some questions about bogofilter 0.13.6&0.15.7

[Mike Lykov]

Hi all.

I install bogofilter on my mail server, when it's version was 0.13.6, and i 
create wordlists from my emails (bogofilter -s, bogofilter -n).

1. Now i upgrade my version to 0.15.7, and I read in mans what wordlist format 
was changed :

 Since  version  0.15.4  bogofilter  has been adding special tags to ALL
       header tokens. Prior to 0.15.4, only tokens in "To:", "From:", "Return-
       Path:",  and "Subject:" lines were given special prefixes. Now the rest
       of the header tokens are given a "head:"  prefix.

How i can re-create wordlists for using it ? Working with this must be more 
accurate, isn't it ? 
Or i can use -H option - it will be same as above ?

2.   In man page: 
 PARSING OPTIONS
       Bogofilter has three special parsing options .....

Where i must use it? In SYNOPSIS it's do not described, but i want to use it 
to be more accurate....

3. Why bogofilter register in wordlists pieces of base64-encoded parts ?
Here the sample :
--Boundary-02=_/Weu++wGDnxQ1AU 1 20030708                                                                  
--Boundary-02=_1gKs+PfUum7GU4d 2 20030708
--cqxox3fnlpmgjstp 1 20030526                                                                              
--cqxox3fnlpmgjstp-- 1 20030526  
--TB36FDmn/VVEgNH/ 2 20030708                                                                              
--TB36FDmn/VVEgNH/-- 2 20030708   

Often i see that the letter with attached file and a little piece if text 
(two-three words) classified as spam, but attach can't be spam! 

I take this letter and classify it by hand (bogofilter -Sn), then it passed 
normally....

I see ver 0.15.8 working with parsing MIME-parts of letters - can it fix this 
annoying feature?

I think bogofilter must not classify attaches at all! (same about headers .. 
%)

-- 
Mike, Samara, Russia
registered linux user #315334
jabber id: combr at jabber.ru