Attempted to make it clear, however, that the prior discussions were
really in a different context, namely when we were trying to resolve
the question "does \xXX denote a code *unit* or a code *point*."

This RFC is proposing a way to side-step that question entirely, at
least in the context of string literals.  (I'll leave aside the
question of if we are really side-stepping it, given our treatment of
byte-string literals.)

These admittedly weaken my "source of confusion" argument.
Nonetheless, I still think the case for `\xXX` for `XX > 0x7F` is
weak.