I'm having a spam problem. One day recently, I was sent several
thousand spam messages all to different email addresses in my domain.
Since all were unknown users, all were forwarded to me personally.
I wish that I could mimic the traditional sendmail behavior. Mail to
unknownuser(_at_)mydomain(_dot_)com should be bounced back to sender.
Here's my tiny .fetchmailrc...
poll best.noir.com with proto IMAP envelope Delivered-To
localdomains noir.com
user "user1" there with password "something"
is rich ice trinity root here
antispam 571 550 501 554
As far as I can tell, there is no difference between mail received and
forwarded for "rich" and mail received and forwarded for "hksS9hQ3San".
Other than potential for rewrite, (rich1=rich2), is there any?
When the number of "bad" addresses is small and well known, I can
explicitly alias those addresses into bounces using sendmail, (awkward,
but feasible). However, when the spammer is apparently automatically
generating the addresses, I appear to be stuck receiving all of them if
I'm using fetchmail.
I'd like to point out two key bits. First, my isp does include a
"delivered-to" header consistently and reliably. Second, sendmail
really does know who is and isn't a valid address in my domain.
Fetchmail's ability to forward all such unknown addresses to me is
useful for debugging, but seems misguided as an attempt to second guess
sendmail.
Suggestions?
--rich