For now everything should look identical. However this moves to using a model for the content of the message and is intended to lay the foundation for improved message content representation, e.g. splitting up a text message in multiple sections and using different delegates for things like code and quotes.
This MR deals with only sending threaded messages. Showing threads will turn up in a follow up. This allows you to start a new thread by clicking reply in thread to a normal message.
You can also do a threaded reply to a threaded message in the main timeline at the moment because those messages aren't shown in a separate thread timeline yet but will be in future.
Move the functionality to cache the contents of a chat bar from the room directly and to a new ChatCache object. This works pretty much the same with a few extra check and balances, this also made it easy to put a test suite around the functionality so I did. The current functionality should be identical to what exists.
This is in prep for threads which will require managing even more caches if we create one per thread.
