On Fri, Aug 06, 2021 at 12:38:24PM -0700, Dave Crocker wrote:
Well, since "Delivered-To:" is for MTA-private use (loop detection) and
Sorry. I missed hearing about the documented rule that restricts it use
this way.
Today, the "Delivered-To" header introduced by Qmail, and adopted in
Postfix and Exim records an internal non-public mailbox address used in
MTA-to-LDA communication, that only happens to coincide with the
incoming public envelope address with some frequency. It is used for
loop detection by the MTA.
(by default) records the internal envelope recipient passed to the LDA,
rather than the input envelope recipient from the sending agent, it was
[not] deemed appropriate to conflate these into a single header.
It's not conflating. It is, arguably, exactly the same use.
I
Note that gmail, for example, merely puts the .... delivery address...
into the field.
Gmail does not appear to rewrite the input recipient address in transit
from MTA to LDA. That's a choice they're free to make.
The claim that this field gets some sort of mystical, obscure token has
yet to be document[ed].
I am communicating its use in Qmail, Postfix an Exim now, for the nth
[and last] time. In these MTAs it is a loop detection token, and not an
address that is expected to be recognised by downstream user agents or
MTAs. The only intended audience is the same MTA (or MTA cluster),
should the message come back around...
Perhaps this information is not welcome, or not sufficiently
authoritative. I can't do anything about that.
What I can say is that I see no compelling case to participate in some
new experimental use of this header. Postfix has:
- X-Original-To: for the inbound external envelope recipient, for
use by fetchmail, ...
- Delivered-To: for recording the MTA-to-LDA handoff envelope, used
to detect loops when LDAs reinject the message downstream.
I guess the maintainers of Qmail, Postfix and Exim are fools, who've
failed to realise that they should be using the same header for both
purposes.
[ With both address forms recorded, fetchmail would have to figure out
whether that's what happened, and which of the two topmost Delivered-To
addresses to use to demultiplex multi-drop mailboxes, after we announce
a suitable flagday when the relevant MTAs join the experiment. ]
I have nothing further to add...
--
Viktor.
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp