The contentFilter plug-in today have a combined black- and whitelist. For instance, by specifying
you will blacklist all “cock”-ish words like “cock”, “cocksucker”, “cc1oo2cc3kk” and so forth, but leave the friendlier words “cocktail” and “cockpit”.
My problem is that I have a huge blacklist containing several thousands patterns (using another chat-server we are migrating from). The number of whitelisted words are much smaller, but it is still problematic to find what regular expression in the blacklist to pre-fix with the corresponding whitelist word (e.g., you need to find .*c+[^a-z]*o+[^a-z]*c+[^a-z]k+. and add (?!cocktail) (?!cockpit) to it, which is manual and error-prone).
One solution would be to explicitly use blacklist and whitelist. The whitelist will be checked after the current check, and override if matching. Is this something more people would like to see in future releases?
Another questions is regarding the size of the “blacklist”. Is there another limit on it than provided by Java String (e.g., something imposed by the property handling of Openfire.).