Rule based spam filter.

[weedy]

04:11:38 <!SmugLeaf> like
04:11:45 <!SmugLeaf> I want iptables for spam
04:12:01 < dx> hmmm HMMM
04:12:09 <!SmugLeaf> X regex in Y field worth Z points
04:12:09 < dx> that sounds fun
04:13:37 <!SmugLeaf> and then one line in config.py that say spam_thresh=20
04:17:20 <!SmugLeaf> but like this also kinda needs huristics, in the perfect world I would like to see what my rule matches
04:17:49 <!SmugLeaf> actually, running the rule against the global trash probably isn't hard
04:17:59 <!SmugLeaf> hmm
04:18:06 < dx> spam doesn't go to the trash right now
04:18:11 < dx> but yeah, good point there
04:18:34 <!SmugLeaf> dx: I would wright rules for shit that gets through
04:18:43 <!SmugLeaf> and in that case
04:18:54 <!SmugLeaf> a drop down would be better
04:19:03 <!SmugLeaf> trash or <list of boards>
04:19:48 < dx> or individual posts too
04:19:50 <!SmugLeaf> not the whole db, but like count(100) DESC
04:20:10 <!SmugLeaf> I guess
04:20:14 < dx> this way i think you don't even need to store any extra info, only send spam posts to trash

We probably need logging of matches.
Also a way to match a string in multiple fields:
13:01:23 < CacheRAM> dx: weedy: suggestion - if you parse the text field for a URL, and find the same URL in the email / name field, it's likely spam. Instaban.