Re: How to parse an AXFR response packet

2001-04-16 Thread Rahmat M. Samik-Ibrahim

Hello:

I have checked the POISSON charter:
   http://www.ietf.org/html.charters/poisson-charter.html
which has no goal at all. Therefore, I can predict
that this thread will go to /dev/null, with or without
using .procmail. Also, this vendetta can be traced back 
to 1992:
   http://ittf.vlsm.org/ietf/16.txt


Lloyd Wood wrote:

 This seems to be a clear example of emergent mailing list 
 behaviour I worried about in recent discussion on Poisson 
 (hence the cc.).

RFC-2418 is a product of the POISSON group. Do you believe
that the WG should revise it? If yes, which part?
 
 Here, XYZZY holds the conflicting roles of chair,

A hollow sound plugh...  :-) 


DISCLAIMER:
   I believe that DJB's tough DNS implementation sucks,
   but that is not the point!

regards,

-- 
Rahmat M. Samik-Ibrahim - VLSM-TJT - http://rms46.vlsm.org
- If ain't broke, ain't fix IT;but I'm broke, so IMFix IT!




Re: How to parse an AXFR response packet

2001-04-14 Thread D. J. Bernstein

[In October 2000, the IESG said that it had approved Bush's rejection of
``SPAM  other postings unrelated to WG.'' Bush is now attempting to cut
short an on-topic discussion. The WG has not authorized Bush's behavior.
Has the IESG authorized it? Can the IESG do this without WG approval?
See http://cr.yp.to/djbdns/namedroppers.html for background.]

I would have my AXFR client discard authority records if I saw even the
slightest justification for doing so. But all we have here is the BIND
company demonstrating how little it cares about compatibility.

The BIND company's Andreas Gustafsson says, in a document that claims to
be a ``clarification'' of the AXFR protocol, that clients ``MUST ignore
any authority section contents.'' Everyone agrees that this is not
required for interoperability: servers must not, and do not, put
anything into the authority section.

The BIND company's Brian Wellington says ``there may be protocol
extensions in the future.'' But it is the responsibility of the
extension to preserve compatibility. When the standards permit a
particular behavior, and that behavior is deployed, it's too late to
come along and say ``Gee, we might want to prohibit that behavior to
make extensions easier.'' Put the extensions on another port.

The BIND company's Mark Andrews says, in particular, that they might
someday extend the authority section in a way that produces incorrect
results with my client. Does he interpret this as a flaw in this
hypothetical extension? No! He demands that I change my client, and that
all my users upgrade.

Wellington claims that changing my AXFR client to discard authority
records would make it ``more compliant.'' More compliant with what?
Nothing in the existing DNS standards requires, or encourages, or even
suggests this behavior.

---Dan