Re: [PATCH 6/6] Due to code rewrites is_color_fixup_supported is not needed anymore in RealizePalette. The code in question should be executed in all cases. This also fixes palette refresh issues when shaders / paletted textures aren't around.

[Roderick Colenbrander]

On Mon, Apr 5, 2010 at 9:52 PM, Stefan Dösinger  wrote:
> Am 05.04.2010 um 20:05 schrieb Roderick Colenbrander:
> > - IWineD3DDeviceImpl *device = This->resource.device;
> > - if((This->resource.usage & WINED3DUSAGE_RENDERTARGET) &&
> > - device->blitter->color_fixup_supported(&device->adapter->gl_info, This->resource.format_desc->color_fixup))
> > - if(This->resource.usage & WINED3DUSAGE_RENDERTARGET)
>
> It's a problem with the existing code, but checking the usage here doesn't make much sense. It should check if the surface is a front buffer on any swapchain(or, as long as we store a P8 surface as ARGB in GL this code also applies to backbuffers)

There are some other places which do the same (both P8 and other
code). It might be nice to rewrite it e.g. using the
p8_primary_render_target call or so perhaps something more generic. In
case of DDraw (without P8 textures) we seem to set
WINED3DUSAGE_RENDERTARGET only for the primary surface on the ddraw
side though.

Would you mind if I rewrote this part at a later stage? It isn't that
critical right now. I mainly rewrote it to get
is_color_fixup_supported out of the way and it also fixed bugs. If you
want I could add a 'TODO' comment, so that it won't be forgotten.

Roderick