Unsophisticated but surprisingly effective new spam attack

/[一-龠]+|[ぁ-ゔ]+|[ァ-ヴー]+|[々〆〤ヶ]+/u

I’m not really familiar with the regex you are using but all the expressions in the square brackets except for the first one are Japanese (though Japanese uses Chinese characters - kanji - too).
 
I’m not really familiar with the regex you are using but all the expressions in the square brackets except for the first one are Japanese (though Japanese uses Chinese characters - kanji - too).
It blocks ranges of unicode characters. All of the characters used in these typed dialects occur in sequence in the unicode character standard.

The square brackets with the - mean "from character A to character X". So [一-龠] means match any character that occurs in the set between (and including) and .

The + means "at least one occurrence of a character in the range" and the | is an "or" so I don't have to write these range matches out as separate regexes. Though you could if you wanted to:

Code:
/[一-龠]+/u
/[ぁ-ゔ]+/u
/[ァ-ヴー]+/u
/[々〆〤ヶ]+/u
Would be equivalent range matches, written as separate rules for each range.

The /u means "do unicode regex matching".

It does potentially block users trying to use these unicode characters in a legitimate way on the forum. But so be it. Greater good and all that.
 
I've been getting hammered by this the last few days, so I wrote this add-on:

It allows you to set the check spam settings by "days since registered" as well as the normal "messages posted".

You don't even need to set the number of days very high, I have mine set to 2 - as that gives me enough time to see the stupid two character test messages and spam clean the offender before they really run riot.
 
Top Bottom