I've got qmail, spamassassin and maildrop working together to provide a per- user / per-domain spam filtering solution. (using a sql database to keep up with which user / whole domain wants spam filtering and checking that DB through external calls from the .mailfilter file).

My next step is get the same thing happening with anti-virus protection. Again, I'm going to do this on a per user / per domain basis using external calls from the .mailfilter file. That part I'm not worried with, I'll just duplicate the mailfilter code from the spam filtering.

What I'd like to know is: has anyone gotten maildrop to work with a virus scanner (i.e. AVP + Virus scanner)?

a few words about the envisioned set-up:

1. qmail invokes maildrop via the .qmail-default file. 2. .qmail-default calls |maildrop .mailfilter 3. the mailfilter says something like this: if (user has virus checking enabled via sql database) { call AVP } 4. (and this is where it gets fuzzy, hang on...) if (AVP returns "Mail Infected") { send mail to trash, email users of failure. } 5. if (AVP returns "Mail Clean") { to "VPOP" }

Obvious questions about this: 1. What does AVP return when a mail is infected? 2. Does AVP return anything when mail is infected? 3. Am I completely nuts and am going about this all wrong?

I'm very interested to hear how anyone is implementing virus scanning *at all* from the mailfilter file.

Rob Short