awaforge

fontEncoding

AGPL-3.0-onlyRelease 2L4

Text encodings for the name table — UTF-16BE and Mac Roman.

Module fontEncoding | Source packages/front/office/fonts/src/primitives/encoding.js | Deps fontErrors | Worker-safe yes

Covers the two dominant encodings in the name table:

  • UTF-16BE — Windows platform (3) encoding 1 (BMP) or 10 (full repertoire), Unicode platform (0).
  • Mac Roman — Macintosh platform (1) encoding 0, 256-byte table.

Other Mac scripts (Japanese, ChineseTrad, …) are not supported: the corresponding name records keep their raw bytes unchanged.

Resolve

const { decodeUtf16Be, encodeUtf16Be, decodeMacRoman, encodeMacRoman } = runtime.resolve('fontEncoding');

API

Method Signature Returns
decodeUtf16Be (bytes: Uint8Array) => string Decodes UTF-16BE.
encodeUtf16Be (str: string) => Uint8Array Encodes (treated as UCS-2 / UTF-16 code units).
decodeMacRoman (bytes: Uint8Array) => string Decodes Mac Roman to Unicode.
encodeMacRoman (str: string) => Uint8Array Best-effort encode — characters outside the table → '?' (0x3F).

Examples

UTF-16BE decoding

const { decodeUtf16Be } = runtime.resolve('fontEncoding');
decodeUtf16Be(new Uint8Array([0x00, 0x41, 0x00, 0x42]));  // 'AB'

Mac Roman

const { encodeMacRoman, decodeMacRoman } = runtime.resolve('fontEncoding');
decodeMacRoman(new Uint8Array([0xC4]));  // '√'
encodeMacRoman('café');                  // Mac Roman bytes

Notes

  • decodeUtf16Be requires an even length — otherwise ParseError('fonts/utf16be-odd').
  • encodeMacRoman is best-effort: any character outside the 256-entry coverage becomes '?' without throwing.
  • For surrogate pairs, encodeUtf16Be writes the raw 16-bit JS code units directly (no BMP expansion).

See also