[Top] [All Lists]

Re: SMTP traffic control

2011-10-29 20:44:09

--On Saturday, October 29, 2011 23:37 +0100 Paul Smith
<paul(_at_)pscs(_dot_)co(_dot_)uk> wrote:

"Invalid recipient" should be a 5xx error, or do you expect
that these addresses will be created within the next few days?

Oh, I agree, but we *often* get ISP smarthost mail servers
giving a 4yz error for 'invalid recipient'. We have to deal
with this as it is our software which is failing to send the
message, and is retrying - the user sees 'invalid recipient'
and thinks our software shouldn't retry, so we have to try to
explain SMTP result codes to them...

(I suspect the server is giving a 4yz because it is doing
call-forward testing whether the remote recipient exists, and
gets no response either from a DNS server or mail server for
that address)

And _that_ is a case where, IMO, an extension in which the
server announced that it did call-forward testing (or other
testing that might yield uncertain results) and therefore might
return one or more reply codes with specialized meanings would
be worth exploring.  I haven't thought much about feasibility
(and won't unless someone were serious about the idea), but the
notion of a server being able to response with a reply that
unambiguously meant "can't verify, will keep trying, check back
in a few hours" feels really elegant to me (especially since I
don't think call-forward testing for recipient addresses is
rational unless the relay maintains a cache of valid addresses).


<Prev in Thread] Current Thread [Next in Thread>