unistr/u8-*: Make Unicode decoder more Unicode Standard compliant.
Based on a remark by Paul Eggert in
<https://lists.gnu.org/archive/html/bug-gnulib/2023-07/msg00120.html>.
* tests/unistr/test-u8-mbtouc.c (test_safe_function): Change expected
results for "non-shortest form" or out-of-range byte sequences. Add new
test cases of incomplete well-formed byte sequences.
* tests/unistr/test-u8-mbsnlen.c (main): Likewise.
* lib/unistr/u8-mbtouc-aux.c (u8_mbtouc_aux): Reject a first byte in the
range 0xF5..0xF7 as invalid. Distinguish incomplete from invalid byte
sequences correctly. For the former, return only the number of bytes in
the maximal well-formed subpart.
* lib/unistr/u8-mbtouc.c (u8_mbtouc): Likewise.
* lib/unistr/u8-check.c (u8_check): Reject a first byte in the range
0xF5..0xF7 as invalid.
* lib/unistr/u8-mblen.c (u8_mblen): Likewise.
* lib/unistr/u8-mbtoucr.c (u8_mbtoucr): Likewise.
* lib/unistr/u8-strmbtouc.c (u8_strmbtouc): Likewise.
* lib/unistr/u8-strmblen.c (u8_strmblen): Likewise.
* lib/unistr/u8-prev.c (u8_prev): Likewise.