David F. Skoll wrote: > [EMAIL PROTECTED] wrote: > >>> Which is why I qualified my reply with "if the sending relay is a valid >>> SMTP client." > >> Maybe we are just arguing semantics but anything that connects to >> my mail server and speaks RFC821 is valid. I might not like what >> it feeds me but that is what ClamAV/SpamAssassin is for. :) > > OK, let me be precise: By "valid SMTP client", I mean one that > generates a DSN in response to a 5xx status code.
Fair enough. >> Then even in the unlikely event of a false positive the sender >> knows. > > This is so unlikely that the backscatter risk outweighs the benefit. I have had it happen. When messages mysteriously go missing and people call me asking where it went I can be rest assured saying that if something was rejected somewhere they should have received a bounce. It makes things easier to debug when there is feedback. What backscatter? If done at SMTP the only person that should be notified is the sender. If that sender goes and does something stupid with my rejection then that is the senders problem. Otherwise there is zero backscatter. Steven _______________________________________________ Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net http://www.clamav.net/support/ml