[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: dueling banjos

This one time, at band camp, Andreas Tille said:
> BTW, the only fals positives I've got after tuning my user_prefs where
> caused by AWL and I have no idea what this means.  I have got a fine
> Mail from a mailinglist (userlinux) marked as:
>    Content analysis details:   (41.8 points, 5.0 required)
>     pts rule name              description
>    ---- ---------------------- --------------------------------------------------
>    -4.9 BAYES_00               BODY: Bayesian spam probability is 0 to 1%
>                                [score: 0.0000]
>     1.5 PLING_PLING            Subject has lots of exclamation marks
>      45 AWL                    AWL: Auto-whitelist adjustment
> I only found sparse documentation about this "Auto"-whitelist feature
> which sometimes adds and sometimes removes scores.  Even if it is off topic
> for this list, could you give some hint (perhaps to some relevant docs)
> how to prevent these false positives.

Auto-Whitelist does a decaying average, IIRC, of the scores of all the
emails from that address.  So, if I sent you an email with the GTUBE
line in it, that would score 1000.  My next message, which by itself
scores 0, now scores 500, and so on.  I find this feature not very
helpful, but YMMV.

It also seems that you have some Whitelist_from: addresses in a
preferences file, and that is messing up your scoring of things.  A
matched whitelist address scores something like -100, making it very
difficult to be picked up as spam.  I think there are some example
whitelist_from's in /etc/spamassassin/local.cf - perhaps you uncommented
them, and that's what doing this?

|   ,''`.					     Stephen Gran |
|  : :' :					 sgran@debian.org |
|  `. `'			Debian user, admin, and developer |
|    `-					    http://www.debian.org |

Attachment: pgphhEwjESbRh.pgp
Description: PGP signature

Reply to: