MH> Is there any way I can talk you into posting that finger server ( or a
 MH> working varient ) so i can try and figure out what I'm doing stupid. 
 
    The source itself is about 13K long, so I'm not sure if it'd be a good
idea to post it. It is on the hobbes cd-rom, and accordingly available by
FTP from the associated sites. I think I found it as fingerd2.zip in
32bit/network.
    Let me know if you can't find it via the above sources. I'll try
trimming it down and posting its skeleton.
    Conversely, any chance of seeing your talk client? I've been using
talk on my Unix host rather than TCP/IP 2.0's talk.exe since the latter
works on a line-by-line mode which isn't too nice towards whoever I talk
to. I've been meaning to hunt down source for ytalk, but the days in
which I'd be able to use it are becoming numbered. :( Actually, email me at
tee{at}ecf.toronto.edu - I can send you the finger daemon from there.
 
    Anyway, the socket-related code in the finger daemon looks virtually
identical to what I've seen under Unix, aside from using recv() rather than
read() and having to call sock_init() before everything. If you can get a
client up and working, I can't imagine what might be going wrong for you.
 
 MH> Are you using emx?
 
   Nope. CSET++ - perhaps the libraries you're using under EMX are at
fault? I'm using the TCP/IP 2.0 Programmer's Toolkit.

--- Maximus/2 2.02p1