From: Brian S. Julin <bri@tull.umassp.edu>
  To  : Tristan Wibberley <bloater@ps.cus.umist.ac.uk>
  Date: Fri, 13 Aug 1999 15:56:58 -0400 (EDT)

Re: GGI and XFree86 and stuff

On Fri, 13 Aug 1999, Tristan Wibberley wrote:
> Can't you let te programmer decide when he/she needs to wait for the
> flush, and give them a function to wait for the flush to complete. This
> way, if the programmer knows that the buffered accels will not disturb
> his/her fb access, they can do both at the same time? Is this too much
> work for too little gain?

If simultaneous buffer access is possible, there will be just an advisory 
lock in place; either the user or the sublib are free in that case to
overlap operations.  If the sublib micromanages it more than is preferred,
it should be easy enough to implement a flag to hand responsibility for
managing the accesses over to the user.  There is already a flush call
in the GGI API IIRC.

--
Brian