As far as I know the forget option removes any tokens added to bayes for that message regardless of whether it was trained in as ham or spam. If you added it as spam by mistake then you would re-add it as ham and vice versa (after running forget).
On Wed, 19 Jul 2006 07:08:06 +0000 (UTC), martin <[EMAIL PROTECTED]> wrote: >Nigel Frankcom <nigel <at> blue-canoe.net> writes: > >> http://spamassassin.apache.org/full/3.0.x/dist/doc/sa-learn.html >> >> http://spamassassin.apache.org/full/3.1.x/dist/doc/sa-learn.html >> >> On Wed, 19 Jul 2006 05:37:29 +0000 (UTC), martin >> <martinnitram <at> excite.com> wrote: >> >> >just wonder, how can make sa-learn/spamassassin to unlearn a message? thx >> > > > thx, and one more question, is the using of --forget parameter prevent >spamassassin trace a message to ham and spam at same time? >
