Parsing Unicode without normalization (as JoshTriplett's solution) is another sort of mádness. (Or is that mádness?) In this case there's an oracle that tells

Parsing Unicode without normalization (as JoshTriplett's solution) is another sort of mádness. (Or is that mádness?) In this case there's an oracle that tells me if I have the right answer, so there wasn't need for anything more rigorous. (Edit: looks like my LATIN SMALL LETTER A WITH ACUTE vs. my LATIN SMALL LETTER A + COMBINING ACUTE ACCENT didn't make it all the way through the system. Haven't researched to figure out where the normalization occurred.)