SUN 4 SPARC with X11/NeWS, TELNET Problems.

[STURTEVANT at ccc.nmfecc.gov]

We are having a problem with SUN X11/NeWS on our SUN 4 SPARC stations.
When we TELNET to our MULTINET VAX/VMS system with NO SCROLL BAR on our
SUN X11/NeWS window, everything works fine.  However, when we enable the
SCROLL BAR on our SUN X11/NeWS window and then TELNET to our MULTINET
VAX/VMS system, pressing return after entering our username causes the
input to be erased. This also happens when we TELNET to other VMS systems
that use a TCP/IP package other than Multinet. 

This does not happen when we TELNET, with the SCROLL BAR enabled, to
either another SUN system or a VAX/ULTRIX system.  In addition I have
heard that this does not happen with MIT's X11 software with the scroll
bar enabled. 

Using our ethernet packet analyzer, I was able to determine:

1.  SUN TELNET sends out a hex "0d 00" (CR NULL) each time the RETURN key
is pressed, with the SCROLL BAR disabled.

2.  SUN TELNET sends out a hex "0a" (LF) each time the RETURN key is
pressed, with the SCROLL BAR enabled. 

Additionally I was able to determine that the TELNET commands, which are
imbedded in the TELNET data stream, were completely identical with the
SCROLL BAR enabled or disabled.  I verified this for each system (MULTINET
VAX/VMS, VAX/ULTRIX and SunOS).  This implies that this problem has
nothing to do with automatic TELNET options negotiation.  The SUN TELNET
seems to unconditionally send CR NULL with the SCROLL BAR disabled, and
send just LF with the SCROLL BAR enabled.  Thus, this appears to be a
problem only with the SUN X11/News TELNET software.  The TELNET RFC (RFC
854) says that you should send CR NULL or CR LF, but there is no provision
for just LF as is done in the SUN X11/NeWS windows with the SCROLL BAR
enabled. 

By the way, the reason this is a problem for VAX/VMS systems is that the
LF character informs the VMS terminal driver to delete-word-left (a line
editing function).  This causes the user's username to be erased when
pressing the return key. 

Unless I hear otherwise and/or that what I am observing is legal, then I
will post this as a bug to SUN.  I welcome all comments and input
regarding this matter.  I also wish to thank all of the people on the
info-multinet mailing list who had responded to Bob Aiken's prior note on
this subject. 

Allen Sturtevant
National Magnetic Fusion Energy Computer Center
Lawrence Livermore National Laboratory
sturtevant at ccc.nmfecc.gov