Re: New options

[Dan Nelson]

In the last episode (Jun 13), Valentin Chopov said:
> thanks for the explanation.
> I'm running access.db, DNSBL ant 2 more milters before spamass-milter
> and maybe this is why this happen. I'll wait for the fix and try
> again. I have a questions regarding the new options -i & -b
> -i is very useful and I'm going to use for our block of networks. My
> question is, is it safe to add 127.0.0.1 (localhost) into this list.
> My understanding is that this will be safe but I need to be sure. The
> reason I'm asking this, is because of some .forward scripts in wich
> case the last hub is the localhost.

Yes, that's safe, since any email coming in on 127.0.0.1 was either
generated locally, or originally came from another IP (and has already
been processed).

> My 2-nd question is about the option "-b" and how it cooperates with
> the "-r". Which mail is forwarded to the bucket - the mail tagged as
> a spam from the SpamAssassin or the mail rejected with "-r" ? I
> personaly preffer the 2-nd one ;)

Unfortunately, you can't do that :)  The libmilter API doesn't allow
you send a reject message back to the sender AND continue delivering
the mail.  You can do both, but it requires you to patch the sendmail
source and rebuild - savannah patch 714 does this.  So the current
behaviour with both flags is

score < required_hits : deliver normally
score between required_hits and -r value : tag
score >= -r value : reject

The only other way to do this would be to have spamass-milter reinject
the message into sendmail itself, which would add another Received:
header and complicate things (at minimum you would have to require -i
127.0.0.1 to ensure that you don't get into a forwarding loop with
yourself)

-- 
        Dan Nelson
        address@hidden