On Mon, 2005-04-25 at 14:45 -0400, Steven W. Orr wrote: > On Monday, Apr 25th 2005 at 13:20 -0400, quoth Bruce Dawson: > > =>On Mon, 2005-04-25 at 12:47 -0400, Steven W. Orr wrote: > =>> Thanks, that's just what I don't want to do. I really do want to reject > =>> the mail before reception is complete. After I get it then they know I > got > =>> it and I have to take the rouble to report it to spamcop. > => > =>This is rather difficult to do because it requires the MTA to start the > =>filtering *before* reception is complete, which means the filtering > =>process has to be able to complete processing before or by the end of > =>reception. > => > =>The only mechanism I know that is capable of this are rather complicated > =>firewall rules that looks for and monitors connections to port 25. These > =>tend to chew up a lot of resources on the firewall machine - especially > =>during a spam storm. > => > =>Most people just punt and use the RBL lists (I do). They have problems, > =>but they do stop *a lot* of useless traffic. (Note that I use > =>list.dsbl.org, sbl.spamhaus.org, dnsbl.njabl.org, bl.spamcop.net, and > =>cbl.abuseat.org, and NOT sorbs! > > Umm, no. That's exactly what milters are all about. Sendmail processes the > incoming data stream through the list of milters. If a milter returns a > failure status code then sendmail will reject the message with a 500 > series error. Works quite well actually. I use RBLs as well as a number of > other tricks in sendmail before I even get to the milter.
Steven: Thanks for the clarification. I was under the impression that the milter is called only after the message had been received. --Bruce
signature.asc
Description: This is a digitally signed message part
