enhancement: -v, -vv, and -vvv

[David Relson]

At 11:19 PM 10/21/02, Allyn Fratkin wrote:

>in versions before 0.7.5, -v used to show the words that bogofilter was
>using along with the calculation to get spamicity.
>
>i notice that in 0.7.5 that -v now only shows the X-Bogosity header
>and -vv is required to show the word lists.  i have no problem with this,
>except -vv also spews out a lot of other debugging information that i
>don't want (e.g., looking up "word" in the good list, etc).
>
>would it be possible to move this extra debugging info to -vvv and leave
>-vv essentially the same as -v used to be without a lot of extra output?
>
>i can work up a set of patches if everyone agrees this is a
>worthwhile thing to do.

Allyn,

Check on the cvs version before doing anything and then double check with 
me (directly).  I've done some work on rationalizing what different levels 
of verbosity do, since the feature freeze for 0.7.5.

1) Using -p or -v produces an X-Bogosity header.

2) Using -pv or -vv also output the 15 extrema tokens and their 
probabilities to stdout.  Hence they will appear in the message after the 
headers.

3) Using -pvv or -vvv expands #2 with a couple more columns giving details 
of the probability calculation.

I have the verbose displays of getting/setting word values turned off.

Note: #2 and #3 are particular to using the Graham algorithm.  The Robinson 
algorithm computes values for many more tokens and would generate much more 
data if treated in the same way.  I am still thinking of the best way to 
deal with this.

David