Speaking of BATV and similar sender rewriting - many kinds of mail-based
server software use the envelope sender address to identify the sending user.
For instance, if I'm going to enable import of mail in Bugzilla, the sender
should be required to have an account or the mail will be rejected (it helps
in case some spam slips through, but if you feel lucky it's of course
possible to create an account for anonymous posts).
The From: field may be the best place to look after all, but as always, it's
not possible to accept some recipients and reject some after DATA. When
rejecting at RCPT it's also easier to tell the sender what went wrong, I
think.
So, which one is better?
1. Look at the From: field and reject the whole message if the address is not
recognized?
2. Look at the From: field and let a bounce be sent, risking collateral spam,
if some of the recipients don't like the addresse, but deliver the mail to
the rest?
3. Look at the From: field and *discard* the recipients that don't like the
address, but deliver the mail to the rest? (The mail sent by Bugzilla would
clearly state that you need an account to comment, or else the mail will be
discarded.)
4. Look at the envelope sender and reject at RCPT, forcing the senders to
disable BATV somehow, sometimes (which also means that they have to skip BATV
checking on incoming bounces from certain addresses)?
5. Same as 4 but let a bounces be sent. This item is mostly included for
completeness.
I'm leaning towards the first option. Most MLM setups use number 2 or maybe 5,
I guess. Or 1 or 4 depending on the level of integration with the MTA.
--
Magnus Holmgren holmgren@???
(No Cc of list mail needed, thanks)
