List Conversion

Eric Seppanen eds at reric.net
Thu Sep 19 21:45:23 CEST 2002


On Thu, Sep 19, 2002 at 03:13:07PM -0400, Doug Beardsley wrote:
> On Wed, Sep 18, 2002 at 09:56:28PM -0700, Adrian Otto wrote:
>  
> > This is actually one of the items on our TODO list. We plan to provide a
> > utility that will be able to import and export word lists from a standard
> > text format. Note that we will not support text-only wordlists for
> > operational use because of the performance penalty required for using them,
> > but there will be a way to take your wordlist from any version of bogofilter
> > and export it so that it can be imported into the current version.
> 
> Good.  This is exactly what I had been thinking.  I totally agree that
> text formats aren't good for the working lists.

Except for lists that are hand-maintained, like ignore lists.  I think one 
of the supported file formats should be plain-text, because it's a pain in 
the ass to go create a new database just because I want bogofilter to 
start ignoring the word "message-id" or "the".

That reminds me: for those concerned about database performance: one of 
the easiest ways to cut down on the number of word lookups is start 
picking out the top several hundred "neutral" words (close to 0.5000) and 
put them in an ignore list.  The ignore list takes precedence over the 
other lists so you can skip lots of db searching.

Oh, except that would mean that you'd all need my ignore-list patch, now, 
wouldn't you :)  My current patches are slipping out of date again; it'll 
be Monday at the earliest before I can bring them up to date.

For summay digest subscription: bogofilter-digest-subscribe at aotto.com



More information about the Bogofilter mailing list