On Mon, 2006-07-03 at 15:17 +0200, Magnus Holmgren wrote:
> >
> > It might also be worth requesting this as a feature. Assuming it's
> > possible to get it with the way that SpamAssassin is called from Exim
>
> The spamd protocol currently offers five modes: PROCESS, which returns the
> full message with added headers, subject to report_safe; CHECK (spamc -c),
> which just returns a yes/no response; SYMBOLS (spamc -y), which returns the
> names of the rules hit; REPORT (spamc -R), which returns the report, as set
> in the SpamAssassin configuration; and REPORT_IFSPAM (spamc -r) which returns
> the report only if the message is considered spam. All except REPORT also
> give the score and threshold.
>
> It currently isn't possible to get the score, list of rules hit, and the
> report in one call, except with REPORT. There is discussion about a new mode,
> which would let the client add the headers according to the configuration.
>
Thanks for this. I played with the spamc/spamd protocol this morning and
tried the SYMBOLS and REPORT commands to see what happens. In our case
though, we configure the 'report' specially and it doesn't include the
SA tests that were used (it's a University policy thing). So only
SYMBOLS showed the tests SA used.
The new spamd mode would do what we want. It could, I suspect, just
return the tests used rather than actually adding the X-Spam- headers.
That way the message itself is left untouched. In our case, we could of
course then add our own X-Spam- headers using the results returned by
spamc (Exim being modified to provide a $spam_status variable as
suggested).
John.
--
---------------------------------------------------------------
John Horne, University of Plymouth, UK Tel: +44 (0)1752 233914
E-mail: John.Horne@??? Fax: +44 (0)1752 233839
