New header token tagging

[David Relson]

On Thu, 25 Sep 2003 10:32:48 -0700
"Greg McCann" <greg at cambria.com> wrote:

> On 9/25/2003 at 1:00 PM David Relson <relson at osagesoftware.com> wrote:
> 
> ...
> >Your conclusions all sound correct to me.  What messages have you
> >been using to keep bogofilter's training up to date?  
> ...
> 
> Known spam received by several spamtrap email addresses is registered
> with "bogofilter -s".
> 
> I register my own incoming email (about 90% ham, 10% spam) with
> "bogofilter -uepl".
> 
> Until upgrading from 0.13.6.2 to 0.15.4, this has worked fine.
> 
> On further investigation, it seems that "-u" updates are no longer
> making any updates to the databases spamlist.db or goodlist.db, though
> "-s" is still updating spamlist.db and -n is still updating
> goodlist.db.  Whether or not the text is in the header field seems to
> be irrelevant.
> 
> If I revert to 0.13.6.2, "-u" again updates the appropriate database.
> 
> 
> Greg

Greg,

"-u" means to score the message and then register it if it's ham or
spam.  Is it possible that you've got two wordlist.db files in different
places?

I just tested "-u" and  it updates as it should.

### Here's my four line test message

[relson at osage src]$ cat msg.david
>From david
To:  david

david


#### Here's the output of "bogofilter -v < msg.david -vv" which shows
the current counts for the tokens in the message.

X-Bogosity: No, tests=bogofilter, spamicity=0.000209, version=0.15.4
                                     n    pgood     pbad      fw     U
"head:From"                        193  0.003431  0.000000  0.000022 +
"to:david"                          30  0.000515  0.000033  0.059698 +
"david"                            119  0.002026  0.000163  0.074601 +
"head:david"                         0  0.000000  0.000000  0.415000 -
N_P_Q_S_s_x_md                       3  1.00e-00  1.41e-05  2.09e-04
                                        1.00e-02  4.15e-01  0.100

#### This next line should change "n" (total spam + ham count) and
"pgood"

[relson at osage src]$ bogofilter -u < msg.david

#### Here's the -vvv output after the change.

[relson at osage src]$ bogofilter -v < msg.david -vv
X-Bogosity: No, tests=bogofilter, spamicity=0.000008,
version=0.15.4.cvs.CVStime_20030925_153854
                                     n    pgood     pbad      fw     U
"head:From"                        194  0.003448  0.000000  0.000021 +
"head:david"                         1  0.000018  0.000000  0.004109 +
"to:david"                          31  0.000533  0.000033  0.057824 +
"david"                            120  0.002044  0.000163  0.074002 +
N_P_Q_S_s_x_md                       4  1.00e-00  7.38e-07  7.64e-06
                                        1.00e-02  4.15e-01  0.100