On 03/02/07, Matthew Hodgson <[EMAIL PROTECTED]> wrote: > On Sat, 3 Feb 2007, Magnus Holmgren wrote: > > Are you sure that you don't want to do spam scanning with the spam condition > > in the DATA ACL, or with SA-Exim? > > Yes, I /think/ so - the reason for forking a new exim and filtering > through spamc once per recipient is so that spamassassin can use a > different user profile for scanning each message, the theory being that > each mail address will receive a different blend of ham/spam, and should > evolve its own corpus and bayesian db to best analyse each message. Or > does that sound nuts to anyone?
No - not nuts. > > I'm also not interested in rejecting spam at SMTP time. As long as that doesn't mean you're going to bounce it later.... > Am I right in > thinking that ACL spam scanning is incapable of making use of > per-recipient user profiles? No, you're not right - with the restriction that it only works with single-reipient messages, you can use the spam = <user> constuct to invoke per-user profiles in SA - as long as you can translate a recipient into a 'user'. Peter -- Peter Bowyer Email: [EMAIL PROTECTED] -- ## List details at http://www.exim.org/mailman/listinfo/exim-users ## Exim details at http://www.exim.org/ ## Please use the Wiki with this list - http://www.exim.org/eximwiki/
