[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: in support of axfr-clarify
At 11:27 PM 11/29/02, D. J. Bernstein wrote:
Mayer claimed ``You can't clarify without breaking something.'' Is there
anyone else here who doesn't understand how ridiculous that claim is?
It's not ridiculous. In this case, whatever you do will result in someone
being in nonconformance with the proposed spec.
BIND company consultant Danny Mayer writes:
> I am not and never have been an employee of either ISC or Nominum.
I apologize for the error. The fact remains that you're not disclosing
your financial interests in BIND.
I have no financial interests in BIND, though I see no need to discuss
my financial interests in anything, least of all to you. If I make use of
BIND in the course of my employment or in pursuance of new
employment, do you consider that a financial interest?
> Since it's ambiguous, people have made different interpretations of
> the spec resulting in compatibility problems.
A typical implementation difference---whether or not it is caused by
different interpretations of the spec---does _not_ cause any failures.
If you would read that sentence again, I didn't say anything about
failures. What you get as a result is that two authorative servers
for the same zone which use AXFR to transfer zones from one to the
other can give you answers if the zone transfer is done where each
server has a different implementation of the protocol based on different
interpretation of the spec.
For example, some AXFR servers insert glue records everywhere they're
used, while some AXFR servers insert each glue record exactly once.
AXFR clients can handle either approach. There are no failures here.
Noone said anything about failures. The discussion is about consistency.
The AXFR clients, as you call them, don't get to pick and choose, it's
the server that is making all of the decisions here.
Explaining that servers can repeat records, and requiring clients to
handle repetitions, would be a clarification of the protocol. It would
not cause any failures. It would save time for future implementors, and
reduce the chance of careless implementors screwing up.
I'm not even sure why you are arguing about this since your idea of zone
transfers use rsync, etc. Why should this even bother you? Why are
you so terrified by this proposal?
Danny
--
to unsubscribe send a message to namedroppers-request@ops.ietf.org with
the word 'unsubscribe' in a single line as the message text body.
archive: <http://ops.ietf.org/lists/namedroppers/>