Re: wined3d: Implement support for gamma corrected (sRGB) texture loading

[H. Verbeet]

On 27/05/07, Phil Costin <philcostin(a)hotmail.com> wrote:
> Hi,
>
> This patch has been a long time in the making. The following test was used
> during the creation of the patch:
>
> http://ati.amd.com/developer/samples/dx9/sRGBTest.html

If it's not too much trouble, a proper wine test would be nice.

The changes to surface_upload_data and surface_download_data look like
they should be separate changes.

In cubetexture.c, and similar code in texture.c and volume.c:
> +    } else if (GL_SUPPORT(EXT_TEXTURE_SRGB) && This->baseTexture.bindCount > 0) {
> +        currentSrgbMode = This->resource.wineD3DDevice->stateBlock->samplerState[This->baseTexture.sampler][WINED3DSAMP_SRGBTEXTURE];
> +        srgbModeChanged = (This->baseTexture.srgbWhenLoaded != currentSrgbMode);
> +        This->baseTexture.srgbWhenLoaded = currentSrgbMode;
> You've got "device" here, no need to use "This->resource.wineD3DDevice"

sRGBmode looks a bit awkward as a variable name, I'd rather see
something like srgb_mode, but I guess that's more a personal
preference.

It would also probably be better to submit/commit this right after a
release instead of right before one.

wined3d: Implement support for gamma corrected (sRGB) texture loading

[Stefan Dösinger]

Appart of two things Henri already said the patch looks good to me(Use device 
in texture.c, break out the changes to surface_upload_data and 
surface_download data into seperate patches).

As for a wine test, I think its not easy to test it reliably since the opengl 
extension gives implementors quite a lot of room and doesn't require exact 
results. I think all we can reasonably test for are the oddnessnes hl2 
requires, like gamma-corrected Q8W8V8U8 and what the rendering outcome is - 
but that should go into a different patch.