wingo
commented
2 years ago - string.new_wtf8 takes a wtf8_policy immediate, to allow for U+FFFD replacement of any invalid WTF-8 subsequence.
- Remove string.measure_utf8, replacing with string.measure_wtf8 with a wtf8_policy immediate.
- string.measure_wtf16 returns length in code units, not bytes.
- Specify little-endian for string.encode_wtf16.