On Mar 16, 2021, at 6:21 PM, Dave Crocker <dhc(_at_)dcrocker(_dot_)net> wrote:
Well, we have decades of operational experience.
Where is documentation that demonstrates this brittleness?
Well, I rather expect that in fact most implementations don't
actually care about the 2nd and 3rd digit that much. In particular,
for a negative 4XX or 5XX reply is not one of the expected ones from
5321 for that command, they'll do the standard thing for a temporary
or permanent error, and if it is, it'll still be a temporary or
permanent error, be one where it may be able to log something more
precise, ... So distinguishing between ignoring those digits and
not ignoring them should be rather difficult.
ietf-smtp mailing list