jv@mhres.mh.nl (Johan Vromans) (02/01/89)
Sorry if this has been asked before, but ... I'm using several Unix systems, and access them via TCP/IP. I use both a Macintosh with NCSA Telnet and a terminal connected to a Bridge CS200 terminal server. My problem is as follows. When I run an application on the Unix machines which switches off CR to LF mapping on input (e.g. stty -icrnl), some of the systems behave correct, but others don't. The Telnet protocol, as defined in RFC854, specifies that the sequence CR-LF is to be treated as a single newline, and that CR-NUL must be used where a CR alone is desired. NCSA Telnet lets you specify this, but the CS200 does not. I did a simple experiment. In the columns, CR-NUL means: NCSA Telnet sends CR-NUL when the CR key is hit, CR-LF means: NCSA Telnet sending CR-LF, and CS200 indicates access via the Bridge CS200 terminal server. "ok" means the application can tell which key (CR or LF) was hit, "-" means: it can not (it only sees a LF). System CR-NUL CR-LF CS200 ------------------------------------- 1. Ultrix 3.0 ok ok ok 2. UTX 2.0 ok - ok 3. HPUX 5.3 ok ok ok 4. HPUX 6.2 ok - - ------------------------------------- I can understand that CR-NUL works ok for all systems. I can understand that systems 2 and 4 adhere to RFC854 and that an application cannot switch off CR to LF mapping because the telnet protocol has already translated it. I don't understand why systems 1 and 3 allow the application to do so, in spite of RFC854. And I cannot understand why systems 1, 2 and 3 can be used via the CS200, but not system 4. Can anyone shed some light into this darkness? Please reply by E-mail, I'll summarize if relevant. Thanks, Johan -- Johan Vromans jv@mh.nl via european backbone (mcvax) Multihouse [A-Za-z ]* [NB]V uucp: ..!mcvax!mh.nl!jv Gouda - The Netherlands phone: +31 1820 62944