Very interesting! I'm curious about the parser's error reporting for malformed input -- does it use Irony's default error reporting, or a custom solution? Are error reports good, bad, or lacking in any way?