Re: Potentially Predatory Pre-Announcement of Possible Vaporware (zinq-djbdns-0.01)

John Levine wrote:
>>> Honestly, what's the requirement?  It compiles fine without it now,
>>> so this appears to be adding a dependency with little benefit.
>> I tend to agree.
> 
> Other than the errno nonsense, which we all know is a one line patch,
> I've never heard any complaints about the portability of djbdns.
> 
>> A better project (IMO) would be to provide well-packaged RPMS.
> 
> Right.  There's a few reasonable patches worth packaging in, e.g., the
> small bug fixes from J. deBPollard.  Beyond that, the recently
> discussed patchlet to handle large responses seems benign and useful,
> and at some point we'll need to think about EDNS0 and maybe even
> DNSSEC.

A couple of further things...

1. If we're talking packaging then it would be worth packaging the tools 
and the servers separately. It would be good to be able to use dnsip, 
etc. on hosts without having dnscache/tinydns installed.

2. I posted a request on the BIND list recently asking for zonefiles to 
use to test a bind2djbdns script that I *will* get round to writing 
soon. I didn't get many responses, but I got one from one "Alan Clegg" 
(from ISC) making the point that djbdns does not support dnssec. He sent 
a follow-up stating:

"BTW -- I sent that file as a bit of an evil prank.  Note that the zone
is DNSSEC "enabled".  DJBDNS/TinyDNS will never be able to correctly
serve that zone.  It does not know how to to provide the correct
responses to queries requesting signed responses.

This is not the best time to be going in reverse (from my point of view)."


So, DNSSEC support would be a "Good Thing" to have.

R.