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

Re: WTH is going on with master's spamassassin?



On Wed, May 19, 2004 at 08:04:50PM -0600, Marcelo E. Magallon wrote:
>  First, 2.60 is old.  Spam detection a la SpamAssassin is effective, but
>  it's also an arms race.  Tests are added, work arounds are added, too.
>  Keeping up with the current version is not just for the sake of having
>  the latest version of everything, it's about keeping up with spammers.
>  Can someone upgrade that?

2.63 has the same tests (although IIRC better scores (?))

3.0.0 will be much better when it's released.
 
>  Second, BAYES_20? BAYES_40? BAYES_10? WTF?  And where's
>  MICROSOFT_EXECUTABLE?  This thing is a virus, but for some reason
>  SpamAssassin isn't seeing the executable that's attached to the
>  message.  My ~/.spamassassin/user_perfs says:
> 
>     score MICROSOFT_EXECUTABLE 7.5
> 
>  it used to say 5.0, but ever since the bayes filter going crazy I
>  bumped it higher.  I'd put it at 100 for all I care, this account
>  shouldn't get MICROSOFT_EXECUTABLEs at all, but if SpamAssassin isn't
>  catching it it doesn't help.

SpamAssassin is NOT a virus checker. Don't try to convince it
otherwise. I agree though that someone should install clamav or
something similar to filter out viruses.

As far as the BAYES thing goes, it looks like your bayesian database
simply thinks that message is ham. This isn't surprising since viruses
don't look like spam.

>  I also bump MIME_HTML_ONLY up to 5 now, if a clueless moron sends an
>  HTML-only message to this account he deserves what he gets.  I'm
>  wondering if that's going to have _any_ effect, since I'm not sure if
>  user's prefs are being read or not.  I don't see anything in
>  master's SpamAssassin configuration that says it isn't.  Someone knows
>  for sure?

AFAICT, it should. (If it's working for MICROSOFT_EXECUTABLE, it
should definitely work for MIME_HTML_ONLY...)

-- 
Duncan Findlay

Attachment: signature.asc
Description: Digital signature


Reply to: