Windows Media Encoder and FreeBSD

[Jonathan S. Keim]

Hello all,

I've been trying without success to get the Windows Media Encoder 8
utility to work under WINE on FreeBSD.  I'm using the snapshot from
20040309 with ttydrv enabled and without a sound driver (a headless,
soundcard-less box), although this problem occurs regardless of whether
the sound driver is set to null or wineoss.drv.

Calling wm8eutil.exe without arguments works properly, displaying the
usage() message.  Adding arguments to get more information from the
usage() messages works.  When I add actual encoding file arguments,
however, the output from the trace (--debugmsg +all, last 20 lines)
looks like this:

0009:Call ntdll.RtlAllocateHeap(20dd0000,00000000,00000830) ret=78001383
0009:trace:heap:RtlAllocateHeap (0x20dd0000,00000002,00000830):
returning 20df0758
0009:Ret  ntdll.RtlAllocateHeap() retval=20df0758 ret=78001383
0009:RET  MSVCRT.malloc() retval=20df0758 ret=0106e672
0009:CALL MSVCRT.free() ret=01074266
0009:Call ntdll.RtlEnterCriticalSection(7803a118) ret=78001075
0009:Ret  ntdll.RtlEnterCriticalSection() retval=00000000 ret=78001075
0009:Call ntdll.RtlLeaveCriticalSection(7803a118) ret=780012af
0009:Ret  ntdll.RtlLeaveCriticalSection() retval=00000000 ret=780012af
0009:Call ntdll.RtlFreeHeap(20dd0000,00000000,20df0758) ret=78001434
0009:trace:heap:RtlFreeHeap (0x20dd0000,00000002,20df0758): returning
TRUE
0009:Ret  ntdll.RtlFreeHeap() retval=00000001 ret=78001434
0009:RET  MSVCRT.free() retval=00000001 ret=01074266
0009:Call kernel32.TlsGetValue(00000003) ret=01064fa0
0009:Ret  kernel32.TlsGetValue() retval=20398d38 ret=01064fa0
0009:CALL MSVCRT.malloc() ret=01074227
0009:Call ntdll.RtlAllocateHeap(20dd0000,00000000,00000430) ret=78001383
0009:trace:heap:RtlAllocateHeap (0x20dd0000,00000002,00000430):
returning 20df0758
0009:Ret  ntdll.RtlAllocateHeap() retval=20df0758 ret=78001383
0009:RET  MSVCRT.malloc() retval=20df0758 ret=01074227

The process jumps to 100% CPU (wineserver remains at 0%).  At this point
the only way to kill the wine process is to kill -9 or ctrl-\. 
Examining the core file generated by the latter creates a core with the
following backtrace:

#0  0x2024344f in init_handler (sigcontext=0x2068e4c8)
    at ../../include/wine/library.h:199
#1  0x202434c1 in setup_exception (sigcontext=0x2068e4c8, 
    func=0x20243703 <raise_segv_exception>) at signal_i386.c:721
#2  0x2024388d in segv_handler (__signal=10, code=65562,
__context=0x2068e4c8)
    at signal_i386.c:963
#3  0xbfbfffac in ?? ()
#4  0x202434c1 in setup_exception (sigcontext=0x2068e6b0, 
    func=0x20243703 <raise_segv_exception>) at signal_i386.c:721
#5  0x2024388d in segv_handler (__signal=4, code=1,
__context=0x2068e6b0)
    at signal_i386.c:963
#6  0xbfbfffac in ?? ()
#7  0x10580b1 in pe_load.0 ()
#8  0x1043a6e in pe_load.0 ()
#9  0x100e433 in pe_load.0 ()
#10 0x10038ac in pe_load.0 ()
#11 0x1002261 in pe_load.0 ()
#12 0x1067123 in pe_load.0 ()
#13 0x204e645c in start_process (arg=0x0) at process.c:763

Any ideas what might be causing this?  This is an easily reproducable
problem, so I can get whatever other traces you might find interesting
very easily.

Jon