[Coco] My first CoCo device (virtual)

[Gene Heskett]

On Wednesday 29 May 2019 09:04:03 pm Stephen Fischer wrote:

> I used a program to split one of the standard libraries into separate
> files.
>
> Then I was able to change the first address the "core" file was
> missing to enable the loader maps to match the "core" dumps.
>
> I used the file list driven merge program to put it all back together.
>
> OS9Gen is that merge program. It does do one other task, write the
> file the DECB "DOS" command reads to start the boot process. I make
> sure that part was done before doing the library put together task.
> Using floppies of course.
>
> I am unsure if the library split program which came from Delphi is
> available now.
>
> http://cococoding.com/sfischer/SHF%20OS-9%20Disks/Z/SHF79.DSK
>
> May have that program, I do not remember the name after so many
> decades.
>
If that one can't be found, there is another that I wrote after being 
challenged by Paul Jerkatis.

"vfy -s os9boot" will do that for you.  And if you've run kernal2dir, you 
also have a directory entry for track, and a vfy -sk kernal 
(miss-spelling intentional) will split that up to its 5 components, the 
rsdos ID header of 6 bytes, the 3 files that make up the working code, 
and the trailer bytes that make it exactly 4608 bytes long, a full 
tracks worth.  That has 100% error catchers because all its other 
functions actually work on the file named insitu. Get a help screen by 
not giving it an argument, or a -character that is not valid.

> Anybody besides me ever look at the loader maps?

Yes, and I will say they are best described as messy and incomplete. 
_Good_ docs and good defs do not exist for some of that. That will have 
to be improved before we can make level 3 work.

It took a week of staring at the includes/defs for the rs232 pack to 
realize that what I was fighting with in doing a mouse driver for a std 
seriel pc rollerball mouse of the day was an off by one in the defs. 
Thats been fixed for close to 20 years now. That mouse driver 
autoconfigs itself to work with both M$ mice or logitech by looking at 
the first packet that includes an mmb push.  Up to then they are 
identical so its a don't care.

> SHF
>
> On 5/28/2019 10:26 PM, Walter Zambotti wrote:
> > I have been playing around with writing a virtual device module for
> > OVCC.
> >
> >
> >
> > It provides a FPU and a GPU.
> >
> >
> >
> > The fpu has all the usual functions and is performing well. The FPU
> > uses DMA to fetch values and write results directly into CoCo
> > memory.
> >
> >
> >
> > The GPU is still in its infancy but it can draw pixels and lines and
> > it also uses DMA to draw directly into CoCo memory.
> >
> >
> >
> > I now need to replace cfloats.a in clib.l with a version that can
> > take advantage of the FPU.
> >
> >
> >
> > I suppose I would need to change the graphics driver and libraries
> > to take advantage of the GPU.
> >
> >
> >
> > Walter
> >
> >
> >
> > Ps OVCC on github is now public so everyone can see the code.
> > WallyZambotti/OVCC


Cheers, Gene Heskett
-- 
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
Genes Web page <http://geneslinuxbox.net:6309/gene>