Thanks for your tips. I had found a little how-to in google in the mean time, and had set KMail accordingly: On Thursday 15 May 2003 11:43, Markus Kohli wrote:

I have exactly that combination running. It's just a tad slow from time to time... So, if you have other filters, put them before the spamassassin-filter.

good idea. had it at the top and it was quite slow. most sort for mailing-lists from suse and debian mostly which generally don't seem to have that much spam (at the moment?)

The SA-Filter has to look about like:

Matches the following: <Any Header> contains . (one dot) Pipe through program: spamc -f Uncheck "Stop processing if filter matches"

Add another filter: Matches the following: X-Spam-Flag equals YES

ok, that seems to be better than <any header> <contains> *****SPAM*****

and define an action as you like (e.g. move to Folder "SPAM" or so).

I'm using spamd (Daemon) and accordingly the client. I've found this to be less ressource-hungry than the spamassassin stand-alone-program.

Hm, at the moment I have it calling the spamassassin -P -F 0, and it was slow, but now that I moved the filter to the bottom it is more acceptable. I only received my first personal spam mail today. So at the moment it is ok. But I will have to rethink your setup once the heavy load starts...

Markus

Kostas

On Thursday 15 May 2003 11:40, Konstantinos Georgokitsos wrote:

I have: <any header> <matches regexp.> . (one dot)

are those equivalent? can't remember regexp's at all.

A dot in a regexp means simply any character. Looks like the contains thingy accepts the dot also as a "joker". Or maybe I'm just lucky that each and every mail contains a dot in a header-field... I'll switch mine to the regexp, probably. Thanks for the hint Markus -- Militant Agnostic--I don't know and you don't either!