John C Klensin writes:
...
Is that more or less what you have in mind?
Yes.
If so, two questions for the group:
(1) Is this worth doing (I think yes -- at least the
C: MAIL FROM:<@foo.com:JQP(_at_)bar(_dot_)com>
needs to be fixed for consistency with other changes).
That's exactly what made me aware of this, so chalk up one more yes.
(2) Note that a side-effect of the famous issue 25 (about NDN
"bounces") is that, if the relay (foo.com) cannot verify either the
forward-pointing address (Jones(_at_)XYZ(_dot_)COM) or the backward-pointing one
(JQP(_at_)bar(_dot_)com) prior to sending back the 250 response to the RCPT
command, it is in big trouble. If we
apply this fix (or even if we don't), does it immediately go into the
Issue 25 list?
I'd try to skirt the issue by saying "foo.com, having received the
message, verifies delivery to jones(_at_)xyz(_dot_)com is permissible and
possible. foo.com now does a DNS lookup on xyz.com."
Arnt