On Tue, 2011-11-22 at 09:26 -0500, Bowie Bailey wrote:
> On 11/22/2011 3:25 AM, ercibrest wrote:
> > X-Spam-Status: No, score=-3.9 required=6.0 tests=BAYES_00,FROM_12LTRDOM,
> > RDNS_NONE,TVD_SPACE_RATIO,TVD_SPACE_RATIO_MINFP shortcircuit=no
> > autolearn=ham version=3.3.2
^^^^^^^^^^^^^
And there is the culprit for the Bayes problem.
> > X-Spam-Report:
> > * -4.7 BAYES_00 BODY: Bayes spam probability is 0 to 1%
> > * [score: 0.0000]
> > * 0.8 RDNS_NONE Delivered to internal network by a host with no rDNS
> > * 0.0 TVD_SPACE_RATIO TVD_SPACE_RATIO
> > * 0.0 TVD_SPACE_RATIO_MINFP TVD_SPACE_RATIO_MINFP
> > * 0.0 FROM_12LTRDOM From a 12-letter domain
According to the RDNS_NONE score, this is sore-set 3, Bayes and network
tests enabled. For auto-learning, the non-Bayes score-set 1 will be
used, with a score for RDNS_NONE even slightly higher. The other rules
are irrelevant, and Bayes of course is not considered.
The default auto-learn threshold for ham is 0.1. This message should
never have been automatically learned as ham.
Your auto-learn threshold settings are terribly messed up. (Possibly the
scores for score-set 1, but that's much less likely.)
> The only thing I see here is that your Bayes database appears to be
> mistrained. It is scoring this message as BAYES_00, which means
> "definitely not spam". At a minimum, you need to manually learn any
> messages that are being scored the wrong way with Bayes. If this is
> happening with all of your spam, you may want to just delete the Bayes
> db completely and start over.
--
char *t="\10pse\0r\0dtu\0.@ghno\x4e\xc8\x79\xf4\xab\x51\x8a\x10\xf4\xf4\xc4";
main(){ char h,m=h=*t++,*x=t+2*h,c,i,l=*x,s=0; for (i=0;i<l;i++){ i%8? c<<=1:
(c=*++x); c&128 && (s+=h); if (!(h>>=1)||!t[s+h]){ putchar(t[s]);h=m;s=0; }}}
