I guess, that's the whole point of the blogpost: when you try to validate email addresses with a regex, you are screwed with unicode domains. It is not clear if

I guess, that's the whole point of the blogpost: when you try to validate email addresses with a regex, you are screwed with unicode domains. It is not clear if you should let them pass, only let them pass if they are properly encoded or not let them pass at all (as the software behind the validator is clueless about them either way). Building the DFA yourself gives you more control.