Iconvex Unicode

iconvex_unicode adds 17 pure-Elixir byte-stream codecs and two logical non-octet Unicode formats to Iconvex. Starting the OTP application publishes the complete name set atomically.

The package includes:

• BOCU-1, CESU-8, SCSU, UTF-1, UTF-EBCDIC, and WTF-8;
• historical 31-bit FSS-UTF (UTF-2);
• draft UTF-5 and whole-hostname UTF-6;
• strict Formal SignWriting in ASCII (FSW);
• explicit Iconvex UTF-16/UTF-32 signature profiles; and
• RFC 4042 UTF-9 and UTF-18, both as packed bitstrings and explicit endian-safe byte transports.

{:ok, compressed} = Iconvex.convert("日本語", "UTF-8", "BOCU-1")
{:ok, "日本語"} = Iconvex.convert(compressed, "BOCU-1", "UTF-8")

{:ok, packed_nonets} = Iconvex.Unicode.UTF9.encode_packed([?A, 0x0391, 0x10330])
{:ok, [?A, 0x0391, 0x10330]} = Iconvex.Unicode.UTF9.decode_packed(packed_nonets)

{:ok, fsw} = Iconvex.Unicode.FormalSignWriting.encode([0x40001])
{:ok, "S10000"} = {:ok, fsw}

Bare UTF-9 and UTF-18 are deliberately not byte-registry names. Use UTF-9-16BE, UTF-9-16LE, UTF-18-24BE, or UTF-18-24LE with Iconvex.convert/4, or call the logical modules for native packed bitstrings.

Development

Use the local Core checkout in this monorepo:

ICONVEX_PATH=../iconvex mix test

The runtime implementation is LGPL-2.1-or-later. Retained source records and Internet-Draft notices are described in NOTICE.