hedrick@geneva.rutgers.edu (Charles Hedrick) (12/16/88)
Note that two new RFC's have just appeared, for telnet options to propagate terminal speed and to toggle flow control. These are intended to allow telnet to provide all of the features of rlogin. At this point we have a telnet and telnetd based on 4.3's versions that provide all of the features of rlogin except password-less login. (I'll be posting diffs shortly.) I'm holding off on password-less login until we complete our upgrade to SunOS 4.0. I hope that I can do something based on the new public key user validation present in 4.0. Rlogin's implementation of password-less login depends upon the BSD concept of privileged sockets. I don't see any obvious way of including that in a protocol intended for general use. During the review stage, one feature of the flow control RFC appeared somewhat controversial: the proposed line mode telnet option will also have provisions for handling flow control. The current RFC is intended to complete the set of facilities needed for reasonable character mode telnet. Line mode telnet will have to deal with a lot more parameters. Once line mode is negotiated on, I would assume it would handle flow control, echo, and everything else for itself. It would be nice to believe that character mode would simply die upon issuance of the line mode RFC. If you believe that, then it might seem unnecessary to be doing new options for character mode telnet. Unfortunately, line mode telnet is going to require kernel support. Past experience suggests that sites without source will have to wait years before they get support for it on all machines. A reasonable character mode telnet can be implemented with the facilities available in BSD 4.3, and possibly 4.2. (The issue with 4.2 is handling telnet sync. Without OOBINLINE it's not clear whether this can really be made to work. The 4.3 version of telnet tries to work on 4.2 systems. However our experience under SunOS 3.2 -- which uses BSD 4.2 networking -- was not entirely good. I ended up adding OOBINLINE to SunOS 3.2.)
henry@utzoo.uucp (Henry Spencer) (12/17/88)
In article <Dec.15.15.03.23.1988.1873@geneva.rutgers.edu> hedrick@geneva.rutgers.edu (Charles Hedrick) writes: >Note that two new RFC's have just appeared, for telnet options to >propagate terminal speed and to toggle flow control... People not on the Internet, who have to get RFCs by more laborious routes, sure would appreciate it if the RFC numbers were given in such announcements. -- "God willing, we will return." | Henry Spencer at U of Toronto Zoology -Eugene Cernan, the Moon, 1972 | uunet!attcan!utzoo!henry henry@zoo.toronto.edu
hedrick@geneva.rutgers.edu (Charles Hedrick) (12/19/88)
Sorry for not giving the RFC numbers. I forgot that not everybody gets the RFC announcements from NIC. They are 1080 Hedrick, C. Telnet remote flow control option. 1988 November; 4 p. (6688 bytes) 1079 Hedrick, C. Telnet terminal speed option. 1988 December; 3 p. (4942 bytes) I can mail them to you if you can't get them from the NIC.
haas@wasatch.UUCP (Walt Haas) (12/31/88)
In article <Dec.15.15.03.23.1988.1873@geneva.rutgers.edu>, hedrick@geneva.rutgers.edu (Charles Hedrick) writes: > Note that two new RFC's have just appeared, for telnet options... > ...the proposed line mode telnet option will also have provisions for > handling flow control... it would be nice to believe that character mode > would simply die upon issuance of the line mode RFC. The X.25/X.28/X.29 impementation that I wrote for the DEC-20 had a line mode that was intelligent enough to work fine for commands, but switching to character mode was unavoidable when you wanted to use a screen editor. The reasons seem pretty fundamental, so I'd be curious to know how you plan to handle that situation. Cheers -- Walt Haas haas@cs.utah.edu ...!utah-cs!haas