Re: [AD] Fw: Acquire/Release speed issue

[Eric Botcazou <ebotcazou@xxxxxxxxxx>]

> My understanding was that video bitmaps were allocated off the (visible)
> screen until a bitmap can no longer be allocated from the screen. It is
> then allocated from a non-visible surface.

My understanding was that video bitmaps never overlapped the screen, except 
the first one with the same dimensions as the screen.

> I see now that the Windows code disagrees with me.

Yes, the Windows code implements the behaviour I am describing (and has done 
so from the bery beginning AFAIK).  Same for the QNX port (implemented by 
me).  But the ports which expose a unique large screen surface (DOS, BeOS, 
Linux, X11) follow the behaviour you are describing.

So we have a problem.  It can be demonstrated by example/exunicod.c: replace 
create_system_bitmap() with create_video_bitmap() line 157.  The example 
still works under Windows, but not under X11 with DGA2 anymore.

Angelo, what is the situation under MacOS X?  I see the driver implements the 
create_video_bitmap method, so you're likely on the Windows side.  But it 
doesn't seem to reuse the screen for the first video bitmap with the same 
dimensions?  If so, how does it support page flipping?

-- 
Eric Botcazou