2.11BSD/src/ucb/tn3270/README

This is the source directory for the tn3270, and the associated mset, command.

Tn3270, etc., talk to IBM hosts running the VM/CMS operating system,
with the University of Wisconsin developed TCP/IP code (this code is
available at a VERY reasonable price to universities direct from the
University of Wisconsin (lhl@wisconsin.arpa is the contact); commercial
customers need to buy the package from IBM).  We use
an ethernet connection to our 3081 (via IBM's DACU), though the Wisconsin
code also supports a proteon ring, and TCP/IP over X.25.

It would be nice if the other TCP/IP on IBM (the UCLA code for MVS, and
Spartacus code for VM) talked the same 3270-over-telnet protocol.

It would be nice if all "3270 emulators" used the same file format
for describing the mapping of key-strokes to 3270 keys (which file format
might or might not be the one we currently use).

It would be nice if the entire MAP3270/mset/map3270 thing
were generalized in some way.

Anyway...

The makefile is, hopefully, self-explanatory.  The one slightly tricky
point is that there are two .s files which are coded for a vax, and
should be used on a vax version of the program.  These routines are
coded in C in "genbsubs.c" for other architecture machines.

We use curses, and the program is designed to talk over a relatively
slow connection between the CPU and the terminal (say 9600 baud).
On single user systems (say a microVax II) with DMA access to memory,
one would probably want to throw away all this optimization, since it
is very CPU intensive.

If you are using this program on a remote machine (though rlogin) note
that boundary conditions (starting up, ending) may look a bit trimmed; the
pty driver is quite aggressive in flushing output on IOCTLs.

Tn3270 needs the 4.3 version of curses in order to run.  In particular,
there is a curses routine with a name which conflicts with an entry
point in tn3270.c.  This was done on purpose, since we want all references
to use our version rather than the one in curses.  In the 4.3 curses, this
doesn't cause any problems.  In other curses, you may get a multiply
defined reference.

Please let me know of any other bugs.

Greg Minshall
...ucbvax!minshall, or
minshall@berkeley.edu

(415) 642-0530