spam.assassin.prefs.conf real file location

Anthony Peacock a.peacock at CHIME.UCL.AC.UK
Wed Nov 23 11:13:11 GMT 2005


> Julian Field wrote on         Tue, 22 Nov 2005 18:09:41 +0000:
> > If we accept that we are having 1 file and a link to it, which
> > should be the real file? The
> > <mailscanner_etc>/spam.assassin.prefs.conf will not be read directly
> > by MailScanner, the <sa_site_path>/ will be read by
> > SpamAssassin as part of its normal initialisation. 
> >  
> > <mailscanner_etc>/spam.assassin.prefs.conf is the real file? 
> > or 
> > <sa_site_path>/ is the real file? 
> >  
> > I prefer the first one as it makes the installation a lot easier to
> > do automatically.
> Reading the sa-talk list as well I find that people quite often get
> confused with spam.assassin.prefs.conf hanging around in a different
> area than /etc/mail/spamassassin. And they train Bayes etc. and then
> later ask on the list why it doesn't kick in. Usually, because they
> trained with the wrong user or because SA was using the wrong prefs
> file (it's own). Same goes for running spamassassin against a mail.
> Usually this will result in different results then via MailScanner.
> They don't know that they have to specify the prefs file if they use
> SA or any of its tools outside of MailScanner. I use the
> /etc/mail/spamassassin/ file as the spamassassin prefs file
> for MailScanner in all my installations. That completely eliminates
> all these problems. I don't see a reason why it should not be used. If
> you provide a "well-behaving" prefs file you should rename the old
> (I think SA scans only files with extension of *.cf in that
> directory) and put yours there. This provides for a clean setup where
> everyone knows where the configuration resides.

Please do NOT delete any existing files.

That would instantly break my SA configuration, which has a lot of 
stuff in to do with SQL Bayes settings, etc.

I vote for option one above, but leave anything else in the 
site_rules_path alone.  By naming the MS file, it will 
be processed after the file anyway.

Anthony Peacock       
CHIME, Royal Free & University College Medical School
All sweeping generalisations are false.

------------------------ MailScanner list ------------------------
To unsubscribe, email jiscmail at with the words:
'leave mailscanner' in the body of the email.
Before posting, read the Wiki ( and
the archives (

Support MailScanner development - buy the book off the website!

More information about the MailScanner mailing list