The string changes where both necessary and correct. There is a difference between bytes and strings and treating them as the same led to so many issues. Thank god I’ve not seen a UnicodeDecodeError in decades.

And the ecosystem agrees.