[Bug 35367] Multiple applications crash due to Wine ole32 code not taking implicit MTA into account (Cyberlink Powerdirector 8, PDFXChange Editor 5.5)

[wine-bugs at winehq.org]

https://bugs.winehq.org/show_bug.cgi?id=35367

Anastasius Focht <focht at gmx.net> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|Cyberlink Powerdirector 8   |Multiple applications crash
                   |crashes during media scan   |due to Wine ole32 code not
                   |(Wine ole32 code must take  |taking implicit MTA into
                   |implicit MTA into account)  |account (Cyberlink
                   |                            |Powerdirector 8, PDFXChange
                   |                            |Editor 5.5)

--- Comment #2 from Anastasius Focht <focht at gmx.net> ---
Hello folks,

I found another app suffering from this: PDFXChange Editor 5.5

Refining summary.

Download: http://www.tracker-software.com/product/pdf-xchange-editor/download

Relevant part of trace log:

--- snip ---
...
002e:err:ole:CoGetObjectContext apartment not initialised
002e:Ret  ole32.CoGetObjectContext() retval=800401f0 ret=107ad4ec
002e:Call ole32.CoInitializeEx(00000000,00000000) ret=107b7a3f 
...
002e:Ret  ole32.CoInitializeEx() retval=00000000 ret=107b7a3f 
...
002b:Ret  PE DLL (proc=0x100113ed,module=0x10000000
L"PDFXEditCore.x86.dll",reason=THREAD_ATTACH,res=(nil)) retval=1
002b:Call PE DLL (proc=0x338d95e,module=0x3320000
L"Bookmarks.pvp",reason=THREAD_ATTACH,res=(nil))
...
002b:Ret  PE DLL (proc=0x338d95e,module=0x3320000
L"Bookmarks.pvp",reason=THREAD_ATTACH,res=(nil)) retval=1
002b:Starting thread proc 0x105b27cd (arg=0x2dd8c58)
...
002b:Call KERNEL32.GetLastError() ret=10015b78
002b:Ret  KERNEL32.GetLastError() retval=00000000 ret=10015b78
002b:Call KERNEL32.WaitForSingleObject(00000104,ffffffff) ret=107b8008
002b:Ret  KERNEL32.WaitForSingleObject() retval=00000000 ret=107b8008
002b:Call ole32.CoGetObjectContext(114baa28,0187e9ec) ret=107ad4ec
...
002b:Ret  ole32.CoGetObjectContext() retval=00000000 ret=107ad4ec 
...
002b:Call KERNEL32.CreateEventW(00000000,00000001,00000000,001a3df0
L"{124380E8-8012-4E0C-B00821E5C6FF7C2F}") ret=101c4775
002b:Ret  KERNEL32.CreateEventW() retval=000003c4 ret=101c4775
002b:Call ole32.CoCreateInstance(1144eb54,00000000,00000004,1144eb44,0187e964)
ret=101c48ef 
...
002b:Call KERNEL32.WaitNamedPipeW(0187e5c8
L"\\\\.\\pipe\\{0DF179B8-96F1-4F3E-9338-DFEEB61B810A}",ffffffff) ret=7e9835f3
002b:Ret  KERNEL32.WaitNamedPipeW() retval=00000001 ret=7e9835f3
002b:Call KERNEL32.CreateFileW(0187e5c8
L"\\\\.\\pipe\\{0DF179B8-96F1-4F3E-9338-DFEEB61B810A}",c0000000,00000000,00000000,00000003,00000000,00000000)
ret=7e983634
002b:Ret  KERNEL32.CreateFileW() retval=000003c8 ret=7e983634 
...
002b:Ret  KERNEL32.GlobalLock() retval=02dbc770 ret=7e962981
002b:Call KERNEL32.GlobalUnlock(02b40182) ret=7e962a38
002b:Ret  KERNEL32.GlobalUnlock() retval=00000000 ret=7e962a38
002b:Call ntdll.RtlAllocateHeap(00110000,00000000,00000010) ret=7e96bb12
002b:Ret  ntdll.RtlAllocateHeap() retval=02d58408 ret=7e96bb12
002b:err:ole:StdMarshalImpl_UnmarshalInterface Apartment not initialized
002b:err:ole:CoUnmarshalInterface IMarshal::UnmarshalInterface failed,
0x800401f0
...
002b:trace:seh:raise_exception code=c0000005 flags=0 addr=0x7e983b26
ip=7e983b26 tid=002b
002b:trace:seh:raise_exception  info[0]=00000000
002b:trace:seh:raise_exception  info[1]=6c61767c
002b:trace:seh:raise_exception  eax=6c617674 ebx=7ea68000 ecx=0187e4a0
edx=00110064 esi=0187e830 edi=0187e914
002b:trace:seh:raise_exception  ebp=0187e718 esp=0187e4a0 cs=0023 ds=002b
es=002b fs=0063 gs=006b flags=00010286
002b:trace:seh:call_stack_handlers calling handler at 0x1132ad90 code=c0000005
flags=0 
...
--- snip ---

Works fine with my rebased proof-of-concept bugfix branch I kept for the
initial report.

NOTE: Although bug 33749 is essentially about the same issue, the summary and
proposed patch only covers a narrow scope which doesn't work for the
applications here.

$ sha1sum PDFXVE5.x86.msi 
0470e3ff74b75f1954a2236fa44749fc4a0cd1a9  PDFXVE5.x86.msi

$ du -sh PDFXVE5.x86.msi 
38M    PDFXVE5.x86.msi

$ wine --version
wine-1.7.35-89-gbad99c9

Regards

-- 
Do not reply to this email, post in Bugzilla using the
above URL to reply.
You are receiving this mail because:
You are watching all bug changes.