On Fri, 2005-01-14 at 19:02 +0100, Hannah Schroeter wrote:
Thus I'd vote for SRS (or similar) to be skipped on SES senders.
That's just a specific instance of a more general rule -- SRS should be
skipped unless it's actually necessary.
A given host should only do SRS where the SPF result of sending the mail
from that host's IP address would be a 'fail', and the next recipient is
known to check SPF.
If the sending domain doesn't publish SPF records, or ends them with
'?all', or has an SPF record which uses the 'exists' mechanism to
express SES in a way which doesn't depend on the IP address, then SRS
shouldn't be done.
--
dwmw2