On Wed, 7 Feb 1996, Naved A Surve wrote: > In message <Pine.SOL.3.91.960207203531.3897B-100000@clark.net> you said: > > If you look deep into the socket-handling routines in circle, you may > > find an input buffer. If you can find this (I'm not sure if it exists in > > circle, I'm just guessing it works this way), and learn how it works, you > > can check the buffer before it gets passed to the command interpreter, > > and just check for the oldest character (the one that was entered first > > since the last buffer purge). > > That won't work. You'd essentially be writting another command interpreter > that gets called before the main command interpreter with no change in the > way it performs. The user still has to hit <cr> on his end to send his > buffer across the network. AFAIK the only way you can handle this situation > is by having the user run a specific client program that handles raw output. Ok if that's true then my basic assumption is wrong, but I think you may be wrong about the user's client not sending one character at a time. The telnet I use (linux stock telnet command) does send one character at a time. If this weren't the case, I wouldn't be able to use pine, or any other hotkey-driven program, over a telnet connection. So, if the client supports character transfers, it must be the server that doesn't, and since pine, lynx, and various other servers support character transfers, it should be possible to do the same with circle, although depending on how it works, you may miss some characters when polling other sockets. I'd say it's possible, but not necessarily worth the trouble. If you get it working I know there are lots of circle imps who would love to see it done. Sam
This archive was generated by hypermail 2b30 : 12/07/00 PST