352438 - Minefield 3.0a1 crashes on startup (Fresh install) [msvcr80.dll]

Status: VERIFIED DUPLICATE of bug 330271

(Firefox :: General, defect)

Description

[RenegadeX]

User-Agent:       Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.8.1b2) Gecko/20060821 Firefox/2.0b2
Build Identifier: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.9a1) Gecko/2006090205 Firefox/3.0a1

I recently installed Minefield 3.0a1 for the first time, using the 'lastest trunk' win32 exe installer (and then later zip version). When the install was done, I was prompted to Launch Minefield. I did so, the main Firefox window appeared and moments later the 'Default Browser Check' appeared on the screen. At that point, the app crashes and a "Firefox.exe has encountered a problem..." dialog appears. When I click for 'more information', it tells me that there is an error in msvcr80.dll. (I will attach a screenshot with the exact details).

If I do a Custom install and choose to install the Feedback (Talkback) extension , the crash is not caught (logged). It also does not show up in Dr. Watson.

I tried it on: 3x WinXP-Pro (SP2) PCs, 1x WinXP-Media Center Edition (SP2), 1x WinXP-Home Edition (SP2). Of the 5 computers, I was only able to get Minefield to run without crashing on the 1 WinXP-Media Center Edition system. Possibly a coincidence, but the XP-MCE system is the only one not running some version of Zone Alarm. But even uninstalling ZA, rebooting and reinstalling Minefield is not successful.

I've uninstalled and reinstalled Minefield and tried various differing install options, but always with the same result. It even crashes in the same place in Firefox Safe Mode.

I made a post on Mozillazine forums: http://forums.mozillazine.org/viewtopic.php?t=462490
and it seems I am not alone. One user reports not being able to run Minefield for over 3 months now. I have tried new 'latest trunk' installs all week long, with no luck - and it seems clear now that it's not going to work. One poster apparently did a System Restore and was finally able to get Minefield to work for them - this also was unsuccessful on my systems. Regardless, Firefox 2.0b2 and all previous versions worked fine on all my systems - Minefield should be no different.

While it obviously is not affecting everyone, this a new, critical issue.

Reproducible: Always

Steps to Reproduce:
1. Install Minefield 3.0a1 Win32 (exe or zip)
2. Launch it
3. Watch it crash before it's done starting.

Comment 1

[RenegadeX]

Attachment: Screencap of Minefield startup error [msvcr80.dll]

Comment 2

[Adam Guthrie]

It's the same crash as the bug Bill mentioned in the MozillaZine thread. It's even crashing in the same place in the module as reported in the other bug (000176d9).

*** This bug has been marked as a duplicate of 330271 ***

Comment 3

[RenegadeX]

Additional information:

Bug 330271, "msvcr80.dll crashes after trying to check for updates" seems to be very similar. Comment #9 in that thread indicates that the originally reported issue (ie: happening on updates) was now happening on startup as well.

A poster on Mozillazine therefore suggested I go into my prefs.js file and disable any of the auto updates. So I tried this, adding the following:
user_pref("app.update.enabled", false);
user_pref("extensions.update.enabled", false);
user_pref("browser.search.update", false);

.. with the same crash as always. Therefore, this is distinctly different from Bug 330271 as it seems that it is not update-related.

Also tried adding:
user_pref("security.enable_tls", false);
user_pref("enable.security_ssl3", false);

.. with no improvement.

Bug 330271 suggested the use of WinDbg to report the debug logs. So I've followed the instructions and this is what I get:

ChildEBP RetAddr  
0012fdd8 7c90e89a ntdll!KiFastSystemCallRet
0012fddc 7c81cd96 ntdll!ZwTerminateProcess+0xc
0012fed8 7c81cdee kernel32!_ExitProcess+0x62
0012feec 78131720 kernel32!ExitProcess+0x14
0012fef4 781319ee MSVCR80!__crtExitProcess(int status = 2014517814)+0x14 [f:\rtm\vctools\crt_bld\self_x86\crt\src\crt0dat.c @ 683]
0012ff2c 78131a36 MSVCR80!doexit(int code = 0, int quick = 0, int retcaller = 0)+0xac [f:\rtm\vctools\crt_bld\self_x86\crt\src\crt0dat.c @ 596]
0012ff3c 0092a5ec MSVCR80!exit(int code = 2088857559)+0xd [f:\rtm\vctools\crt_bld\self_x86\crt\src\crt0dat.c @ 398]
WARNING: Stack unwind information not available. Following frames may be wrong.
0012ffc0 7c816fd7 firefox!jpeg_free_large+0x56fd
0012fff0 00000000 kernel32!BaseProcessStart+0x23

Hopefully that is of some use.

Comment 4

[Peter van der Woude [:Peter6]]

don't dupe bugs against UNCO's please !!!

Comment 5

[timeless]

duping an unconfirmed against an unconfirmed is perfectly acceptable. especially if it already has developers trying to look at it.