[ Only replying to the list, because you apparently block mails sent
via Comcast. About that, see:
http://ars.userfriendly.org/cartoons/?id=20040421 ]
On May 26, 2004, at 10:07, Scott Call wrote:
> The thing about dspam is it's a training based system only so it needs
> to
> modify the body of every message passing through to add a signature,
> so in
> case that message is indeed spam, you can report it back to the dspam
> system.
>
> the default behavior for dspam is to call the MTA again and do the
> delivery.
In that case, something like Exiscan-ACL would not really help you
either. The way it works with SpamAssassin is that it passes on a copy
of the message, and obtains SA's feedback into $spam_score and
$spam_report, as well as its exit status in the "spam" condition. In
other words, SA's message modifications (header markup) are not
incorporated back into the message.
You may need to write your own transport, perhaps invoking a pipe to a
wrapper around DSPAM. Depending on how/where you store your mail (BSD
mailboxes? Maildirs? Cyrus?), you could let that wrapper handle the
delivery directly or indirectly.
-tor
