On Thu, Mar 11, 2010 at 8:46 AM, Martin Gregorie <mar...@gregorie.org> wrote: > That 'male enhancement junk' advert may well contain something that > could be the basis of an additional rule - don't omit *anything* in > future, at least until you understand how to write custom rules. > Spammers often use an algorithm to generate their destination websites. > This algorithm often generates patterns that can be matched with an SA > rule. However, it may be reasonable to obscure your own and/or your > user's address, e.g. by changing it to u...@example.com.
I did omit my user and domain. That is a test domain I own but doesn't route nor is it in production. It is not the domain I am actually using live. > In fact, when I ran your message through SA 3.3.0 the standard rules > gave a score of 5.2 even without the body text. That is enough to treat > it as spam if you were using the default required score. Why did you > change your required score to 6.3? That is a pretty specific value. I don't understand this. If you ran my exact same message through SA and got a score of 5.2 (omitting the actual URL), how come my headers show a score of 0? X-spam-status: No, score=0.0 required=6.3 tests=FREEMAIL_FROM, RCVD_IN_DNSWL_NONE,TVD_SPACE_RATIO,T_DKIM_INVALID autolearn=ham version=3.3.0 As for me, I don't think I changed any values but perhaps my memory is not serving me well. I checked my /etc/mail/spamassassin/local.cf file and I show: rewrite_header Subject *****SPAM***** required_score 6.31 report_safe 1 use_bayes 1 use_bayes_rules 1 bayes_auto_learn 1 Is this wrong? Should I change this to reflect something else? I think this was the default when I installed and configured SA on my server. > Since this spam was sent via Yahoo webmail you could also try passing it > to ab...@yahoo.com, though if they're still routing ab...@yahoo.com > to /dev/null that won't do any good apart from maybe making you feel > better.
