Closed Bug 40324 Opened 24 years ago Closed 24 years ago

Crash launching Win 98

Categories

(SeaMonkey :: Installer, defect, P1)

Product:
SeaMonkey
Component:
Installer
Platform:
x86
Windows 98
Type:
defect

Tracking

(Not tracked)

Status:
VERIFIED FIXED

People

(Reporter: esther, Assigned: ssu0262)

References

Details

(4 keywords, Whiteboard: [dogfood+][nsbeta2-]) 

Using build 2000-05-23-10 on win98 only, I crash after the splash screen when 
launching netscap6.exe.   Does not happen on Nt 4.0 or win95 per Grace.
adding keywords, no workaround found yet.
Keywords: crash, nsbeta2, pp, regression
agree- just tried both installs (typical and custom) on 98 and crash both times- 
after splash screen- error in xpcom.dll
It crashes on my PIII running Win98, but launches ok on my PII running Winnt 4.0 
SP6.0.
umm.. isn't this a duplicate of bug 40306 ?

on win2000 and 98, i can see it crashing. Related perhaps to bug 40306, as i 
crash on the following console line:

Profile Manager : Profile Wizard and Manager activites : End

*** Bug 40332 has been marked as a duplicate of this bug. ***
ok, on today's may 24 win32 mozilla build, I crash just like yesterday. 

I crash on line Profile Manager : Profile Wizard and Manager activites : End
in console at module <unknown>

Marking this blocker, smoketest keyword and setting to m16. This makes win32 
unusable.
Severity: critical → blocker
Keywords: smoketest
Priority: P3 → P1
Target Milestone: --- → M16
why is this assigned to me?  reassigning to default component owner.
Assignee: granrose → ssu
adding me and shrir to Cc: list...we need to get this fixed ASAP...its
problematic in today's 5/24 build also.
chaning summary to "crash launching mozilla", netscape6 in a subject won't get 
much attention.
Summary: Crash launching netscap6.exe → Crash launching mozilla
*** Bug 40427 has been marked as a duplicate of this bug. ***
This looks very much like a dup of 40294 except it's not on Linux.  Are these 
bugs related?
Can we get a stack or a talkback log, please?
For Win98...

The last working build for me was 2000-05-22-09-M16.

The following builds all crash at startup:
2000-05-23-10-M16
2000-05-23-21-M16
2000-05-24-09-M16

It makes no difference whether I run the installer versus unzipping.  TalkBack 
does not engage when crash occurs.
could be a dupe of 40294. but, no -profilemanager is need to crash. Plus, it
seems to have been fixed by gbush's comment.

I have almost the same problems on W2K but only with the optimized build. The 
debug builds are starting smoothly, so a stack trace is not possible.

It stops at GKGFXWIN() (don't know if that helps)
I have a debug build going right now.  Once done, I'll try to get a better stack 
trace under Win98.

(fyi: this is not an installer issue)
Putting on [dogfood+] radar.
Keywords: dogfood
Summary: Crash launching mozilla → Crash launching Win 98
Whiteboard: [dogfood+][nsbeta2-]
the stack is destroyed, I only get two frames, one of which is in GKGFXWIN, and 
above that is just garbage :(
I got a debug build under Win98, but it doesn't look like it's crashing the 
same way/place.  Here is a snippet of the console window that I get now:

=======================================================================
*** Registering xpconnect components (all right -- a generic module!)
*** Registering xpconnect test components (all right -- a generic module!)
*** Registering nsSoftwareUpdate components (all right -- a generic module!)
*** Registering nsPrefModule components (all right -- a generic module!)
nsNativeComponentLoader: autoregistering succeeded
registering (all right -- a JavaScript module!)*** Deferring registration of sam
ple JS components
************************************************************
** NOTE: This report will only be printed in DEBUG builds.**
* Call to xpconnect wrapped JSObject produced this error:  *
[Exception... "Factory not registered (may be tried again) [nsIModule::registerS
elf]"  nsresult: "0x80040155 (NS_ERROR_FACTORY_REGISTER_AGAIN)"  location: "<unk
nown>"  data: no]
************************************************************
-*- sidebar: registering (all right -- a JavaScript module!)
*** Registering sample JS components
nNCL: registering deferred (0)
nNCL: registering deferred (0)
Profile Manager : Profile Wizard and Manager activites : Begin
Profile Manager : Command Line Options : Begin
Profile Manager : Command Line Options : End
ProfileManager : GetProfileDir
ProfileName : default
ProfileDir  : D:\TREES\SEAMONKEY\MOZILLA\DIST\WIN32_D.OBJ\Users50\default
ProfileManager : GetProfileDir
~nsProfile
Unloading component.*** Unloading sample JS components
-*- sidebar: Unloading component.
CanUnload_enumerate: skipping native
###!!! ASSERTION: Component Manager being held past XPCOM shutdown.: 'cnt == 0',
 file d:\trees\seamonkey\mozilla\xpcom\build\nsXPComInit.cpp, line 634
###!!! Break: at file d:\trees\seamonkey\mozilla\xpcom\build\nsXPComInit.cpp, li
ne 634
==========================================================================

It looks like right after profile manager is launched, it's destructor is then 
called.  Mozilla then asserts, then simply quits with an exit code of 1.  Does 
any above ring any bells for anyone?

fixed, oh finally fixed.
Status: NEW → RESOLVED
Closed: 24 years ago
Resolution: --- → FIXED
Problem fixed.  Pinkerton just checked in a couple of files in gfx/src/windows 
that fixes this bug: nsScreenManagerWin.cpp and nsScreenWin.cpp

I'll let pinkerton elaborate on the resolution (if he wants to).

Thanks also to Bhuvan Racham, Steve Elmer, and Don Melton who help me track down 
this bug too.

The fix will show up in tomorrow's build.

Update: Build 2000-05-24-21 works for me now on Win98.  I'll let Grace verify it 
though.
Working on Windows 2000 too...

*fine*
Update: Build 2000-05-24-21 works for me now on Win98 installed on Dell PIII.
:)  I can successfully launch the commercial build.
build 2000-05-24-21 launching on my Win98
Status: RESOLVED → VERIFIED
200052508 mozilla nightly build, it is fixed. good work!
Product: Browser → Seamonkey
You need to log in before you can comment on or make changes to this bug.