Well, if you look at my other question, it should be obvious how I'm
planning to work around it :)
I'm working up a delivery script (or just succesive pipe them from the
default delivery, but i think it'll be easier to do a script so I can
add/subtract "filters" without bouncing courier) that will use spamassassin
and anomy.
I like spamassassin because all it does is add ******SPAM****** to the
subject line if it thinks it's spam. Then each user can set up what to do
in their server side filters for maildrop. IE put into a "spam" bucket,
delete, check for known addresses, etc.
Anomy is sort of the same... it will catch normal virus's and delete them
from th emessage (using fprot/mcafee/etc), but will also "cleanse" dangerous
html from messages.
Both of them take stdin, and go to stdout, so I figured my delivery would
just be:
| /usr/local/bin/spamassassin | /usr/local/bin/anomy |
/usr/local/courier/bin/maildrop
----- Original Message -----
From: "Alessandro Vesely" <ves...@tana.it>
To: "Sam Varshavchik" <mrs...@courier-mta.com>
Cc: "courier" <cour...@lists.sourceforge.net>
Sent: Saturday, January 19, 2002 2:45 AM
Subject: [courier-users] Re: courierfilter
Sam Varshavchik writes:
Bill Williamson writes:
Question I have is, can a filter change the text of an email, or just
block/pass?
Just block/pass.
That is not enough. E.g. for sanitizing html
messages against swf or js viruses that can
really act in the preview window.
The next question then is: is anything planned
to overcome this filter limitation, or is there
any workaround, or is it a design decision?
Let me just note that putting some antivirus or
sanitizer filter *before* courier, one could not
use black list filtering based on relay address.
Have a bright day
Ale