On Fri, Oct 24, 2003 at 02:36:19PM -0400, Greg A. Woods wrote:
> [ On Friday, October 24, 2003 at 10:12:28 (+0100), Tony Finch wrote: ]
>> Subject: Re: [Exim] Refuse connection if no MX for sending host
>> On Thu, Oct 23, 2003 at 12:03:01PM -0400, Greg A. Woods wrote:
>>>>> There _should_ always be an MX for every domain name that is a host name
>>>>> of a mailer, and the target name of the MX should be the same domain
>>>>> name:
>>>> Where does it say that?
>>> Before I answer that (because you're sure as heck not going to like the
>>> answer!), let me ask you this:
>>> Why the hell wouldn't you publish an MX for every mail server?!?!?!?!?
>> Because there are no valid email addresses at that domain.
> Strictly speaking there is always one valid e-mail address at every host
> which answers SMTP, and even all of those which might speak SMTP -- at
That's not an answer to what Tony wrote. There is a valid email address
at that host, but there is no valid postmaster@<hostname> address. The
valid address is the unqualified postmaster address.
> least when they might speak as a relay (i.e. ignoring mail readers which
> should probably use LMTP not SMTP for sending outgoing messages, but
> which have not since historically LMTP was not specified when they
> became common).
So this is a complete diversion from what Tony actually wrote, but used
to justify your own ravings.
>> See my previous
>> comment about not supporting email delivery to arbitrary computers.
> I never said anything about supporting e-mail delivery to arbitrary
> computers. If you re-read my very simple question above you'll
> hopefully understand that I asked only why you wouldn't publish an MX
> for the host domain name of every _mail_server_ -- i.e. those systems
> which do already accept e-mail delivery by way of SMTP.
Because they don't accept their hostname as a valid domain-part. What
part of this is difficult to understand? Tony said it in your quoted
message, I said it (though you snipped it), and you've raved on about
postmaster addresses without really answering any of the questions that
people have put to you about your outrageous assertions. You haven't
answered anything yet, you've attempted, badly, to fob people off. If
you're going to make assertions, you have to back them up with something
more solid than just "if you go read RFC2821 for yourself, you'll
understand". I have read RFC2821 and I still don't agree with your
assertions.
Tony asked you a perfectly valid question, "where does it say that you
have to have MX records for the host name of every host that has other
MX records pointing at it". You have, so far, totally failed to answer
this question, instead trying to divert people's attention away from
the fact that you are unable to answer it.
MBM
--
Matthew Byng-Maddick <mbm@???> http://colondot.net/
