Scott Lamb schrieb am Thu, Apr 25, 2002 at 11:27:10PM -0500: > Scott M Likens wrote: > > >This is quite Sickening, RBL is a MTA implementation not needed to be > >done via Sieve, and as for spamassasin you can always write decent > >header checks and body checks for postfix to use. I am sure there is > >the same option in sendmail. > > > If you have some idea how I could accomplish my stated goals on the MTA > side, please share. I've given my reasons for this approach. Why do you > feel so strongly that this belongs in the MTA? > > Here's one I haven't mentioned yet - some people want to do 'fileinto > "INBOX.spam"' instead of outright blocking spams. With something purely > on the MTA side, this isn't possible at all - Cyrus is what decides > where the mail gets filed. This is possible with the X-RBL-Warning, but > I've pretty thoroughly stated why I think that's inadequate. I won't > repeat myself.
I think both sides may have their share. MTAs are definitely the place where to apply recognition patterns on the incoming mail stream. Thus an MTA will _mark_ a mail as "gold", "sending-mta-is-on-rbl" or "garbage" or <whatever>. This is a) efficient and b) does no harm to users (but may violate some RFC I've forgotten that states that MTAs should only alter envelopes, never the mail itself?). Sieve/some user controlled filter on the other hand are the place to apply a _policy_ to the marks set above. If a user likes spam - there he goes. If a user wants to have every "gold" mail printed immediately on his local lpr: he can do that. I think MTAs should only apply and en- force policy if the mail system itself is in danger of falling apart i.e. due to 10.000 spam mails/h that are blasting it away. A side note: the MTA should see that the headers it sets (the "marks") are not insertable by other MTAs/users - it should always replace them. If not, some nasty people who know something about the way your MTA marks mails may kill all messages for someone who does /dev/null to, say, "garbage" marked mails. Regards, Birger
