Hector Santos wrote:
GA: That would be entirely up to the receiving server. I am
GA: certainly not going to assume that there are no cases where
GA: this would happen, that could result in lost mail.
I will go on the record, not that it means anything, that this 4.2.5
interpretation is incorrect and does cause mail lost, clearing born
from the 8 year old gaffe in 2821 which allowed 5yz retries. I have no
doubt about it, and none of you wish to acknowledge.
Hector, I'm struggling to see any case where doing a retry on a message
which couldn't be delivered to a temporarily unavailable recipient could
lead to a message lost, unless there are other issues going on as well.
What you seem to be saying is that if I send a message to one person and
that message is rejected after the data with a 5xx error, then that
exact same message should not be attempted to be sent to anyone else.
This is obviously wrong, as many people will block messages over a
certain size, or containing attachments, or whatever.
What you're saying is that:
----------------
RCPT TO: d(_at_)domain(_dot_)com
250 OK
RCPT TO: r(_at_)domain(_dot_)com
4xx mailbox full, try later
DATA
354 go ahead
.....
.
5xx attachment prohibited
-------------------
is NOT identical to splitting it into two separate recipients:
-------------------
RCPT TO: d(_at_)domain(_dot_)com
250 OK
DATA
354 go ahead
.....
.
5xx attachment prohibited
RCPT TO: r(_at_)domain(_dot_)com
4xx mailbox full, try later
(then some time later)
RCPT TO: r(_at_)domain(_dot_)com
250 OK
DATA
354 go ahead
.....
.
250 message accepted
--------------------
What the rest of us are saying, is that they ARE identical. In your
case, the message would not be delivered to 'r', ever. In everyone
else's case, the message would be delivered to 'r', but later.
To me, it seems obvious that putting two recipients on the same message
should behave exactly identically to sending the same message to two
individual recipients.
--
Paul Smith
VPOP3 - POP3/SMTP/IMAP4/Webmail Email server for Windows