> Huh. I was sure I'd already reviewed this.
>
> It's conservative - compatible() will return false in lots of cases where
> DisplayBuffers would not need to be invalidated - but so are all the real
> hardware implementations.
yes thats the motivation.. I think we could reduce the invalidating cases in the hardware implementations, i.e. we could make the position of the db not be part of the db - but just an information the compositors get from the configuration. Then we can come back here and maybe reiterate on the behavior of fake display.
> Huh. I was sure I'd already reviewed this.
>
> It's conservative - compatible() will return false in lots of cases where
> DisplayBuffers would not need to be invalidated - but so are all the real
> hardware implementations.
yes thats the motivation.. I think we could reduce the invalidating cases in the hardware implementations, i.e. we could make the position of the db not be part of the db - but just an information the compositors get from the configuration. Then we can come back here and maybe reiterate on the behavior of fake display.