Thank you for telling me. I hadn’t considered it, as I recently rehauled the backend for messages filtering.
The flow for them is like so now:
Message -> (ex: **hi**)
Process with rich text (ex: **hi** --> <b>hi</b>)
Strip rich text and maintain information: (ex: <b>hi</b> --> hi)
Filter stripped text
-- When requested for a user:
Reapply tags to filtered text (ex: hi --> <b>hi</b>)
Distribute message accordingly to clients
Basically, I used to handle all of the rich text on the client, but now after moving it to the server, the client only has the XML formatted version and doesn’t know that it’s not a regular message. To fix it, I’ll probably have to just cache what the message originally was or something along those lines. Thank you for bringing this to my attention.
That being said, it should still work if you directly add it in the settings menu. All of this was to combat filter bypassing, a problem that I had seen in games using the chat. It was possible to bypass by making colored messages back-to-back with different letters and the chat would have no idea you’re actually swearing.