On 2000-09-13 at 09:30 +0100, Nigel Metheringham gifted us with:
> and intend when I get time (!) to extend it
> to allow perl callouts at receive time. However remember many MTAs
> don't take permanent error codes at end of DATA so things will get
> worse - you will actually get the stuff being sent to you several
> times. You could just close the link, but then the sender *will* send
> you the message again... and again... Since you have had to accept the
> message already, just throw it now.
You have perl callouts. And Philip Hazel has noted that "more hooks" is
on the wishlist.
So if you're already rejecting the mail because it's spam, _you_ _know_
that the source host has been compromised[1]. None of the "do I trust
RBL list XYZ?" business. So if you're doing all this, you could add the
IP to a database which is used for host_reject. Then return a code
which causes Exim to close the link.
Then clean out the database periodically.
?
[1] If it's a machine materially owned by a spammer, then it's still
compromised. :^)
--
A science is said to be useful if its development tends to accentuate the
existing inequalities in the distribution of wealth, or more directly promotes
the destruction of human life - Godfrey Hardy, A Mathematician's Apology, 1941
