I'm having a problem that I don't understand. After wrangling with the unicode conversion for a while, I have results that are mostly right, but a couple of characters are off.
'\xe3\x82\xbf\xe3\x82\xb3\xe3\x82\xbf\xe3\x82\xa2' should equal
'\xe3\x82\xbf\xe3\x83\xa2\xe3\x82\xbf\xe3\x83\xaf'.
Specifically, what should be \x83 is being encoded as \x82, even though there is a properly encoded \x82 character. The majority of the characters being correct makes me think my logic is correct, but I'm still encountering a unicode bug somwhere. Any insights would be greatly appreciated!

I agree! There should probablt be a special case for that happening, wouldn't be too dificult to fix.