On 3/6/2013 7:22 PM, LuKreme wrote: > /dev/rob0 opined on Wednesday 06-Mar-2013@17:26:02 >> On Wed, Mar 06, 2013 at 11:52:35AM -0700, LuKreme wrote: >>> The bad word begins with u and then is followed by n, s, u, b, an >>> archaic word meaning a person who is employed in writing, and then >>> a final d. >>> >>> u, n, s, u, b, >>> scribe >>> d >> >> Cute. :) >> >>> I [bad word] from a mailing list but the company continues to send >>> emails, so I was considering adding them to my header_checks.pcre >>> file >>> >>> /^Received:.*cinemark\.com / >>> REJECT You refuse to respect [badword-d] requests, welcome to >>> the blacklist. >>> >>> But I thought, before I do this, I better double check that this is >>> the best way to do this. >> >> Almost surely not. You probably want a check_client_access >> restriction to reject all mail from that[those] IP address[es]. Even >> a check_sender_access would be better. > > I have no way of knowing all the IPs, they use some remailer service, and I > don not want to block the remailer because they are not the problem. >
You don't reject based on the IP, you use the client hostname, very likely the same name you're rejecting in your header check. Or use a check_sender_access map with the envelope sender address. Most remailers use a client or envelope sender name something like foo.remailer.com where foo is a unique identifier for that customer, allowing you to reject mail from a specific business without blacklisting the whole remailer. Anyway, if your request isn't honored then it's fair to hold the remailer accountable too. And it's bad form to ask if there's a better way to do something and then argue with the correct answer. >> A good rule of thumb: never do something in the message content if >> you can accomplish the same thing with the envelope. Another one: >> header_checks(5) are rarely useful. > > I find the date checks useful (to reject messages with future/past dates). I used to do that too. Didn't catch much extra spam, but I did discover that my coworkers correspond with a surprising number of folks who can't set their PC to the right year. Maybe you'll have better results. -- Noel Jones
