Peter,

> That is excactly the problem. I do not have different levels set for
> different users. They are all the same, getting their seetings from a
> mysql db as described in the READMEs.
>
> I just looked a the amavisd log, as you suggested. The problem seems to be
> the penpal function I introduced, when I needed it for the bounce-kill
> feature. The one who did get the mail had earlier written a
> mail to the originator of the mail, that was unfortunately declared spam.
> Thus the person, who got the mail resulted in fewer points (5.1), but
> that only counted for him and not the other two, their rating created
> (8.269)

Ah, yes, that explains it.

> What is weird though, is that it gives the same number of points for both
> cases (bottom of the lof excerpt, where it says "Passed SPAM" and "Blocked
> SPAM".
>
> I guess this is just an unfortunate event, or is there something I can do
> about this? I have in the meantime thrown out the FUZZY_OCR check, because
> it doesn't seem to make much difference at the moment, and I might have
> had the score set too high.

PenPalsSavedFromKill 8.269-3.160,
  <[EMAIL PROTECTED]> -> <[EMAIL PROTECTED]>

[...]
Passed SPAM, AM.PDP-SOCK [62.64.120.200] [85.223.218.20]
  <[EMAIL PROTECTED]> -> <[EMAIL PROTECTED]>,
  ... Hits: 5.109, ...

Blocked SPAM, AM.PDP-SOCK [62.64.120.200] [85.223.218.20]
  <[EMAIL PROTECTED]> -> 
  <[EMAIL PROTECTED]>,<[EMAIL PROTECTED]>, 
  ... Hits: 5.109, ...

The only problem here is a misleading log entry, claiming 'Hits: 5.109'
for the 'Blocked SPAM' entry for two recipients, even though the
actual score for those two (no pen pals match) was 8.269.

If you would use per-recipient log (keeping $log_recip_templ at
its defaut value, not setting it to undef), you would receive
three log entries, one for each recipient, and the scores would
be logged individually per recipient.

The more usual $log_templ top-level log entry tries to summarize
results (e.g. score in this case) across all recipients, and in
this case reported a minimal score value for all three recipients.
Admittedly the 'Blocked SPAM' could do better and only summarize
the score across the two blocked recipients. It seems more work
than it is worth to get a small improvement there (it would still
be wrong/misleading if the two recipients obtained different scores
for some other reasons (soft-whitelisting, their own pen-pals hit, etc).

If per-recipient scores in the log are essential, keep the
$log_recip_templ enabled, and perhaps disable the $log_templ.

  Mark

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
AMaViS-user mailing list
AMaViS-user@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/amavis-user
AMaViS-FAQ:http://www.amavis.org/amavis-faq.php3
AMaViS-HowTos:http://www.amavis.org/howto/

Reply via email to