Mailing-List: contact dev-help@tomcat.apache.org; run by ezmlm
Precedence: bulk
Reply-To: "Tomcat Developers List" <dev@tomcat.apache.org>
Received-SPF: pass (asf.osuosl.org: local policy)
From: Jim Jagielski <jim@jaguNET.com>
Message-Id: <200601301359.k0UDx7402956@devsys.jaguNET.com>
Subject: Re: AprEndpoint and IPv6
To: dev@tomcat.apache.org
Date: Mon, 30 Jan 2006 08:59:03 -0500 (EST)
Reply-To: jim@jaguNET.com
In-Reply-To: <200601301247.54889.mailing-tomcat-dev@schoenhaber.de> from
 "Markus =?iso-8859-1?q?Sch=F6nhaber?=" at Jan 30, 2006 12:47:54 PM
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit

Markus =?iso-8859-1?q?Sch=F6nhaber?= wrote:
> 
> Jean-frederic Clere wrote:
> > Jean-frederic Clere wrote:
> > > Jim Jagielski wrote:
> > >> There's a bug report (37788) regarding allowing AprEndpoint
> > >> to use APR IPv6 addresses. Their patch is almost right, but
> > >> instead the value should be to use APR_UNSPEC instead
> > >> of APR_INET6 (or the current APR_INET) to allow APR
> > >> to correctly determine IP version and do a graceful
> > >> recovery... As well as handle cases where APR wasn't
> > >> built with IPv6 support.
> > >>
> > >> I'd like to commit that change... comments?
> > >
> > > -1: It also cores on my machine.
> > > +++
> > > #
> > > # An unexpected error has been detected by HotSpot Virtual Machine:
> > > #
> > > #  SIGSEGV (0xb) at pc=0xa808eb5e, pid=10005, tid=3084708160
> > > #
> > > # Java VM: Java HotSpot(TM) Client VM (1.5.0_06-b05 mixed mode)
> > > # Problematic frame:
> > > # C  [libapr-1.so.0+0x19b5e]  apr_socket_bind+0x2e
> > > #
> > > +++
> > >
> > > APR_INET6 works OK with the same configuration.
> >
> > With the following patch:
> > +++
> > Index: src/network.c
> > ===================================================================
> > --- src/network.c       (revision 373299)
> > +++ src/network.c       (working copy)
> > @@ -280,7 +280,7 @@
> >      GET_S_FAMILY(f, family);
> >      GET_S_TYPE(t, type);
> >
> > -    if (family > 0) {
> > +    if (family >= 0) {
> >          TCN_THROW_IF_ERR(apr_socket_create(&s,
> >                           f, t, protocol, p), a);
> >      }
> > @@ -290,7 +290,7 @@
> >      a = (tcn_socket_t *)apr_pcalloc(p, sizeof(tcn_socket_t));
> >      a->sock = s;
> >      a->pool = p;
> > -    if (family > 0)
> > +    if (family >= 0)
> >          a->net = &apr_socket_layer;
> >      a->opaque   = s;
> >      apr_pool_cleanup_register(p, (const void *)a,
> > +++
> > It works.
> >
> > Comments?
> 
> With the above patch it works for me too.
> 

Ahhh. Good catch on searching for a family check. Now do
see why it didn't dump core for me :)

-- 
===========================================================================
   Jim Jagielski   [|]   jim@jaguNET.com   [|]   http://www.jaguNET.com/
	    "If you can dodge a wrench, you can dodge a ball."

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@tomcat.apache.org
For additional commands, e-mail: dev-help@tomcat.apache.org