On Thu, Oct 23, 2003 at 12:35:23PM +0200, Benedict Verheyen wrote: > So, if i understand this correctly, fetchmail will stream the mail to exim > while it is fetching it or does it download it first anyway?
I'm no mail guru, but I don't _think_ that's right.
I believe what happens is that fetchmail downloads the whole mail (but
doesn't yet delete the original from the remote mailbox), passes the
whole thing to your local mailserver, and _only_ after receiving
confirmation that your local mailserver knows what to do with it _then_
it deletes the original from the remote mailbox.
Mailfilter has the advantage that you can do some filtering without
downloading the DATA of the message (just the headers). And naturally,
as you pointed out, this is more limited than full-body filtering
methods like spamassassin.
But you can't apply content-based filtering without downloading the
content.
You could always use both, though. Pre-connect with mailfilter to get
rid of swen and the most obvious spam, grab the survivors with fetchmail
and have the local mailserver (or each user's procmail setup) feed them
through spamassassin to do the really thorough checking.
And I'm pretty sure that it does not make any sense to set up your local
MTA with "reject" rules if you're using fetchmail. The spammer has
already successfully delivered the spam to your remote mailbox, and
fetchmail has already downloaded it for you. Rejecting fetchmail's
connection to the MTA at that point is kinda useless.
Cheers!
--
,-------------------------------------------------------------------------.
> -ScruLoose- | To hell with Saddam <
> Please do not | and may he quickly be joined by Bush. <
> reply off-list. | - Salam Pax <
`-------------------------------------------------------------------------'
pgp00000.pgp
Description: PGP signature
