RE: Running the x18 simulator from Xwindows Colorforth?

--- Roman Pavlyuk <[email protected]> wrote:

> Hi Ray,
> 
> One of key differences between native and Windows
> ColorForth memory models
> is in the fact that it's impossible (or I just do
> not know how?) to allocate
> memory below 64K limit under Windows (using
> VirtualAlloc). So if there's
> code that's using absolute addresses, it won't work
> properly. Usually here
> are variables where you can just set base address,
> but I did not play with
> x18, so I cannot tell you any details.
> 
> BR,
> Roman
> 
> P.S. I'm glad to see some activity inthis list :)

Hello Roman.

I actually got this working.  I'm sorry.  I thought
I had reported that to the list.  Basically I change
how "block" works.  Here's the trick.  Magenta
variables actually point to a memory location inside
the block of code where they are defined.  So if
you put a magenta variable at the beginning of a
block, then when you access it the address it returns
is 1 plus the block offset.

Look at the block definition in "block 30".

http://www.dnd.utwente.nl/~tim/colorforth/ether-ip-udp/oninshiko/net.html

I changed the code:

( colors etc )
: block 100 * ;
: white ffffff color ;
.
.
.

To read:

var blk 0 ( colors etc )
: block -30 + 100 * blk + -1 + ;
: white ffffff color ;
.
.
.

After doing this all code that uses "block" works
just fine.  All of the x18 experimentation /
development I've done recently I did using 
Windows ColorForth, including a new/improved
x18 disassembler.

Regards,

John M. Drake

__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com