gh-96268: Fix loading invalid UTF-8 (GH-96270)

This makes tokenizer.c:valid_utf8 match stringlib/codecs.h:decode_utf8.

It also fixes an off-by-one error introduced in 3.10 for the line number when the tokenizer reports bad UTF8.
(cherry picked from commit 8bc356a7)

Co-authored-by: Michael Droettboom <mdboom@gmail.com>