On January 11, 2005 08:02 am, Bill Medland wrote: > Any idea what is going wrong in the OXID handling here?
See below for theory > > Is "a" the first thread of the new process? Is the problem that it should > be "b"? > > (I've removed a few extraneous traces) > (This is after the COM_CreateApartment fixes patch) > > 0009:trace:ole:create_marshalled_proxy > rclsid={4dfceca2-156a-4e02-9d02-2e5253ad1d01}, > iid={00000001-0000-0000-c000-000000000046} > 0009:trace:ole:create_marshalled_proxy waiting for > \\.\pipe\{4DFCECA2-156A-4E02-9D02-2E5253AD1D01} > 0009:trace:ole:create_server activating local server > 'L"G:\\runtime\\a4wlpmgr.exe -Embedding"' for > {4DFCECA2-156A-4E02-9D02-2E5253AD1D01} > 0009:trace:ole:create_marshalled_proxy waiting for > \\.\pipe\{4DFCECA2-156A-4E02-9D02-2E5253AD1D01} > 0009:warn:ole:create_marshalled_proxy Could not open named pipe to broker > \\.\pipe\{4DFCECA2-156A-4E02-9D02-2E5253AD1D01}, le is 2 > > 000b:trace:ole:DllMain 0x540000 0x1 0x1 > > 0009:trace:ole:create_marshalled_proxy waiting for > \\.\pipe\{4DFCECA2-156A-4E02-9D02-2E5253AD1D01} > 0009:warn:ole:create_marshalled_proxy Could not open named pipe to broker > \\.\pipe\{4DFCECA2-156A-4E02-9D02-2E5253AD1D01}, le is 2 > 000b:trace:ole:CoInitializeEx ((nil), 0) > 000b:trace:ole:CoInitializeEx () - Initializing the COM libraries > 000b:trace:ole:RunningObjectTableImpl_Initialize () > 000b:trace:ole:COM_CreateApartment thread 0xb is entering the multithreaded > apartment 000b:trace:ole:CoRegisterClassObject > ({4dfceca2-156a-4e02-9d02-2e5253ad1d01},0x405701a0,0x00000004,0x00000005,0x >412054) 000b:trace:ole:CoMarshalInterface (0x4006b7a0, > {00000001-0000-0000-c000-000000000046}, 0x405701a0, 0, (nil), 1) > 000b:trace:ole:CoGetStandardMarshal > ({00000001-0000-0000-c000-000000000046},0x405701a0,0,(nil),1,0x51fd84) > 000b:trace:ole:CoMarshalInterface Using standard marshaling > 000b:trace:ole:CoMarshalInterface Calling IMarshal::MarshalInterace > 000b:trace:ole:StdMarshalImpl_MarshalInterface > (...,{00000001-0000-0000-c000-000000000046},...) > 000b:trace:ole:start_apartment_listener_thread apt->listenertid=0 > 000d:trace:ole:apartment_listener_thread Apartment listener thread starting > on (\\.\pipe\WINE_OLE_StubMgr_0000000a00000000) 000b:trace:ole:CoGetPSClsid > () riid={00000001-0000-0000-c000-000000000046}, pclsid=0x51fd18 > 000b:trace:ole:CoGetPSClsid () Returning > CLSID={00000320-0000-0000-c000-000000000046} > 000b:trace:ole:CoGetClassObject > CLSID: {00000320-0000-0000-c000-000000000046}, > IID: {d5f569d0-593b-101a-b569-08002b2dbf7a} > 000b:trace:ole:COMPOBJ_DLLList_Add > 000b:trace:ole:PSFacBuf_CreateStub > ({00000001-0000-0000-c000-000000000046},0x405701a0,0x51fd10) > 000b:fixme:ole:StdMarshalImpl_MarshalInterface table marshalling > unimplemented 000b:warn:ole:get_stub_manager_from_object Could not map OXID > a00000000 to apartment object 000b:trace:ole:register_ifstub constructing > new stub manager > 000b:warn:ole:register_ifstub Could not map OXID a00000000 to apartment > object in COM_CreateApartment (Yes, that new area of code we are now executing!!!): When we enter that apartment presumably we should be altering some of the MTA's properties like the tid and oxid. Is there a function for it already? Or else, what locks the MTA? -- Bill Medland mailto:[EMAIL PROTECTED] http://webhome.idirect.com/~kbmed