neo-neo-bblessed/lib/unicode.ts

A fork of neo-blessed (which is a fork of blessed) with bug fixes and maintenance.

/**
 * unicode.ts - modern unicode utilities using npm packages
 * Copyright (c) 2025, Contributors (MIT License).
 * Replaces embedded unicode libraries with proper npm dependencies
 */

import * as eastasianwidth from 'eastasianwidth';
import * as unicode from 'unicode-properties';

/**
 * Wide, Surrogates, and Combining
 */

function charWidth(str: string | number, i?: number): number {
  const point = typeof str !== 'number' ? codePointAt(str, i || 0) : str;

  // nul
  if (point === 0) return 0;

  // tab
  if (point === 0x09) {
    // Default tab width is 8, can be overridden by environment variable
    return +(process.env['BLESSED_TAB_WIDTH'] || 8);
  }

  // 8-bit control characters (2-width according to unicode??)
  if (point < 32 || (point >= 0x7f && point < 0xa0)) {
    return 0;
  }

  // Use unicode-properties to check for combining characters
  if (
    unicode.getCategory(point) === 'Mn' ||
    unicode.getCategory(point) === 'Me' ||
    unicode.getCategory(point) === 'Mc'
  ) {
    return 0;
  }

  // Use eastasianwidth package for proper width detection
  const width = eastasianwidth.eastAsianWidth(String.fromCodePoint(point));

  switch (width) {
    case 'F': // Fullwidth
    case 'W': // Wide
      return 2;
    case 'H': // Halfwidth
    case 'Na': // Narrow
    case 'N': // Neutral
      return 1;
    case 'A': // Ambiguous
      // Handle CJK Ambiguous based on environment variable
      if (process.env['NCURSES_CJK_WIDTH']) {
        return +process.env['NCURSES_CJK_WIDTH'] || 1;
      }
      return 1;
    default:
      return 1;
  }
}

function strWidth(str: string): number {
  let width = 0;
  for (let i = 0; i < str.length; i++) {
    width += charWidth(str, i);
    if (isSurrogate(str, i)) i++;
  }
  return width;
}

function isSurrogate(str: string | number, i?: number): boolean {
  const point = typeof str !== 'number' ? codePointAt(str, i || 0) : str;
  return point > 0x00ffff;
}

/**
 * Use unicode-properties for combining character detection
 */
function isCombining(str: string | number, i?: number): boolean {
  const point = typeof str !== 'number' ? codePointAt(str, i || 0) : str;

  const category = unicode.getCategory(point);
  return category === 'Mn' || category === 'Me' || category === 'Mc';
}

/**
 * Code Point Helpers - using native implementations when available
 */

function codePointAt(str: string, position?: number): number {
  if (str == null) {
    throw TypeError();
  }
  const string = String(str);

  // Use native implementation if available
  if (string.codePointAt) {
    return string.codePointAt(position || 0) || 0;
  }

  // Fallback implementation
  const size = string.length;
  const index = position ? Number(position) : 0;

  if (index !== index) {
    // better `isNaN`
    return 0;
  }

  // Account for out-of-bounds indices:
  if (index < 0 || index >= size) {
    return 0;
  }

  // Get the first code unit
  const first = string.charCodeAt(index);
  let second;

  if (
    // check if it's the start of a surrogate pair
    first >= 0xd800 &&
    first <= 0xdbff && // high surrogate
    size > index + 1 // there is a next code unit
  ) {
    second = string.charCodeAt(index + 1);
    if (second >= 0xdc00 && second <= 0xdfff) {
      // low surrogate
      // http://mathiasbynens.be/notes/javascript-encoding#surrogate-formulae
      return (first - 0xd800) * 0x400 + second - 0xdc00 + 0x10000;
    }
  }

  return first;
}

function fromCodePoint(...codePoints: number[]): string {
  // Use native implementation if available
  if (String.fromCodePoint) {
    return String.fromCodePoint(...codePoints);
  }

