It looks like it's also rot13'ing the the HTML codes in some weird way.

It looks like Jim was trying to do the "exple[i][/i]tive" trick, which at some point the board turned into "exple<i></i>tive", which for some reason got turned into "exple&gt;i&lt;&gt;/i&lt;tive", which the new code turned into the crud that appears there. I don't know why it multiplies itself, though.

If I understood what was changing it into the "&gt;"-style character entities, I might have a better idea how to avoid it. Maybe it's holding it that way in the database where they get changed back to "<"s when written out. The obvious thing to do would be to avoid changing HTML character entities, but I don't see a really efficient way to do that at the moment. You can't do it with your matching code very easily, and you can't do it with the javascript very easily, not that I think that's the place where it matters anyway.

I think you'll somehow have to match all the stuff inside the spoiler tags that isn't an HTML character entity, but you'll have to do that in a loop, and, it would seem, some sort of recursive addressible regex. I'll leave that up to someone more familiar with PHP.

That colored-text thing is looking more attractive all the time, huh?