Re: [Exim] Mail problem

Page principale
Supprimer ce message
Répondre à ce message
Auteur: Matthew Byng-Maddick
Date:  
CC: exim-users
Sujet: Re: [Exim] Mail problem
On Sat, Nov 24, 2001 at 03:03:25PM -0500, Dave C. wrote:
> On Sat, 24 Nov 2001, Matthew Byng-Maddick wrote:
> > On Fri, Nov 23, 2001 at 12:46:29PM -0500, Dave C. wrote:
> > > While theoritically using the user@[ip.ad.re.ss] format is technically
> > > valid, it is a terribly archaic manner of addressing, and should be
> > > obsoleted, if it isnt officially already, at least on the public
> > > Internet. No mailhost I run will ever accept that as a sender or
> > > recipient address, local or remote.
> > Then you are in breach of the RFCs.
> > >>> RCPT TO:<postmaster@[ip.ad.re.ss]>
> > MUST be accepted, as must
> > >>> RCPT TO:<postmaster>
> Oh indeed not! Fully qualified addresses with domains only, thank you.


No.

> We dont accept mail for hosts, only for domains. Which of the several
> thousand domains that we host postmaster would you like to send to?


I want it to go to the person who manages the machine I connect to. As
the RFC implies.

> I do recall something about their must be a postmaster address at every
> _domain_ that one runs mail for, but nothing like you mention. Do you


Well, perhaps you ought to read up the spec, then.

> have a reference for these requirements (as a curiosity only, so I know
> which one to reference as I write a new one obsoleting it)


Ah, right, you know enough about mail delivery to do this, then? This is
not apparent from the rest of your messages to the list. The correct
document is RFC2821 dated April 2001.

The paragraph you will want to look at is particularly the Syntax: part
of 4.1.1.3 of this document.

> As someone previously mentioned, site policies are not dictated by RFC.


This is true, however most of those are SHOULD rather than MUST. If you
want to interoperate, you play by the rules, if you don't want to
interoperate, not my problem.

> Especially RFCs from decades ago that took no account of the
> current nature of the Internet.


I'm amazed that you think "April 2001" is decades ago, and that there are
no mentions of policy in the document in question. Perhaps you ought to
actually read it? I think that if you think you know best then you ought
to join the standards track for mail.

MBM

-- 
Matthew Byng-Maddick         <mbm@???>           http://colondot.net/