[PATCH] d3dx9_36: Add stub for D3DXComputeNormalMap.

[Zebediah Figura]

On 06/03/18 15:53, Matteo Bruni wrote:
> 2018-03-04 10:31 GMT-06:00 Zebediah Figura <z.figura12 at gmail.com>:
> 
>> diff --git a/dlls/d3dx9_36/texture.c b/dlls/d3dx9_36/texture.c
>> index b81f3a1..8d8a70e 100644
>> --- a/dlls/d3dx9_36/texture.c
>> +++ b/dlls/d3dx9_36/texture.c
>> @@ -1919,3 +1919,12 @@ HRESULT WINAPI D3DXSaveTextureToFileInMemory(ID3DXBuffer **dst_buffer, D3DXIMAGE
>>
>>      return hr;
>>  }
>> +
>> +HRESULT WINAPI D3DXComputeNormalMap(IDirect3DTexture9 *texture, IDirect3DTexture9 *src_texture,
>> +        const PALETTEENTRY *src_palette, DWORD flags, DWORD channel, FLOAT amplitude)
>> +{
>> +    FIXME("texture %p, src_texture %p, src_palette %p, flags %#x, channel %u, amplitude %f stub.\n",
>> +            texture, src_texture, src_palette, flags, channel, amplitude);
>> +
>> +    return D3D_OK;
>> +}
> 
> I guess this is for bug 38352? Do you know if it actually helps any of
> those games? Is the D3D_OK return necessary?
> 
> I'm a bit uncomfortable with the patch at the moment. Ideally I'd like
> an actual implementation with tests (or even just some tests),
> although that is maybe asking too much?
> 
> 

Looking at the associated bug report, it seems that the stub probably
isn't enough, so yes, this probably shouldn't go upstream in this form.

I only submitted it because it looked like an easy target for
upstreaming, but perhaps removal from staging would be better, as it
fairly clearly isn't adding anything.