Has definately changed the error message names, possibly the parser. Also check lexer against new tokenizer spec.
Log in to post a comment.