On Fri, 26 Mar 2004, Meng Weng Wong wrote:
What's been proposed is
A X) B -> C
But it seems easier to just let the failure occur like with
A -> B X) C
[Assigning those label really cleared up this discussion for me. Thanks.]
So you're suggesting that when incoming mail gets an SPF fail, then
C should return 551 with the recipient address unchanged?
Then, when the reject gets back to A, it will have the address that
was attempted to be forwarded. That way, B won't have to do anything,
and the onus is on A to handle the resending, either manually or
automatically. Furthermore, C doesn't need any extra table or lookups.
He simply reflects back the recipient address he just received when the
sender is not authorized. So this could be a generic feature of spf milters
and such. If the message was in fact spam, a 551 is as good as a 550.
Thought: if A handles the resending automatically, then it could be
helpful for the MTA to keep a cache of forwarding translations, to avoid
bugging B on subsequent communications. But that leaves the issue of
not noticing when B changes the forward. (Cache expires?) When bugging
B, the traffic looks like this:
A -> B X) C
-> A -> C
Question: How does the 551 work with multiple recipients?
--
Stuart D. Gathman <stuart(_at_)bmsi(_dot_)com>
Business Management Systems Inc. Phone: 703 591-0911 Fax: 703 591-6154
"Very few of our customers are going to have a pure Unix
or pure Windows environment." - Dennis Oldroyd, Microsoft Corporation