I am using CALLBACKS to communicate between a VB App and an app Written with VC6.0.

Here is the problem: When the (English) App was written, it was not coded using unicode.
Well the obvious solution would be to set _UNICODE and go about my life. Well this
would also cause about 3E09 errors and I would have to sack a small village to get
the budget or the time to do this. :)

This works, but what I want to know is why!

Why does memcopy copy the data from the BSTR into ProcLine, which is a TCHAR correctly?

BSTR = *OLECHAR = *WCHAR

#ifdef _UNICODE
TCHAR is a WCHAR
#elif defined (_MCBS)
TCHAR is a char
#YADA YADA

//shared data
TCHAR ProcLine[256];
///...later in the DLL
int _declspec (dllexport) CALLBACK PutProcLine(BSTR line)
{
memset(ProcLine,0,sizeof(ProcLine));
memcpy(ProcLine,line,sizeof(ProcLine));
procdtevent.Lock();
return 0;//success
}

int _declspec (dllexport) CALLBACK GetProcLine(BSTR *line)
{
_bstr_t tmp = *line;
TCHAR tline[288];
_stprintf(tline,"%s",ProcLine);
_bstr_t ret = tline;
*line = ret;
procdtevent.SetEvent();
return 0;
}

I mean ProcLine would be riddled with NULLs unless memcopy castes each short in the
BSTR to byte and life is ok because of little endian.