[Pan-users] Re: Filtering non-Latin script postings

[Jack Spaar]

On Fri, 15 Jun 2007 16:32:30 -0700, JCA wrote:

> Sometimes I come across postings in Arabic or some other languages with
> scripts other than the Roman one. How can I filter these postings out,
> assuming that the subject line (or the message body) is entirely written
> in one of such scripts?

I use the following entry in my scorefile to ignore subjects that contain 
either only non-roman-alphabetics, or all non-alphabetics preceded by 
"re:".  It doesn't filter subjects with a mix of roman and non.  
Tweak the newsgroup appropriately if you try it.

--Jack

%BOS
%Filter out anything containing only non alphabet in subject line
%Note: pan is case insensitive for character matching.
[gmane.linux.usb.*]
Score:: -9999
        % Any of the below must match
        ~Subject: [a-z]+
        {:                                      % ALL below must match
          Subject: ^Re:
          ~Subject: ^Re:.*[a-z]
        }
%EOS