Should contain a few words on * Chevrotain lexing (first fit, etc) * How to create a multi-mode lexer * How to create tokens with custom lexing patterns