(moved to SMTP list)
--On Sunday, 07 September, 2014 16:30 +0000 Arnt Gulbrandsen
<arnt(_at_)gulbrandsen(_dot_)priv(_dot_)no> wrote (on
apps-discuss(_at_)ietf(_dot_)org):
I don't think what postfix does is a protocol violation. The
4xy and 5xy reply codes were assigned according to the same
principles, and so _should_ be similar apart from the
permanent/temporary bit.
If it makes you happy, ignore that comment, which is not
reflected in the draft.
However, it seems to me (and to Ned -- see discussion on
apps-discuss) that 5321 (and predecessors) have been very clear
that the fourth digit is for fine distinctions only. It would
be correct to say that 5321 does not explicitly prohibit
changing the first digit of a code and then assuming that they
results are otherwise the same, but I can find nothing in it
that justifies or authorizes that behavior.
In any event, since 5321 does not specify a 456 code, assignment
and use of 556 should eliminate any possible issues. If
something turns it into a 456, that would both violate the
intent of nullMX (and the code itself) and force any client that
does not assign special meanings to codes not in 5321 (as
amended) back to the "use only the first digit" rule.
So, if 521 was a problem because of some Postfix behavior, use
of 556 eliminates that problem. As a bonus, it provides a
distinction between "that host does accept mail and it told me
so" and "that host doesn't accept mail and something or someone
else told me so". Given all of the configuration errors we have
seen in the past, I think that distinction might prove useful
and is almost certainly not harmful. My preferring to justify
the added code on that basis rather than on what Postfix does
--"violation" or not-- may be considered just a matter of taste.
best,
john
_______________________________________________
ietf-smtp mailing list
ietf-smtp(_at_)ietf(_dot_)org
https://www.ietf.org/mailman/listinfo/ietf-smtp