🚀 go-pugleaf

For some reason, RTS remains asserted after I close the tty.  I fopen the file,
apply ioctl to it, getc for a while, restore the original ioctl state, fclose
and exit.  RTS was inactive before I ran the code, went active for the run of
the program, and remained active after the program terminated.  I've done this
on three different ports, with consistent results.  HUPCL is set in both the
original (system-generated) and application ioctl blocks.  Subsequent runs of
the program show the entire ioctl block to be the same as before the first run.
The device is opened for reading only (it originally has been coded for read and
write).  Any ideas?

Oh, I just discovered that doing a ^C on "cat /dev/tty__"  produces the same 
effect.  The equipment on the other end is wired DCE, with HPSM lines 5, 6 and
8 (CTS, DSR and DCD) ganged to the DCE's DSR (line 6).  The equipment's DSR
is inactive at the time.  Pulling the equipment end of the cable makes no
difference.  There are no loopback's on the Masscomp end.  The only other lines
in use are RD, TD, SigGnd, RTS (obviously) and DTR, which each do straight to
the corresponding (same-numbered) pins on the other end of the cable.

				-Jim Gonzalez	gonzalez@bbn.com
				BBN Labs, Inc.	(acoustics, not communications)
				Cambridge, MA