[VDPAU] Interpretation of the VdpOutputSurfaceRender{Bitmap, Output}Surface spec

[Stephen Warren]

Emil Velikov wrote at Tuesday, August 26, 2014 11:31 AM:
> Hello list,
> 
> Having a closer look at the specification [1] of the two functions one can
> observe the following
> 
>   "...If source_surface is NULL..."
> yet it seems that the type of source_surface is a 32 bit unsigned:
> 
>   typedef uint32_t VdpBitmapSurface
>   typedef uint32_t VdpOutputSurface
> 
> Which begs the question - by NULL do the spec actually mean 0 or
> VDP_INVALID_HANDLE (0xffffffffU) ?

The spec should say VDP_INVALID_HANDLE not NULL.

> I'm suspecting that either interpretation
> might cause buggy behaviour in some (any?) programs that depends on the
> particular functionality.

Given that the NVDIA implementation almost certainly is coded to actually
implement a check for VDP_INVALID_HANDLE not a check for NULL, I hope that
applications are already passing the expected value here, or they'd be
experiencing unexpected behavior here, or the API is simply returning an
error.

-- 
nvpublic