[comp.protocols.misc] Are Bridge terminal servers brain damaged where Telnet, Xoff is concerned

rsharpe@flinders.cs.flinders.oz (Richard Sharpe) (08/23/88)

We are having problems with Bridge Terminal Servers.

1. Do Bridge properly implement the Telnet Protocol?

   I wrote a small c program to place a TCP/IP connection to the Telnet Port
   (ambiguous I know because the SIO ports are also called ports) of a
   specific IP port on one of their boxes (CS200 running TCP/IP 20000), and
   when I saw the SGA option (WILL SGA and DO SGA) I responded that I would
   not and that he should not (ie, DONT SGA and WONT SGA). 

   I then dutifully sent GAs at the end of each line, but never got one back
   from the Telnet server in the Terminal Server. In fact, it seems to me that
   it acted as if SGA was in fact in action. I expected to see a GA back from
   the server if it had no data to send to me.

   Can someone comment on this? Is my understanding of GAs and the SGA option
   at fault?

2. When ports are in Listen state (or their DeVice type is HOST), they ignore
   flow control (Xon/Xoff) if they are not presently connected. This seems to
   be the most Brain Damaged activity I have ever seen, perhaps because I am
   trying to implement a Printer Symbiont on VMS that would need this
   behaviour. 

   Can anyone comment? Is there a way around this problem?

   Just briefly, we want to share the printers among several hosts (VMS and
   Unix), so a print job consists of: Connect to the port via Telnet, ship the
   data down, then disconnect to give some one else a chance. In writing this,
   I found that if you disconnected too early, you loose the last block or so,
   which is why I was looking at the GA route. I was kind of hoping that GAs
   could be used as an indication that all the data I had sent had been given
   to the printer (but not necessarilly printed), so I could send more data,
   or could disconnect with impunity.

   In any case, if there is currently no connection (from a host) to the port,
   and some user walks up and takes the printer off-line, the printer sends an
   Xoff. If a print request starts up at that point, it will be lost (or at
   best, only half of it will print, etc).

Any help appreciated. Replies to me please, and I will summarize to the net.

Regards
Richard Sharpe (rsharpe@flinders.oz)