"Jesse D. Guardiani" <[EMAIL PROTECTED]> writes:

> I don't think the reverse verification process would ever work against
> a qmail MTA since the bounce doesn't happen at the SMTP conversation
> level. (Or does it? Maybe I'm confused?)

Yes, you are correct.  Documentation on this Postfix feature is
limited, but I was able to find a SENDER_VERIFICATION_README file
which says:

  Limitations
  ===========
  Postfix probes the nearest MTA for the sender domain without
  actually sending mail. If that MTA accepts the recipient, then
  Postfix assumes that the address is deliverable, even when the
  address will bounce AFTER that MTA accepts it.

So this feature won't be able to effectively probe a qmail server
since qmail accepts the recipient during the SMTP negotiation.
However, most MTAs don't have this behavior, so this feature will be
able to reject most unreturnable mail.  This includes Sendmail, Exim,
Postfix, Exchange Server (e.g, hotmail.com), CommuniGate Pro, etc.

> Let's talk hypothetically here: If I implement some Python code in a
> professional manner that effectively distinguishes between
> autoresponders and bounces, but is a totally optional feature, would
> you include it?

Sure, I'll keep an open mind if you figure out a reliable way to do
this.  I won't add a hack though, even if it's optional, because it
could lead to users inadvertently losing mail.  This issue is not well
understood, so it's likely many users would enable such a feature
without thinking through its ramifications.
_________________________________________________
tmda-workers mailing list ([EMAIL PROTECTED])
http://tmda.net/lists/listinfo/tmda-workers

Reply via email to