[AD] Allegro 3.9.22

[Shawn Hargreaves <shawn@xxxxxxxxxx>]

... is on http://www.sunsite.auc.dk/allegro/wip.html.


Stefan: there is still something wrong with screen sub-bitmaps under 
Windows. The new exswitch example bombs out, and I've traced this to 
whenever it tries to call rectfill() or blit() on a screen sub-bitmap: the 
calls to textout() and putpixel() work fine. Looks like just the hardware 
accelerated routines are going wrong here.


Windows people: the MSVC asm might or might not work better now. Let me know 
how you get on, and fingers crossed!


Everybody: the GFX_SAFE system is much more flexible now. I think perhaps 
too flexible: I'm not sure whether I like it. The current implementation is:

   First try to autodetect for the exact mode requested.

   If that fails, if GFX_SAFE_ID, GFX_SAFE_W, etc, are defined, use those. 
   Otherwise try to autodetect for a 0,0 resolution.

   Various Linux drivers (VGA and fbcon) know how to handle 0,0 size, and 
   will return whatever screen mode seems most sensible to them.

   If the graphics driver returns a color depth other than the one 
   requested, graphics.c contains some magic bank switch hooks to make it 
   work in a sensible way.

This will give you a working safe mode pretty much whatever, but I have 
major reservations about the automatic color depth emulation. That's a 
useful thing if you have a program expecting an 8 bit mode and want to run 
it with a vesafb device that is stuck in truecolor, but the performance is 
so bad, I'm worried that this is not a good thing to do automatically.