Date   : Sun, 27 Oct 1991 13:11:27 GMT
From   : theory.TC.Cornell.EDU!gould@tcgould.tn.cornell.edu (EWD)
Subject: Re: Sayer's UP/M program

In article <k8A4dDy@quack.sac.ca.us> mrapple@quack.sac.ca.us (Nick Sayer)
writes:
>No need to do that. The emulator handles Z-80.
>Now the emulator wasn't fully debugged, so there
>aren't any guarantees. I don't even have the source
>anymore.
>
Had a lot of fun with that upm, hacking for it's own sake.  As I reall
there were amazingly few errors, like xoring swapped with an and in one
of the series, and some incorrect flag states, I forget.  

On some of the fast machines it will give respectable performance, 
especially for disk intensive tasks.

The only way I was able to debug it properly while hacking for all
the speed I could find was to set up test programs on a real Z80 CPM
machine, dump the console output back to unix, then run the same program
on the emulator, and use diff to munch the megabytes.  That sure zeroed
in on some obscure errors in a hurry.

In the end though, all I really wanted was to bring unix tools to the
CPM code, compile, and debug cycle.  To that end all it took was
some utilities to switch the conin/conout from the real console to
the serial port and back, a simple file transfer utility, and then a
rudimentary scripting utility on the unix side.  It's a lot of fun
to type "make" and dog that little puppie. Meanwhile, in another window,
you can work on the next step.  It's like revenge for all those hours
spent plinking a little and waiting for the stinker to be ready for me,
"Mush! Mush! Spin that floppy, compile that code!  No rest for you!
Now mush, mush!"


C976,

  Eliot


Eliot W. Dudley                       edudley@rodan.acs.syr.edu
RD 1, Box 66
Cato, New York   13033                315 626 2878


End of INFO-CPM Digest V91 Issue #182
*************************************