> 
> Jim - 
> 
> I hope you had a good Christmas!

Yes, it was great.  Hope you had a great one also.  Have a happy
New Year!

> 
> 
> I sent you some sample files to try on the Disassembler.
> They are compressed in a file in the /incoming area.

O.K.

> 
> Rename asi.tz to asi.tar.Z

Did that.  Actually I did a tar -xvzf asi.tz.

> 
> There is a README file that contains a description of the files.
> If you have any questions, call me at (813)628-4447 x40.
> 
> Or if I am not there ask for Brian Hammond.  He put the samples together 
> the other day.  
> 
> Thanks for the help.
> 
> Randy Galliher
> 

Randy,

I have put up the file disasm.tgz which is a tared and zipped file of
the disassembler output of the various flavors of object files.  The
files end in .ds and are of three types.  *.ds are your *.o files, *17.ds
are your *.17 files, and *nb.ds are my files from nbcas.  They all look
pretty close.  The file commac is the merged macro library of m.mac65 and
m.mpxmac.  Nbcas does not use the ma2 file.  You can cat them together
by doing "cat m.mac65 m.mpxmac >newmac" and then ddumping the file and
patching the first 0f00 record (EOF) to 0000 and you have a combined
library.  Use that when you use nbcas.

Please look at the output and tell me what you think.  What changes would
you like to see?  I currently do not have access to an Alpha to try this
on, (NASA is closed for the holidays), but if you want to try to compile
it, I have included disasm.c and the makefile to build it.  It may not
work, but what the hell, give it a try.

You said before that unblk did not work.  Is this still the case?  You
use it by redirecting stdin for the file to be unblocked.  I.e. do
"unblk <file >newfile".  It should work.  The input file must be a
binary copy of the mpx blocked (stored or saved) editor file.

Does nbcas run correctly for you?  How about nbcc?  You have not said if
they work.  If these don't work, what is the version of your Alpha O/S.
These were built on the latest Encore machines.  Can't look at the rev
right now.

There are a couple of problems yet to resolve with the disassembler.  Bit
fields are set to 0-7 and the target address is offset by 0-3 bytes.  This
looks strange for your datapool items.  Double word references have an
offset of 2.  Not sure if I can fix these.  I selected a relocation address
of 0x20000 for the dsect before disassembly to make sure there is a difference
for absolute references.  It is easy to change.

I also found a small problem with nbcas, it can not handle more than 9600
symbol references when doing a cross reference.  If this is a problem, turn
off listing or remove the -X from the nbcas command line.  I fixed it (doubled
the value) if you need another version.  I should have one next week.

So,  let me know what you think.

Jim Bevier
jbev@jbsys.com


** James C. Bevier - J B Systems - 10136 Horizon View - Morrison, CO 80465 **
** E-mail: jbev@jbsys.com - Web: www.jbsys.com - Telephone: (303) 697-1038 **
*** Mountain Living at it's Best -&- Still Searching for that pot of Gold ***
