You are viewing a plain text version of this content. The canonical link for it is here.
Posted to apache-bugdb@apache.org by Marc Slemko <ma...@znep.com> on 1997/07/21 08:20:02 UTC

config/620

The following reply was made to PR config/620; it has been noted by GNATS.

From: Marc Slemko <ma...@znep.com>
To: apbugs@apache.org
Subject: config/620
Date: Mon, 21 Jul 1997 00:11:31 -0600 (MDT)

 
 
 
 ---------- Forwarded message ----------
 Date: Sun, 25 May 1997 17:49:53 -0400 (EDT)
 From: Christopher Neitzert <ch...@razorfish.com>
 To: marc@hyperreal.org
 Cc: apache-bugdb@apache.org, marc@apache.org
 Subject: Re: Changed information for PR config/620
 
 On Sun, 25 May 1997 marc@hyperreal.org wrote:
 
 > Synopsis: FIX TO BUG #605
 > State-Changed-From-To: open-feedback
 > State-Changed-By: marc
 > State-Changed-When: Sun May 25 14:34:23 PDT 1997
 > State-Changed-Why:
 > That does not make any sense.  SOCK_STREAM is not a 
 > variable, but a constant defined in a system header
 > file.  On Solaris, it should be in /usr/include/sys/socket.h:
 > 
 > #define NC_TPI_COTS     2               /* must agree with netconfig.h */
 > [...]
 > #define SOCK_STREAM     NC_TPI_COTS     /* stream socket */
 > 
 > What do your include files show?  Unless you can say that
 > Apache is doing something, the problem appears to be in
 > your include files.  Apache does nothing with the
 > SOCK_STREAM define other than use it twice.
 from http_main.c
 --snip--
     if ((s = socket(AF_INET,SOCK_STREAM,IPPROTO_TCP)) == 2) {
         log_unixerr("socket", NULL, "Failed to get a socket, exiting
 child",
                     server_conf);
         exit(1);
     }
 --snip--
 it was set to -1, I changed it, it compiled, I havent had a chance to
 stress test it yet though.
 
 >From /usr/include/sys/netconfig.h
 
 --snip--
 #define NC_TPI_COTS     2
 --snip--
 
 
 I see what you are saying, and then it must be with _my_ include files
 though nearly every solaris platform I have access to would not
 run the binary due to this error.  I am not sure if i've abnormal
 solaris boxen or this is common.  The Apache would run untill
 it made the socket call then exit with an error of no permission
 to bind the socket.  --maybe a small paragraph in the faqs
 would save others from pulling hair like i was(for all of ten minutes).
 
 cheap the faith!
 
 christopher neitzert
 
 --
 christopher neitzert
         [ systems administrator ]
                                         107 grand street   | 3rd floor
                                 new york   | ny 10013
        w w w . r a z o r f i s h . c o m