Thanks for your help. I thought that maybe the message had been lost amongst the auto reply complaints.
I have used what you supplied. Again thanks for the help. On Sat, 2005-06-18 at 04:15 +0200, guenther wrote: > > I notice that in my /usr/sbin folder there are 3 clam related files. > > > > 1..clamav-milter > > 2..clamd > > 3..clamsmtpd > > > > I am trying to create a filter for evolution to scan for viruses. I was > > able to create a filter for spam by pointing to spamc. I presume it is > > either one or two above. But which one does the work? > > Neither of them. They are not intended to be run by a user anyway. > > Have a look at the ClamAV related executables in your /bin directory. > Oh, and *please* have a look at the documentation... > > 'clamscan' can scan data streams, which is necessary in your case. > Unfortunately there is no client for the 'clamd' daemon provided AFAIK, > that takes data streams -- which would speed up scanning. > > A Filter condition like this works for your purpose: > > "Pipe to Program" "/bin/clamscan --quiet -" "returns" "1" > > The dash is necessary to use it on data streams, the --quite option > prevents scanning reports on STDOUT. See 'man clamscan' for more > details. > > > A warning about Evolution Filters and STDOUT: > > Although a quick test even without --quite just did work for me, I > vaguely remember a bug at some time, that output on STDOUT may "rewrite" > the mail. Did not do this for me. You should test this anyway, before > running this on valuable mails... > > > A related note: Evolution 2.2.x comes with SA integration. That is, > there is a convenient option to use SA to filter for SPAM. It uses > spamc/spamd if available, and there are buttons to train Bayes by > explicitly learning mails. There even is a "Junk Test" Filter. No need > at all to create a filter for this purpose on your own... > > ...guenther > > _______________________________________________ http://lurker.clamav.net/list/clamav-users.html
