does the X-Bogosity line mess up spamicity calculations?

[Chris Wilkes]

On Wed, Aug 04, 2004 at 11:54:45PM -0300, Trevor Smith wrote:
> On August 4, 2004 9:52 pm, Chris Wilkes wrote:
> 
> > David pointed this out to me a while ago:
> >
> > # s=/tmp/mail.txt
> > # echo "X-Bogosity: No, tests=bogofilter, spamicity=0.000000,
> > version=0.91.4" > $s # echo "Hi There" >> $s
> > # echo "" > $s
> > # echo "Body" >> $s
> > # cat $s
> > # bogofilter -I $s -vvv
> >
> > You can see it skips the X-Bogosity header.
> 
> hmph.
> 
> I'm not able to reproduce any output from the above instructions. I am *VERY* 
> inexperienced with linux in general so bear with me.

No problemo!

> 1) Should there be a double greater than on line 4 so we append instead of 
> overwrite?
> # echo "" > $s

Yep, you're right.  See also David's way with cat <<EOF.  I would of
liked to put that on one like with a
  echo "Hi There\n\nBody" >> $s
but I couldn't get \n to make a newline and so I ended up with that mess
up there.

> 2) bogofilter doesn't do anything with the last line given except generate an 
> error:
> Extra arguments given, first: /tmp/mail.txt. Aborting.
> Perhaps it should be:
> 
> bogofilter -l -vvv < $s

Did you use a *capital* -I in that?  That says "use the next argument
($s) as your email file"

Otherwise you could say 
  bogofilter -vvv < $s
like you had (no -l in there)

> 3) this generates a report, shows that it's not spam, but I don't see anything 
> that illuminates anything about how it processes or doesn't process the 
> X-Bogosity line:
> 
> X-Bogosity: No, tests=bogofilter, spamicity=0.520000, version=0.92.2
>                                       n    pgood     pbad      fw     U
> "head:Body"                           0  0.000000  0.000000  0.520000 -
> N_P_Q_S_s_x_md                        0  0.000000  0.000000  0.520000
>                                          0.017800  0.520000  0.375000

Well *if* I would of posted the ">>" correctly above you would of seen
about the same thing.  This is a report of what bogofilter sees as the
tokens in your email.  Granted this one's pretty dull so you should do
it with a normal email with the "X-Bogosity" header in there.  You'll
see that it doesn't show up in the report, while normal headers (like
"from:blah") do.

Chris