Special characters in #hashtags

The original way I did it used Regex, which is about a zillion times better than the current string-based implementation (and would fix this problem) except that some people had problems with it due probably to server configuration issues. I'd like to switch back to Regex entirely, but I will probably just add an option to use it instead.

I'm having some trouble getting the regex to work with unicode. I opened a forum topic on a PHP Regex board though, and once I get an answer there the rest is fairly trivial.

I just committed a fix to CVS that completely switches to Regex. I didn't leave an option for the string-based method because it was too restrictive. The new parsing:

• Allows special UTF8 characters in hashtags and usernames.
• Corrects a problem where a status like "#the #therapy" would match "#the" twice and "#therapy" never.
• Allows surrounding hashtags and usernames with square brackets to support tags with word-break characters. For example, [#hello world] now matches the tag #hello world where it previously would only have matched #hello.

[#hash-tag] works, but #hash-tag doesn't. I'm going to keep it that way since a hyphen is a word-break character and Twitter doesn't accept hashtags with hyphens either.

Actually, #hash-tag seems to work too, but I think Views might have a problem with the hyphen character in arguments or something like that... will look into it.

Yeah, it turns out that Views has a bug with taxonomy term name arguments where the term name has a hyphen in it. There's not much I can do about that, unfortunately.

I have no idea what the actual problem is, I just asked on IRC. There's probably already an issue open in the Views queue, and if not that's likely the best way to go.