Okay, we're not using a thread to load chunks anymore; just do 'em inline.  Performance feels identical and it's far less complicated this way
1 file changed