Date: Mon, 20 Apr 92 21:38:48 GMT
From: kuku AT acds DOT physik DOT rwth-aachen DOT de (Christoph Kukulies)
To: djgpp AT sun DOT soe DOT clarkson DOT edu
Subject: blits
Status: O

Although 1.06 is released I still have a 1.05 related question 
(havn't had the time to install this big chunk and as I'm presently
a little bit under pressure I'm cautious and don't want to run 
unneccessarily into unwanted trouble):

I have a hand rolled blit routine which memsets bits in 0xd0000000 VRAM.
Presently I see strange things happen. (bitmaps are sometimes garbled
and I wonder if it's a 1.05 problem, a general problem, a problem induced
by myself - the latter is the most probable one :-), anyway, I'm
getting a segmentation violation in pointer 0x004a8000 at 40:8e2c.

DJ, does this ring any bells? I mean, the pointer 004a8000 looks so special.
Somehow I have in mind that I read across a comment that a memory access
into mapped VRAM could cause a off screen memory reference. I'm not
sure whether this was in the DJGPP programs - it could also have been else-
where.

But if there are known problems in that area in 1.05 it could be a reason
for me to switch as fast as possible to 1.06.

--Chris