option names
Eric Seppanen
eds at reric.net
Mon Sep 16 19:39:22 CEST 2002
On Mon, Sep 16, 2002 at 07:57:34AM -0400, David Relson wrote:
>
> I've been thinking about option names and list names. In ESR's initial
> release, he had options 's' and 'n' for registering spam and
> non-spam. Several releases later, he added 'S' and 'N' for moving words
> from one list to another. 'n' and 'N' were renamed to 'h' and 'H'. Now
> Eric has added 'i' for ignore-lists.
No, I added "-c <filename>" for plaintext-to-db list conversion (an
unfortunate requirement for ignore-list right now). But I'm not attached
to it, and I'll use whatever, "--convert" is fine if that clears up
the alphabet pollution.
> My question is: How do we handle multiple lists/filters?
>
> We can't keep adding single letter flags, because that quickly becomes
> unmanageable and because no two people are likely to want the same
> lists. One way would be for bogofilter to have a config file which
> contains info on the lists, e.g. list name, file name, weight, good/bad,
> etc. Maintaining the lists might be done with add/change/delete command
> line flags which require a list name (matched against the config file).
For multi-list, I'd always assumed a .bogofilterrc file would be needed.
The idea is something vaguely like this:
# add_list_db file weight
add_list_db /var/bogo/spamlist.db 0.8
add_list_db /var/bogo/hamlist.db -1.5
add_list_db myspamlist.db 1
add_list_db myhamlist.db -2
# add_ignore_db file weight
add_ignore_db myignore.db
I don't know what the best way of parsing such a file would be.
For summay digest subscription: bogofilter-digest-subscribe at aotto.com
More information about the Bogofilter
mailing list