Characters in UTF-8 text

Decided

D-0069, settled by the language’s author: part of spec §18. Older text cites it as 18.69.

Positions are bytes; characters are decoded by iteration. String keeps UTF-8 as it always did, and gains chars() (and for c in s), char_count, pop, is_char_boundary, slice and truncate, every cut checked against a character boundary. Rejected: indexing by character, which is a linear walk disguised as s[i], and UTF-32 storage, which quadruples memory for the text a game mostly holds. Invalid bytes decode as U+FFFD rather than panicking, since only unchecked code can make them (§14.4)