On December 01, 2003 01:06 am, Aaron Young wrote:
> On Mon, 1 Dec 2003, Pedro Sam wrote:
>
>
> > My messages are in mbox format, so I report/learn spam by:
> >
> > "formail -s spamassassin -r < SPAM_MBOX"
> >
> > But when I dump the contents of the bayes database, I found many tokens
> > unique
to SA headers and markup.
>
>
> I found some header stuff but no other markup.
>
>
> > As a compromise, I now proprocessing spam with an explicit "spamassassin
> > -d",
before I pass it to "spamassassin -r". I'm just wondering if
> > anyone else is experienceing this, and maybe file a bug report.
Try passing the dump of sa-learn to "sort -n" ... It'll present the most hammy
token first, then the most spammy tokens. When tip me off was that I saw
"BAYES_99" as the most spammy token in the database, which is odd, because I
don't pass SATalk into sa-learn.
> Does spamassassin -d actually work for you? Ever since I upgraded to 2.60
>
the -d option doesn't remove any markup. It just seems to pass the
> message through unaltered.
spamassassin -d works fine for me... it just takes up more cpu to start 2
instances of perl per message reported. Yes, I'm also using 2.60 from CPAN
--
Electrocution, n.:
Burning at the stake with all the modern improvements.
-------------------------------------------------------
This SF.net email is sponsored by: SF.net Giveback Program.
Does SourceForge.net help you be more productive? Does it
help you create better code? SHARE THE LOVE, and help us help
YOU! Click Here: http://sourceforge.net/donate/
_______________________________________________
Spamassassin-talk mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/spamassassin-talk
