HELP UTF8                    Code-point operations over UTF-8 byte strings

    uses utf8;

Poplog strings are byte strings; this library treats their bytes as
UTF-8 text.  Validation is strict (RFC 3629): overlong encodings,
surrogates, values above U+10FFFF, truncated sequences and stray
continuation bytes are rejected.  Everything except utf8_valid
mishaps on malformed input.

utf8_valid(S) -> BOOL

utf8_length(S) -> N
    Code points, not bytes: length('日本語') is 9, utf8_length is 3.

utf8_code(I, S) -> CODEPOINT
    The I-th code point (1-based).

utf8_substring(I, N, S) -> S
    N code points starting at the I-th — the code-point analogue of
    substring.

utf8_explode(S)
    Push every code point on the stack (pair with consutf8 or count
    with #| ... |#).

consutf8(N) -> S
    Build a UTF-8 string from N code points on the stack:
        consutf8(16:65E5, 16:672C, 2) -> '日本'
    Rejects surrogates and out-of-range values.

utf8_appcodes(S, P)
    Apply P to each code point in order.

Interplay with LIB * JSON: json_parse decodes \uXXXX escapes to
UTF-8 bytes, so utf8_* is the right toolkit for examining what it
returns.

See also: LIB * STRUTILS (byte-level operations), pop/extern/unicode
(C codecs for legacy encodings), tools/tests/test_utf8.p.

--- pop/help/utf8