rokicki@navajo.STANFORD.EDU (Tomas Rokicki) (08/08/86)
[ blitter blitter blitter blitter ]
About that code I just posted. Looking in the RKM volume 2, it says
under WaitBlit():
BUGS Because of a bug in agnus, this code may return too soon when
blitter has in fact not started the blit yet, even though
BltSize has been written. This most often occurs in a heavily
loaded system with extended memory, HIRES, and 4 bitplanes.
They are exactly correct; I have found the 4 bitplanes not to be
a requirement, however. I know of no fix to my code that I posted.
So, if you run that LIFE, or use those blit routines, either do not
use extended memory or any other programs which require a lot of
blitter.
I just tried it under 1.2 beta 6; I can't seem to repeat the bug!
It repeats easily under 1.1, but I can't find any mention of the bug
fix in the docs. Perhaps the bug was a side effect of something else.
All around, I like 1.2 beta 6; all of my programs seem to work under
it. You should see the spectacular crashes Archon provides under 1.2
beta 6 with extended memory! I thought only the TRS-80 Coco did that.
It seems to work if I unplug the extended RAM though.
Does anyone know if the WaitBlit() routine has been fixed? It seems
to work for me . . .