Of course I'm addicted to using SpamLookup 2.0, but its Keyword Filter doesn't recognize regular expressions with multi-bytes strings and Unicode properties/blocks/scripts which are supported by Perl 5.8. Therefore it's not easy to train SpamLookup for rejecting comment/trackback spams in foreign languages, especially in asian languages.
The following is a patch for enabling Unicode support in SpamLookup 2.0.
If you are using a Linux box, it is easy to apply. Just download or copy it into your MT directory and type as follows:
patch -p0 < SpamLookup2.0-encode.patch
Once you applied this patch, you could write regular expressions enpowered by Unicode support, as Keyword Filter rules.
For example, to reject comments/trackbacks with Hiragana strings, just as follows:
Or to accept only Latin-1 comments/trackbacks, you can do as follows: