e-mail addresses

[David Relson]

At 08:59 AM 3/6/03, Boris 'pi' Piwinger wrote:

>Hi again!
>
>Today is dumb question day#-)
>
>I was just wondering how much a specific e-mail address
>would make a mail look like ham. So I used bogoutil -w to
>find out. Actually there was not a single hit. So I thought,
>that my address 3.14 at piology.org would show up. It did not.
>Well, bogoutil -d ~/.bogofilter/goodlist.db |grep piology
>should show something, but actually there are only very very
>few entries. Are To and From and the like completely ignored?
>
>pi

pi,

'@' is not allowed in tokens so a typical address, i.e. john at example.com, 
becomes "john" and "example.com".  The easiest way to see what the lexer 
does is a command like "echo john at example.com | bogolexer".  If you want 
just the token printed (without the other messages), use the '-p' (print) 
option.

Additionally, the parser tends to ignore numbers, with the exception of IP 
addresses.  So "3.14 at piology.org" becomes just one token - "piology.org".

If you run "bogoutil -w ... piology.org logic.univie.ac.at" I'd expect you 
to see some large numbers (counts).

When I look at my wordlists I see:

                    spam  good
piology.org           0    16
logic.univie.ac.at    0   117

David