RDROYA01@ULKYVX.BITNET.UUCP (07/12/86)
The addresses in the background module are absolute because on my system
everything in the O/S is absolute. Even the bios is linked in at an
absolute address, a real pain. But I think you can overcome this problem on
a relocatable system with a little work. First some labels
My module CPMLIB Offset from cpm (in CPMLIB)
=========================================================
_ccpmain _main $1f4a
log_dsk _log_dsk $79a2
ssp stack $617e
[this is in the revised sources]
dmaddr [none] wherever your bios stores the current dma
address
Where to put the code. I think it could be linked into the bios in a
relocatable system. You will need to arrange for the caller to pass the
address of an output routine (_outchar in mine) and the address of the
program's base page (*_base in 'C' progs) to the trap #1 handler. It
would then store those addresses until needed. It would resize the tpa so
that h_tpa was $100-$200 less than base, and later load the address of the
supplied output routine into an address register, e.g. into a0, and jsr (a0)
instead of jsr _outchar. This would also allow any program you write to
work although two programs calling trap #1 at once would surely bomb since
it assumes one to be the controller and the other a parasite. This really
was written to be part of the editor.
Another alternative is to use the offsets above to discover the relative
addresses of the presently equated values. This would work if your system
allows a user program to find the location of the O/S. For instance the
CPMLIB _traphnd (trap #2 handler) is $228 above _main. If you know the trap
handler's address or if you know the address of the cpm module, then you can
calculate the absolute addresses needed.
BTW: currently, I am back into xLisp. I'd like to set it up to run in
tandem with the editor so I can add some GNU_MOCK_LISP type functions to my
system. Does anyone have any experience with Mock Liasp and Xlisp who can
point out potential gotcha's? It looks like Mock is closer to Franz than to
Common Lisp. Is this correct?
ATTENTION: R. Heller, I emailed a newer source file that explains some of this
direct to you but got back some errors on the mail attempt. Did you ever
receive? I'll go ahead and post the new sources to the net as they deal
quite well with the disk change, disk logging, and dma buffer problems I had
earlier.
Still hacking away,
Robert Royar
rdroya01@ulkyvx.bitnet