  // Fallback implementation
  const MAX_SIZE = 0x4000;
  const codeUnits: number[] = [];
  let highSurrogate: number;
  let lowSurrogate: number;
  let index = -1;
  const length = codePoints.length;

  if (!length) {
    return '';
  }

  let result = '';

  while (++index < length) {
    const codePoint = Number(codePoints[index]);

    if (
      !isFinite(codePoint) || // `NaN`, `+Infinity`, or `-Infinity`
      codePoint < 0 || // not a valid Unicode code point
      codePoint > 0x10ffff || // not a valid Unicode code point
      Math.floor(codePoint) !== codePoint // not an integer
    ) {
      throw RangeError('Invalid code point: ' + codePoint);
    }

    if (codePoint <= 0xffff) {
      // BMP code point
      codeUnits.push(codePoint);
    } else {
      // Astral code point; split in surrogate halves
      // http://mathiasbynens.be/notes/javascript-encoding#surrogate-formulae
      const adjustedCodePoint = codePoint - 0x10000;
      highSurrogate = (adjustedCodePoint >> 10) + 0xd800;
      lowSurrogate = (adjustedCodePoint % 0x400) + 0xdc00;
      codeUnits.push(highSurrogate, lowSurrogate);
    }

    if (index + 1 === length || codeUnits.length > MAX_SIZE) {
      result += String.fromCharCode(...codeUnits);
      codeUnits.length = 0;
    }
  }

  return result;
}

/**
 * Regexes for compatibility
 */

const chars = {
  // Double width characters that are _not_ surrogate pairs.
  wide: new RegExp(
    '([' +
      '\\u1100-\\u115f' + // Hangul Jamo init. consonants
      '\\u2329\\u232a' +
      '\\u2e80-\\u303e\\u3040-\\ua4cf' + // CJK ... Yi
      '\\uac00-\\ud7a3' + // Hangul Syllables
      '\\uf900-\\ufaff' + // CJK Compatibility Ideographs
      '\\ufe10-\\ufe19' + // Vertical forms
      '\\ufe30-\\ufe6f' + // CJK Compatibility Forms
      '\\uff00-\\uff60' + // Fullwidth Forms
      '\\uffe0-\\uffe6' +
      '])',
    'g'
  ),

  // All surrogate pair wide chars.
  swide: new RegExp(
    '(' +
      // 0x20000 - 0x2fffd:
      '[\\ud840-\\ud87f][\\udc00-\\udffd]' +
      '|' +
      // 0x30000 - 0x3fffd:
      '[\\ud880-\\ud8bf][\\udc00-\\udffd]' +
      ')',
    'g'
  ),

  // Regex to detect a surrogate pair.
  surrogate: /[\ud800-\udbff][\udc00-\udfff]/g,

  // Basic combining regex (simplified, using unicode-properties for actual detection)
  combining:
    /[\u0300-\u036f\u0483-\u0489\u0591-\u05bd\u05bf\u05c1-\u05c2\u05c4-\u05c5\u05c7\u0610-\u061a\u064b-\u065f\u0670\u06d6-\u06dc\u06df-\u06e4\u06e7-\u06e8\u06ea-\u06ed]/g,
};

// All wide chars including surrogate pairs.
const charsWithAll = {
  ...chars,
  all: new RegExp(
    '(' +
      chars.swide.source.slice(1, -1) +
      '|' +
      chars.wide.source.slice(1, -1) +
      ')',
    'g'
  ),
};

// Backward compatibility - these were exports in the original

// Legacy export compatibility
const combining = {};
const combiningTable: number[][] = [];

// Initialize combining lookup for backward compatibility
function initializeCombining() {
  for (let i = 0; i <= 0x10ffff; i++) {
    const category = unicode.getCategory(i);
    if (category === 'Mn' || category === 'Me' || category === 'Mc') {
      (combining as any)[i] = true;
    }
  }
}

// Lazy initialization
let combiningInitialized = false;

function ensureCombiningInitialized() {
  if (!combiningInitialized) {
    initializeCombining();
    combiningInitialized = true;
  }
}

// Legacy exports for CommonJS compatibility
const combiningProxy = new Proxy(
  {},
  {
    get(_target, prop) {
      ensureCombiningInitialized();
      return (combining as any)[prop];
    },
  }
);

// Export everything properly for ESM
export {
  charWidth,
  strWidth,
  isSurrogate,
  isCombining,
  codePointAt,
  fromCodePoint,
  charsWithAll as chars,
  combiningProxy as combining,
  combiningTable,
};