> Not necessarily. The correct framebuffer needs to be bound for

> subsequent draws, but that may very well be a different framebuffer

> than the one currently bound. Invalidating STATE_FRAMEBUFFER ensures

> that context_apply_draw_state() will bind the correct framebuffer even

> if the bound render targets didn't change since the previous draw.

> No, the invalidated state is going to be applied again by wined3d

> "state manager", if needed.

Thanks for explaining, that's a clever setup.

I'll send an updated patch.

Regards,

Fabian Maurer