Re: [AD] GGI patch for 3.9.21

[Jason Wilkins <fenix@xxxxxxxxxx>]

> Since Allegro now has functions to control switching behavior, it'd be
> best if those functions did the same thing with all targets.  What
> worked for the GGI people with svgalib should work for us.  The hunter
> becomes the hunted :)

I don't know if its possible to make it work exactly the same everywhere.
On the BeOS, I do not think that it is possible to tell the OS that you
can not switch.  So, I hope that no one decides to write internal Allegro
code that assumes that switching can be turned off while it executes,
because it will crash on the BeOS.

I think I can block a switch for about 2 seconds, but after that something
decides that my program should cause a page fault.  So, I guess some code
can be written that can be guaranteed not to be screwed during a page
fault, but it must execute quickly, and never hold the screen for an
indefinite period of time.  On the BeOS, it would just cause an attempt to
switch to be blocked until the OS decided the program needs to die (which 
is about 5 seconds).  I think this is because some other threads get antsy
and start doing things they shouldn't when a switch is happening ^_^

> -- 
> tjaden@xxxxxxxxxx - http://psynet.net/tjaden/ - ICQ: 4588779

--
         The Phoenix - President of The Artistic Intuition Company
    Caelius * Zen-X * Mirror Reflex * Runica * X-Domain * Infinite Realms
                          http://www.io.com/~fenix