Xref: utzoo comp.os.os2.programmer:552 comp.protocols.tcp-ip:15385
Path: utzoo!utgpu!news-server.csri.toronto.edu!cs.utexas.edu!sdd.hp.com!wuarchive!mit-eddie!uw-beaver!milton!yoda.eecs.wsu.edu!wbonner
From: wbonner@eecs.wsu.edu (Wim Bonner)
Newsgroups: comp.os.os2.programmer,comp.protocols.tcp-ip
Subject: Re: How do I detect an active TCP/IP (IBM) under OS/2?
Message-ID: <1991Mar24.091547.9174@eecs.wsu.edu>
Date: 24 Mar 91 09:15:47 GMT
References: <8281@umd5.umd.edu>
Reply-To: wbonner@yoda.UUCP (Wim Bonner)
Organization: Washington State University, Pullman
Lines: 23

In article <8281@umd5.umd.edu> dzoey@terminus.umd.edu (Joe Herman) writes:
>from the socket() (or sock_init()) call.  However, if inet is not
>functioning the call to socket() causes the program to exit.  Since
>the program can still function without TCP/IP, this is not a desired
>behavior.

Is the socket call in a DLL?  If so you may be running into a similar problem
I've had with my network programming in that all of the LANMAN calls are 
actually located in DLLs, and so the first call kills the program.  The solution
that was suggested to me, was to try to demand load the call and if that was 
successful, assume the network was loaded and make the call.  If you cannot 
demand load the call, the DLLs are not available, so don't make any network 
calls.

Unfortunately, I'm not at my own computer right now, so I can't post the 
suggested code.  I'll try to remember to put it up tomorrow when I can get to my
computer.

Wim.
-- 
|  wbonner@yoda.eecs.wsu.edu  |
| 27313853@wsuvm1.csc.wsu.edu |
|  72561.3135@CompuServe.com  |