Hi,

On Tue, 07.09.2004 at 09:58:43 -0400, Ted Zlatanov <[EMAIL PROTECTED]> wrote:
> On 5 Sep 2004, [EMAIL PROTECTED] wrote:
> > I would like to know whether it would be possible to add spam filter
> > and virus scanner with LDAP support. The idea is to let users choose
> > if they want their email scanned or not. In the web interface they
> 
> Could you look up their LDAP record when you're about to deliver the
> mail, so the choice is made in the global script without involving
> qmail-ldap?  The lookup should be cached, or the LDAP connection
> should be cached, but the basic idea is that you always call the
> scanning script but it doesn't always use CPU cycles.  Implementing
> this is pretty easy.

this is an end-of-pipe-approach.

Some people would prefer to prevent alleged SPAM from entering the
network altogether. Short of hacking qmail-smtpd, gatewaying it with
something like an rblsmptd clone that handles the SPAM problem using
LDAP would be the way to go, imho.

> In our environment we allow users to set up a .procmailrc, and then
> LDAP will tell qmail-ldap to use procmail for those specific users (as
> a delivery program).

Ugh. I'm no longer a fan of procmail... but of course, asking
LDAP in qmail-scanner or SpamAssassin could be done, too. I'm not
convinced that pooling connections and/or caching results is possible
with either directly asking LDAP out of qmail-scanner or procmail since
both are fired off in a separate, stateless process for each email that
arrives.


Best,
--Toni++

Reply via email to