[no subject]

[HIC at MIT-AI (Howard I. Cannon)]

I have implemented and tested the ASCII portion of the
TOPS-20 file job.  The binary stuff should go pretty smoothly
too.

A) I found two bugs in TOPS20:
   1) In MTOPKS, there is a CAILE number of bytes,300 followed by
      an error.  This doesn't let you .MOPKS a packet longer than
      300 bytes.  THIS IS A CROCK!  When I patched it to a reasonable
      length (like 750), the filejob works fine.  The check should 
      really be 4*CHPMXW.

   2) Given that this was wrong, it was supposed to give me an
      error.  Instead, it returned from the MTOPR as if it had
      succeeded, and left FILLCK locked, so that subsequent attempts
      to do ANYTHING to the JFN caused hangage (like, GDSTS, ;J, or
      INFO FILES).  Patching FILLCK to -1 made the hangage go away.
      I don't understand enough about the TOPS20 error system to
      debug this problem, so I merely don't get errors!!

B) ZWEI needs to be hacked fairly grossly.  It knows about ITS
   filenames all over the place.  I've implemented the instance-based
   filename stuff, which all seems to work, and now the editor needs
   to use it.  I would be willing to do the work, but since I didn't
   want to start hacking without MMCM's approval, I haven't done
   more than just poking around.  Since hacking the editor may point
   out deficiencies in the filename scheme, I have also refrained
   from hacking other references to the now obsolete filename
   functions, but will do so after we finish witht he editor.  I'd
   like to sit down with you, Mike, perhaps tonight, and win with
   this stuff (or this afternoon, I don't really understand what
   phase I'm on!).

All in all, I think this stuff will be up by the beginning of term,
as promised.
--Howard