[mythtv] Updated patch for directshow filters with MythTV (for CoreAVC)

[Alan Nisota]

Dev Null wrote:
> Using the latest patch 1.4 and CoreAVC 1.3 always getting Seg Fault
> no matter what channel... Please help !!!
>
> The system credentials are:
> x32 processor e6600 opensuse 10.2 x32 myth head revision 13169
>
> avcdecoder 1.3 pro on StarzHD
> ...
> tid:a2dffb90 GetStdHandle(0x1228) => 0x10085c50
> tid:a2dffb90 GetFileType(0x1228) => 0x3 = pipe
> tid:a2dffb90 TlsGetValue(0)
> tid:a2dffb90 TlsGetValue(0)
> tid:a2dffb90 TlsGetValue(1)
> tid:a2dffb90 InitializeCriticalSection(0x8940c54)
> tid:a2dffb90 LockResource(0x20) => 0x20
> tid:a2dffb90 GetLastError() => 0x2
> tid:a2dffb90 TlsGetValue(0)
> tid:a2dffb90 TlsGetValue(1)
> Called unk_RestoreLastError
> Segmentation fault
>   
The problem is that 'RestoreLastError' is not defined, and it returns 
improperly in this case, I believe.  However, the weird thing is that at 
the same point in loading the module, I see:
tid:b6904900 GetFileType(0x1228) => 0x3 = pipe
tid:b6904900 TlsGetValue(0)
tid:b6904900 TlsGetValue(0)
tid:b6904900 TlsGetValue(1)
tid:b6904900 InitializeCriticalSection(0x89fe134)
tid:b6904900 SetHandleCount(0x20) => 1
tid:b6904900 GetLastError() => 0x2

which implies that something is different here (this is just during 
loading the ax file, so the stream itself is irrelevant at this point).  
When did you download 1.3.0?  It is possible they have changed it 
without updating the version number, but I have compared a version I 
recently downloaded to the original that I got a couple weeks ago, and 
they both behave the same way.  You are sure this is the Pro version and 
not Standard?  I have never tried the Standard version before.
If you like, you can send me your CoreAVCDecoder.ax file (directly, not 
through the list), and I can see if I can reproduce using that (thus 
eliminating the question of whether it is the codec or the system that 
is causing the issue).  But whether you choose to do so or not is 
entirely up to you.  Alternatively, I could try to implement 
RestoreLastError, but I'd basically be flying blind trying to resolve 
the issue.