> The conversion packed pixed->bitplan need about 2 Megs of ram
> The conversion address pack->address bitplan take 700Kos
> 
> The tables for conversion are created when the server is lauched
> so you to wait about 5 seconds....

Is there any way to do this algorithmically?  I have 8M of core, and the
fact that 2M of RAM disappears is QUITE evident :-)

Also, about the "top bitplane is 1/2 wide" bug - it's got nothing to do with
bitplanes.  I have written a graphics package in raw XLib, so I know exactly
which calls are being made.  Rectangles rendered with a certain stipple
pattern work fine (in black & white), triangles rendered in solid black
work fine, but lines rendered in solid black are "1/2 width".  So it's
something wrong with certain rendering routines (prob. lines) and not the
bitplane handling itself.


-- 
David Jones, M.A.Sc student, Electronics Group (VLSI), University of Toronto
email: dej@eecg.utoronto.ca, finger for more info/PGP public key

------------------------------------------------------------------------------