On 10/19/2011 10:00 PM, Alex wrote: > Hi, > >>> I have a fedora15 system with spamassassin-3.3.2 and clamav-0.97.2, >>> and also using the clamav-unofficial-sigs. and I've just realized the >>> score for catching one of the listed domains is only 0.2. >>> >>> X-Spam-Status: No, score=3.444 tagged_above=-100 required=5 >>> tests=[AV:INetMsg.SpamDomain-2w.t67f_com.UNOFFICIAL=0.1, >>> BAYES_50=0.8, >>> HTML_MESSAGE=0.001, HTML_MIME_NO_HTML_TAG=0.377, >>> MIME_HTML_ONLY=0.723, >>> RCVD_IN_BRBL_LASTEXT=1.449, RELAYCOUNTRY_LOW=0.5, >>> RP_MATCHES_RCVD=-0.504, SPF_HELO_PASS=-0.001, SPF_PASS=-0.001] >>> autolearn=no >>> >>> Is that typical? Can you recommend a more suitable score? >> >> Spamassassin scores are quie optimized. But you can fine tune the >> scores based on your own requirements after observing the trend for a >> few days > > I think I assumed that you knew too much about my level of > understanding. I'm familiar with local.cf and building my own SA > rules. > > However, I don't know where the original definition of the clamav > rules are listed. Where is that 0.1 actually defined?
Those look like scores from amavisd-new, which has special code to turn a clamav spam detection into a SpamAssassin score. (Normally, clamav is separate from SA; any detection results in a reject.) In amavisd-new, the score added (or whether to just go straight to quarantine) is controlled in the amavisd.conf file. See the amavisd-users list or docs for details. There are likely other filters or milters that do similar things. -- Noel Jones _______________________________________________ Help us build a comprehensive ClamAV guide: visit http://wiki.clamav.net http://www.clamav.net/support/ml
