Tony Finch <[EMAIL PROTECTED]> writes:

> This isn't feasible for site-wide installations of SpamAssassin like
> mine (more than 30,000 users). Academic users use forwarding a lot
> because they move around so much.

In the end, I used "-notfirsthop" logic for those DNSBL tests I modified
which is close to or better than "-firsttrusted" in terms of results, but
doesn't suffer the same loss of hits when mail is relayed or forwarded.

Daniel

-- 
Daniel Quinlan
http://www.pathname.com/~quinlan/

Reply via email to