Reminds me of writing my own csv parser that implemented escapes properly. The one everyone else went with of course was written in regex, so it was faster… But broke if there were escaped newlines.