[PATCH 03/10] wined3d: Use GL_DEPTH_COMPONENT16 for 16 bit depth texture formats.

[Matteo Bruni]

2017-12-06 11:00 GMT+01:00 Józef Kucia <jkucia at codeweavers.com>:
> From: Matteo Bruni <mbruni at codeweavers.com>
>
> Signed-off-by: Józef Kucia <jkucia at codeweavers.com>

> -    {WINED3DFMT_D16_UNORM,              GL_DEPTH_COMPONENT24_ARB,         GL_DEPTH_COMPONENT24_ARB,               0,
> +    {WINED3DFMT_D16_UNORM,              GL_DEPTH_COMPONENT16,             GL_DEPTH_COMPONENT16,                   0,

FTR this has always been "wrong" as far as git history goes. I suspect
this was intentional and that the original reason is moot at this
point.
I can think of two possible motivations: working around depth buffer
precision issues, probably due to the different depth range in d3d vs
OpenGL, and avoiding color / depth buffer bit depth mismatch which was
not supported in some old Nvidia GPUs. The former should be fixed by
ARB_clip_control and the latter is something that d3d applications
also have to take care of, which means that we shouldn't need to do
anything special.

Anyway:
Signed-off-by: Matteo Bruni <mbruni at codeweavers.com>