> Alright, I think we've gotten to the bottom of this in the email chain.
>
> We really need a shim EGL platform to adapt EGL_KHR_mir_platform to
> EGL_KHR_android_platform. This adaptation should be done in the shim, not
> within the android client platform.
>
> I'd prefer a shim-layer to be proposed instead of this, just on the basis of
> it having 2-step initialization, deferred errors from ErrorDriverInterpreter
> and introducing temporary confusion about MirRenderSurface (which sparked all
> that email discussion)
I'd rather tackle the shim idea after we have MirRenderSurface merged in, since it is internal refactoring, and only affects Android. Whereas, this is blocking Mesa upstreaming.
> Alright, I think we've gotten to the bottom of this in the email chain. mir_platform to android_ platform. This adaptation should be done in the shim, not rpreter
>
> We really need a shim EGL platform to adapt EGL_KHR_
> EGL_KHR_
> within the android client platform.
>
> I'd prefer a shim-layer to be proposed instead of this, just on the basis of
> it having 2-step initialization, deferred errors from ErrorDriverInte
> and introducing temporary confusion about MirRenderSurface (which sparked all
> that email discussion)
I'd rather tackle the shim idea after we have MirRenderSurface merged in, since it is internal refactoring, and only affects Android. Whereas, this is blocking Mesa upstreaming.