Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!gatech!purdue!haven!uvaarpa!murdoch!usenet
From: randall@Virginia.EDU (Ran Atkinson)
Newsgroups: comp.protocols.tcp-ip
Subject: Re: have you compiled NCSA Telnet w/Turbo C?
Keywords: Turbo C NCSA Telnet
Message-ID: <1991Feb8.192749.1743@murdoch.acc.Virginia.EDU>
Date: 8 Feb 91 19:27:49 GMT
References: <1991Feb07.214707.16739@ecst.csuchico.edu>
Sender: usenet@murdoch.acc.Virginia.EDU
Reply-To: Ran Atkinson <randall@Virginia.EDU>
Followup-To: comp.protocols.tcp-ip
Organization: Computer Networks Laboratory, University of Virginia
Lines: 43

In article <1991Feb07.214707.16739@ecst.csuchico.edu>, 
	madams@ecst.csuchico.edu (Michael E. Adams) writes:

>I won't waste your time detailing the aggravation I
>experienced sorting out the tel23src.zip rats nest of files
>from NCSA.
>
>Still, I would love to hear from anyone who has successfully
>compiled the Telnet programs with Turbo C.

For several months late last spring and summer, I was working with
the NCSA Telnet 2.3 Beta sources under TC/TC++.  The folks at NCSA
should have changed some of the sources by now based on my comments
back then and a fairly recent 2.3 Beta shouldn't require too much
effort to compile with TC.

Back then, the main problems were that the sources were laden with
non-ANSI header files that are unique to MS C.  In many cases, the
header <stdlib.h> should replace one or more MS C headers.  Also,
there were some incorrectly placed #defines that kept TC from seeing
headers that it needed to see and caused it to see headers that it
shouldn't see.  I think that most of the actual C source will be fine,
but the headers and #defines will take some tweaking.  You might
also want to turn off the "no prototype" warning since the order of
declaration of the functions in some of the files causes them to
be defined after all other references to them.

I think that if/when the NCSA folks switch to a newer compiler than
MS C 5.1, it will be easier for them to make NCSA more ANSI conforming
and hence TC/TC++ compatible.  (This might already have happened.)

I'm not actively working on the NCSA sources at the moment due to
other projects, but I think that all would benefit if problems with
getting the NCSA Telnet sources to compile with mainstream (Borland,
Watcom, Microsoft, Zortech, etc.) compilers were reported with suitable
fixes.  I've been impressed with the quality of response from the NCSA
folks to technical problems and questions.

I've posted rather than replied because I think that this might be
of more general interest.  

Ran Atkinson
randall@Virginia.EDU