httpd-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "David Reid" <dr...@jetnet.co.uk>
Subject Re: cvs commit: apache-2.0/src/lib/apr/network_io/unix sockets.c
Date Sun, 19 Nov 2000 01:06:23 GMT
Hmm, have to look at it.  IPv6 has a macro that checks for this case and so
we need to have different logic anyway...  If I don't fix it I'm sure Jeff
will... :)

david

----- Original Message -----
From: "Greg Stein" <gstein@lyra.org>
To: <new-httpd@apache.org>
Sent: Sunday, November 19, 2000 1:05 AM
Subject: Re: cvs commit: apache-2.0/src/lib/apr/network_io/unix sockets.c


> On Sat, Nov 18, 2000 at 03:32:16PM -0000, dreid@locus.apache.org wrote:
> > dreid       00/11/18 07:32:16
> >
> >   Modified:    src/lib/apr/network_io/unix sockets.c
> >   Log:
> >   Hide an instance of AF_INET6 if we don't have IPv6 and start adding
> >   code to allow BeOS to use this file.
> >...
> >        }
> >   @@ -246,6 +250,7 @@
> >                sock->local_port_unknown = 1;
> >            }
> >            /* XXX IPv6 to be handled better later... */
> >   +#if APR_HAVE_IPV6
> >            if (sock->local_addr->sa.sin.sin_family == AF_INET6 ||
> >                sock->local_addr->sa.sin.sin_addr.s_addr == 0) {
> >                /* not bound to specific local interface; connect() had
to assign
>
> Um. Does the #if accidentally prevent the second part of that condition
from
> executing?
>
> For example, what if the family was IPv4 and s_addr == 0. ??
>
> [ I don't see the surrounding code or know the detailed semantics (thus
I'm
>   deferring the analysis to you), but the above change looks fishy ]
>
>
> Cheers,
> -g
>
> --
> Greg Stein, http://www.lyra.org/
>


Mime
View raw message