TODOS : For the future version Support many encodings (Only UTF-8 for now...) Make a lazy lexical analyzer with a token stream