Re: RIPE DNS Hostcount February 1993

[Sam Wilson ercm20@localhost]

> > [ zone transfers adding glue records which are not in the zone files ]
> 
> Those "glue records getting transferred" problems seems to appear on many
> systems, Bind 4.8.3, but also Sun's named. ...

I wouldn't treat Sun's named as representative of anything.  This
feature is in all unpatched versions of BIND (which Sun's named is). 
The best known fix is the 'Don Lewis Jumbo Patch' (details attached
below). 

> I'm waiting for Bind 4.9 to see if this problem is persistant.
> Perhaps could Paul Vixie indicate us if this issue has been
> adressed, or just even known, in 4.9 ?

This problem is certainly known.  I can't speak for Paul Vixie, and I
haven't seen 4.9 yet, but I'd be very surprised if it wasn't fixed
there. 

Sam Wilson
Network Services Division
Computing Services, The University of Edinburgh
Edinburgh, Scotland, UK
------------------------------------------------------------------------
From edcastle!dcl-cs!glasgow!daemon 10 Jun 92 04:28:01 GMT
Path: edcastle!dcl-cs!glasgow!daemon
From: gdonl@localhost (Don Lewis)
Newsgroups: mail.bind
Subject: Re: The "infamous" Don Lewis patches
Message-ID: <9206100428.AA26122@localhost
Date: 10 Jun 92 04:28:01 GMT
Sender: bind-request@localhost
Organization: Glasgow University Computing Science Dept.
Lines: 59
Approved: usenet@localhost
In-Reply-To: mandrews@localhost (Mark Andrews) "The "infamous" Don Lewis patches" (May 6, 1:16pm)
Resent-Message-Id: <9206100505.AA11938@localhost
Resent-Date: Tue, 9 Jun 1992 21:28:01 -0700
Precedence: bulk
Resent-To: bind@localhost
Resent-From: bind-request@localhost
Original-Sender: bind-request@localhost
X-Mailer: mail-news 2.0.5

On May 6,  1:16pm, Mark Andrews wrote:
} Subject: The "infamous" Don Lewis patches
} 
} Time to refresh my memory. How many Don Lewis patches are there to bind4.8.3
} and what are their "names"?
} 
} I know of the lame delegation patch with a message id of
} <9102132021.AA05427@localhost, but what are the others?

That's my BIND Jumbo patch (combining all my other patches and well as
any other patches I picked up from the net).  This is the last patch
set I published, since I changed jobs  month later and lost my Internet
access :-(.

Please note that there is a minor bug in part of the patch (which disables
one of the bug fixes).  The if test and goto added around line 210 in
ns_forw.c should be inserted between the preceeding for loop and if test.

** 208,213 ****
--- 221,228 ----
                for (dp = np->n_data; dp != NULL; dp = dp->d_next) {
                        if (dp->d_type != T_A || dp->d_class != class)
                                continue;
+ if (dp->d_type == T_CNAME && dp->d_class == class)
+       goto skipserver;  /* XXX - Hack for now */
                        /*
                         * Don't use records that may become invalid to
                         * reference later when we do the rtt computation.

should be:

** 208,213 ****
--- 221,228 ----
                for (dp = np->n_data; dp != NULL; dp = dp->d_next) {
+ if (dp->d_type == T_CNAME && dp->d_class == class)
+       goto skipserver;  /* XXX - Hack for now */
                        if (dp->d_type != T_A || dp->d_class != class)
                                continue;
                        /*
                         * Don't use records that may become invalid to
                         * reference later when we do the rtt computation.



Also, the original patch description failed to mention the TRACEROOT
compile time option to toss out bogus looking root server names
(whatever.ARPA, etc.).

} 
} Are they available for anonymous ftp somewhere?
} 

The patch file mentioned above (complete with bug) is available as:
	slopoke.mlb.semi.harris.com:pub/semi/named/diffs-vs-4.8.3


Don "Truck" Lewis              Phone: +1 916 478-8284   Silicon Systems
Internet: gdonl@localhost       FAX:   +1 916 478-8290   138 New Mohawk Road
UUCP: {uunet,tektronix!gvgpsa.gvg.tek.com}!ssigv!gdonl  Nevada City, CA  95959