ID3DXEffect::SetValue

[Luis Carlos Busquets Pérez]

The current implementation includes a check on param->bytes >= bytes

specifically:

if (data && param->bytes >= bytes)
{

...

}

So if bytes<param->bytes returns D3DERR_INVALIDCALL.

However, the MS version does not act that way: If the parameter is 
shorter than the length of the data, then the data that fits in the 
parameter is put and returns D3D_OK

Therefore, the check of the size should be eliminated and in memcpy 
instead of referring to bytes, it should refer to min(bytes, param->bytes)

Luis

[Austin English]

On Mon, Feb 6, 2012 at 09:18, Luis Carlos Busquets Pérez
 wrote:
> The current implementation includes a check on param->bytes >= bytes
>
> specifically:
>
> if (data && param->bytes >= bytes)
> {
>
> ...
>
> }
>
> So if bytes<param->bytes returns D3DERR_INVALIDCALL.
>
> However, the MS version does not act that way: If the parameter is shorter
> than the length of the data, then the data that fits in the parameter is put
> and returns D3D_OK
>
> Therefore, the check of the size should be eliminated and in memcpy instead
> of referring to bytes, it should refer to min(bytes, param->bytes)
>
> Luis

Please send a patch and/or file a bug at http://bugs.winehq.org. The
same applies to your other report.

Thanks!
Austin

[Rico Schüller]

Am 06.02.2012 18:18, schrieb Luis Carlos Busquets Pérez:
> The current implementation includes a check on param->bytes>= bytes
>
> specifically:
>
> if (data&&  param->bytes>= bytes)
> {
>
> ...
>
> }
>
> So if bytes<param->bytes returns D3DERR_INVALIDCALL.
>
> However, the MS version does not act that way: If the parameter is
> shorter than the length of the data, then the data that fits in the
> parameter is put and returns D3D_OK
>
> Therefore, the check of the size should be eliminated and in memcpy
> instead of referring to bytes, it should refer to min(bytes, param->bytes)

Yes, that looks partly correct, but the check needs to be in there and 
should be "if (data && param->bytes <= bytes)" and the copied size 
should be param->bytes. It doesn't show up, because there isn't a test 
in the test suite for this case, yet. So the trace was correct 
(http://www.winehq.org/pipermail/wine-devel/2011-December/093489.html) 
but not the if and the memcopy. :-) Well, I've overseen this. Feel free 
to send a patch, otherwise I'll do that tomorrow.

Cheers
Rico