peter green wrote: > > ok tonight i'll try to code a sanitiseing proxy, mscaht connects to > localhost my program connects on to the irc server then the program will > parse the traffic flowing from the irc server back to mschat and rebuild it > in a form mschat can handle > > > > > > This basically means you end up having to special case this (is there a > > > colon? Then theres no part message. Is there no colon? Then theres a > >part > > > message) which seems to be nothing but extra work. > > > >Huh? You're confused. The RFC says that if there's a :, then all text > >following the colon, even field separators (white space), is part of the > >final argument. In other words, : is a sentinel for the last argument > >in a command. > > > > > And theres exactly the same `issue' for clients rejoining during a split > >- > > > I forget which way around it is but a normal client join and a netsplit > > > join are different - one is : prefixed, the other is not. > > > >*sigh* not too surprising--they're generated by different routines in > >the source. I hope to be able to consolidate these in the future. > >-- > >Kevin L. Mitchell <[EMAIL PROTECTED]> > >
plugwash clearly doesn't understand the RFC, and thinks about it in the same broken way as some client coders. "a b :c" == "a b c" and means: 3 tokens, "a", "b", "c". interpreting it in any different way is wrong. beware