On Fri, 03 May 2002, Craig Sanders wrote:
> On Fri, May 03, 2002 at 02:01:45PM +1000, Matthew Hawkins wrote:
> > I would then use something like a transport map to pick up the people
> > that want to be exempted, and feed them via the route that does not do
> > header/body checks in cleanup(8). But please check the doco, that's
> > just off the top of my head and (since I haven't got that snapshot yet
> > and tried it out) could be wrong.
>
> i have no idea if that would work or not but it sounds like it would
> defeat one of the two advantages(*) that body_checks and header_checks
> have over filtering in the delivery agent - i.e. that messages can be
> rejected in the SMTP session and thus avoid having to queue bounces
> (most of which will be undeliverable).

header and body checks are done by cleanup, not smtpd, so the advantage
is already defeated [*]. In any case, what we're doing is simply
applying a different cleanup daemon instance for a particular user (or
users) That's what this functionality was added to the snapshot for. My
only uncertainty is whether or not transport_maps is the best place to
specify the list of morons to subject to spam. I've only read the blurb
for this new snapshot feature and not the code or any supporting doco.
It could also possibly be done using aliases - I'm just not sure how the
import to the second cleanup daemon would be done that way. At least
using a transport map you can say:
    joe_bloggs smtp:[127.0.0.1:20025]
with master.cf setup to run your second cleanup daemon on that one.

> in short, you may as well use procmail to do this kind of filtering. or
> maildrop. or mailagent. or similar programs. that's what i used to do
> before body/header checks appeared in postfix.

And what do the people who cannot run a filtering MDA do? For example,
pre-sieve Cyrus IMAPd on a box with no local users bar the necessary
service ones. I for one am glad Postfix has a way to help those people.

[*] just to confuse people, the advantage is not lost at all. smtpd is
calling cleanup and rejecting inline AFAIK, and I can't see how
specifying an alternate cleanup daemon is going to affect that -
except that this cleanup daemon is not doing header/body checks and
hence won't have anything to reject in the first place ;)

-- 
Matt
-
To unsubscribe, send mail to majordomopostfix.org with content
(not subject): unsubscribe postfix-users

• Messages sorted by: [ date ] [ thread ] [ subject ] [ author ]