Quoting Antonino Daplas ([EMAIL PROTECTED]):
> Hi,
> 
> I posted before on the possibility of the Overlay layer supporting a
> backbuffer.  I was thinking of possibilities on how to implement
> buffermode change in reallocate_surface(), and how to add support for
> DLBM_BACKVIDEO in allocate_surface().  

I added support for DLBM_BACKVIDEO in allocate_surface() in the CVS version.

> For reallocate_surface(), how about just deallocating the old surface
> and then allocate a new one with the new buffer format? And for
> allocate_surface(), just pass a DSCAPS_FLIPPING to
> dfb_surface_create()?  

The surface shouldn't be destroyed, because IDirectFBSurface for a layer's
surface would then return DFB_DESTROYED and you would have to call
IDirectFBDisplayLayer::GetSurface() again.

However, destroying the old buffers and allocating new ones would be sufficient.

> As for the gfxdrivers' FlipBuffers() implementation, most will probably
> just require a call to dfb_surface_flip_buffers() followed by an update
> of the hardware buffer pointer registers.

That's right.

> I have implemented the changes using this line of thought, and they work
> for me. Still, I'm not too sure if these changes will introduce
> difficulties, or if a more elegant way exists for doing this.  Any
> comments?

See above for difficulties with surface destruction.

-- 
Best regards,
  Denis Oliver Kropp

.------------------------------------------.
| DirectFB - Hardware accelerated graphics |
| http://www.directfb.org/                 |
"------------------------------------------"

                            Convergence GmbH


-- 
Info:  To unsubscribe send a mail to [EMAIL PROTECTED] with 
"unsubscribe directfb-dev" as subject.

Reply via email to