On Thu, 2004-03-18 at 10:47 -0500, Stuart D. Gathman wrote:
On Thu, 18 Mar 2004, David Woodhouse wrote:
On Wed, 2004-03-17 at 23:27 -0500, Stuart D. Gathman wrote:
I hope you followed that. It is no big deal yet, because usually all the
failing bounces from client servers are attempts to tell spammers that
the mail can't be delivered.
Stop there. You are sending bounces to spam, instead of just rejecting
it at SMTP time? Fix that first.
When I know it is spam, I just reject and discard it. However, I can't
be sure it is spam just because the recipient doesn't exist. Often,
legitimate users misspell email addresses.
True. But you don't need to accept the mail and then subsequently bounce
it; you can just reject it. Unless, of course, it comes from an
authenticated source, perhaps by TLS+AUTH, in which case it's probably
OK to accept it before making the decision.
A system should not generate bounces in normal operation.
At some sites I use a content filter (DSPAM wrapped for Python), but
this has its own problems with false positives.
Still, why is it causing bounces and not just rejection?
--
dwmw2