14 bytes	
 	rpcrt4.dll!__SEH_epilog4_GS()  + 0x57dab bytes	
 	rpcrt4.dll!__InitializeDLL@12()  + 0x293 bytes	
 	rpcrt4.dll!_InitializeDLL@12()  + 0x23 bytes	
 	ntdll.dll!_LdrpCallInitRoutine@16()  + 0x14 bytes	
 	ntdll.dll!_LdrShutdownThread@0()  + 0xde bytes	
 	ntdll.dll!_RtlExitUserThread@4()  + 0x2a bytes	
 	msvcrt.dll!__endthreadex()  + 0x23 bytes	
 	msvcrt.dll!__endthreadex()  - 0x25 bytes	
 	msvcrt.dll!__endthreadex()  + 0xf2 bytes	
 	msvcrt.dll!__endthreadex()  + 0xf2 bytes	
 	kernel32.dll!@BaseThreadInitThunk@12()  + 0x12 bytes	
 	ntdll.dll!___RtlUserThreadStart@8()  + 0x27 bytes	
 	ntdll.dll!__RtlUserThreadStart@8()  + 0x1b bytes

Also I don't know what 'breakpad ID' is.

If you load Firefox and go to www.thesun.co.uk the crash happens immediately so you should be able to obtain the stack trace yourself that way, if the information above is not enough.

Comment 4

[timeless]

about:crashes would give you a breakpad id if you were using a trunk build.

sadly i can't figure out what you got from visual studio. the second stack is probably from a random irrelevant thread. and i can't tell if you have symbols for the first stack (you should, otherwise it shouldn't be able to give line numbers, but it's too incomplete).

irc.mozilla.org has people who can talk you through windbg if there are really problems.... but if you have problems w/ that doc, please feel free to explain them here, someone will work to improve it.

Comment 5

[Paul]

When I attempt to do this and I get to the part where I choose "Go" from the Debug menu, often Firefox does not load and instead I get some sort of violation in the Debug window and this is even before I have attempted to repro the bug.  If I load Firefox, then complete the steps in the document, then repro the bug, instead of getting the stack trace I get the information listed in the FAQ section, however I do have the "Debug child processes" checkbox checked, so I'm not sure what the issue is.

I have never use WinDBG before yesterday so it could be possible that I am doing something wrong, however I am following the instructions as listed in that doc, however the behaviour is not what the document expects.

I will try this again on an XP machine as I have been trying to get the stack trace on a Vista machine.

I still think it would be much faster for a developer to just repro this and get the stack trace themselves as it takes a total of about 30 seconds to repro the bug.  Load Firefox, type in www.thesun.co.uk into the address bar, hit enter and once it reaches the page, Firefox crashes.

It could indeed be some sort of a Flash bug but even if it is, it would be nice if it didn't crash Firefox.

Comment 6

[timeless]

fwiw, after a lot of work, i managed to reach your page w/ enough support to see the content (trunk), and i didn't crash.

no, you'll just have to live w/ being walked through the debugger.

w/ debug child processes you might (will?) see more than one process in the view>process and threads window, you would want to select the one w/ more threads (it's probably lower in the list..)

Comment 7

[Paul]

I have tried this on Firefox 3 and it is not crashing.