Tony Hansen | 8 Mar 00:10 2007

update to SMTP/RFC2821 about to become available

There is an update to RFC 2821 (Simple Mail Transfer Protocol, aka SMTP)
about to become available from the secretariat in an internet-drafts
repository near you. It will be discussed on the ietf-smtp <at>
mailing list.

Here are the guidelines to follow for comments on it:

We want to advance RFC2821/SMTP to Draft Standard.  Thus, we don't want
to design anything "new" or "better" for SMTP as part of this particular
effort, which would require recycling at Proposed Standard -- in fact,
we want to be conservative in minimizing changes.   As with any movement
from DS to PS, changes should be limited to:

   1)	clarifications of the spec where there are discrepancies, and
   2)	removal of unused features.
   3)	The fixing of "bugs" should be limited to those that have a
	*very strong* consensus.

*Do not* bother raising the issue of reorganizing the document once
again, and/or breaking it up into separate pieces. Such emails will be
summarily ignored.

*Do not* ask us to reopen the NDN question (whether it is reasonable to
accept a message and then send a message-type rejection, and prohibit
that behavior "unless the sender can be guaranteed to be authentic").
While quite a worthy topic, this question will not be addressed in
2821bis. >>>Instead<<<, write an internet-draft that is an *UPDATE* to
2821/2821bis, and we will be happy to review such an I-D on the
ietf-smtp <at> mailing list.

(Continue reading)

Arnt Gulbrandsen | 28 Mar 22:06 2007

2822bis and "" <at>

In 2822 that address is apparently valid. Should that be so?

I could understand it being a valid obs-this-or-that address. But it 
seems to be quite simply valid. Message parsers must parse it, messages 
senders may generate it.


Frank Ellermann | 31 Mar 21:19 2007

Re: 2822bis and "" <at>

Arnt Gulbrandsen wrote:

 ["" <at>]
> In 2822 that address is apparently valid. Should that be so?

FWIW, "" is not allowed as <id-left> in I-D.ietf-usefor-usefor-12
and in an I-D about news URIs based on that approved USEFOR RFC.

It's still allowed in the 2822upd-00 I-D, that draft doesn't yet
support "canonical Message-IDs" as specified in the USEFOR RFC.

An interesting note in 2822 and 2822upd-00 is the "semantical
content" of a quoted-string, it's (no surprise) an empty string
for "".

I'm unsure what a <local-part> that's "semantically empty" might
be, apparently (quote) "it is simply interpreted on the particular
host as a name of a particular mailbox" (unquote 3.4.1).

And SMTP can transport it, so that appears to be fine, ignoring
the question of canonical Message-IDs for the moment.  Better
than "\e2J" <at> for e = ESC, or other NO-WS-CTL cases ;-)