Sure, which is why I suggested to make the task complex enough so that the "usual" *nix text tools aren't enough, or significantly more complex to use then a re

Sure, which is why I suggested to make the task complex enough so that the "usual" *nix text tools aren't enough, or significantly more complex to use then a real parser. I mean it's a basically a rehash of "how do I parse HTML with regexes".