Module string/rune

Unicode code point type (rune) for representing individual characters.

A rune is a newtype over u32 holding one scalar value — Go's rune, Rust's char. from_u32 is the only constructor that validates, and it rejects the surrogate range D800..=DFFF as well as anything above 10FFFF, so a rune in hand is always encodable as UTF-8. Runes are values: Eq, Ord and Send are all defined by comparing the code point.

This module gives a PRIMITIVE type its inherent methods from outside the prelude (the prelude imports nothing and has no c_include, so it cannot host them); std/math.yo follows the same pattern.

Stability

unstable — seven of the eleven methods are ASCII-ONLY under names Rust reserves for the Unicode behaviour: is_whitespace, is_digit, is_alphabetic, is_uppercase, is_lowercase, to_lowercase and to_uppercase all test or shift the ASCII range and pass every other code point through unchanged. Each says so in its own doc, but the NAMES are the problem — Rust spells these is_ascii_* / to_ascii_* and keeps the unprefixed names for the table-driven versions. The audit's plan (plans/STD_API_STABILIZATION.md §4, "Text") is to rename them to is_ascii_* / to_ascii_* and add real Unicode versions backed by std/string/unicode.yo's tables. That rename is breaking, so the module cannot be frozen before it lands. rune itself, from_u32, to_u32, is_ascii and the named constants are settled.