core-js/internals/url-percent-coding.js

Standard library

'use strict';
// TODO: in core-js@4, move /modules/ dependencies to public entries for better optimization by tools like `preset-env`
require('../modules/es.string.from-code-point');
var getBuiltIn = require('../internals/get-built-in');
var uncurryThis = require('../internals/function-uncurry-this');

var fromCharCode = String.fromCharCode;
var fromCodePoint = getBuiltIn('String', 'fromCodePoint');
var $encodeURIComponent = encodeURIComponent;
var $parseInt = parseInt;
var charAt = uncurryThis(''.charAt);
var push = uncurryThis([].push);
var replace = uncurryThis(''.replace);
var stringSlice = uncurryThis(''.slice);
var exec = uncurryThis(/./.exec);

var FALLBACK_REPLACER = '\uFFFD';
var VALID_HEX = /^[0-9a-f]+$/i;
// a surrogate pair is matched first, so a one-unit match is always a lone surrogate
var SURROGATE = /[\uD800-\uDBFF][\uDC00-\uDFFF]|[\uD800-\uDFFF]/g;

var parseHexOctet = function (string, start) {
  var substr = stringSlice(string, start, start + 2);
  if (!exec(VALID_HEX, substr)) return NaN;

  return $parseInt(substr, 16);
};

var getLeadingOnes = function (octet) {
  var count = 0;
  for (var mask = 0x80; mask > 0 && (octet & mask) !== 0; mask >>= 1) {
    count++;
  }
  return count;
};

var utf8Decode = function (octets) {
  var codePoint = null;
  var length = octets.length;

  switch (length) {
    case 1:
      codePoint = octets[0];
      break;
    case 2:
      codePoint = (octets[0] & 0x1F) << 6 | (octets[1] & 0x3F);
      break;
    case 3:
      codePoint = (octets[0] & 0x0F) << 12 | (octets[1] & 0x3F) << 6 | (octets[2] & 0x3F);
      break;
    case 4:
      codePoint = (octets[0] & 0x07) << 18 | (octets[1] & 0x3F) << 12 | (octets[2] & 0x3F) << 6 | (octets[3] & 0x3F);
      break;
  }

  // reject surrogates, overlong encodings, and out-of-range codepoints
  if (codePoint === null
    || codePoint > 0x10FFFF
    || (codePoint >= 0xD800 && codePoint <= 0xDFFF)
    || codePoint < (length > 3 ? 0x10000 : length > 2 ? 0x800 : length > 1 ? 0x80 : 0)
  ) return null;

  return codePoint;
};

var replaceLoneSurrogate = function (chunk) {
  return chunk.length === 2 ? chunk : FALLBACK_REPLACER;
};

// https://url.spec.whatwg.org/#percent-decode
/* eslint-disable max-depth -- ok */
var decode = function (input) {
  var length = input.length;
  var result = '';
  var i = 0;

  while (i < length) {
    var decodedChar = charAt(input, i);

    if (decodedChar === '%') {
      if (charAt(input, i + 1) === '%' || i + 3 > length) {
        result += '%';
        i++;
        continue;
      }

      var octet = parseHexOctet(input, i + 1);

      // eslint-disable-next-line no-self-compare -- NaN check
      if (octet !== octet) {
        result += decodedChar;
        i++;
        continue;
      }

      i += 2;
      var byteSequenceLength = getLeadingOnes(octet);

      if (byteSequenceLength === 0) {
        decodedChar = fromCharCode(octet);
      } else {
        if (byteSequenceLength === 1 || byteSequenceLength > 4) {
          result += FALLBACK_REPLACER;
          i++;
          continue;
        }

        var octets = [octet];
        var sequenceIndex = 1;

        while (sequenceIndex < byteSequenceLength) {
          i++;
          if (i + 3 > length || charAt(input, i) !== '%') break;

          var nextByte = parseHexOctet(input, i + 1);

          // eslint-disable-next-line no-self-compare -- NaN check
          if (nextByte !== nextByte || nextByte > 191 || nextByte < 128) break;

          // https://encoding.spec.whatwg.org/#utf-8-decoder - position-specific byte ranges
          if (sequenceIndex === 1) {
            if (octet === 0xE0 && nextByte < 0xA0) break;
            if (octet === 0xED && nextByte > 0x9F) break;
            if (octet === 0xF0 && nextByte < 0x90) break;
            if (octet === 0xF4 && nextByte > 0x8F) break;
          }

          push(octets, nextByte);
          i += 2;
          sequenceIndex++;
        }

        if (octets.length !== byteSequenceLength) {
          result += FALLBACK_REPLACER;
          continue;
        }

        var codePoint = utf8Decode(octets);
        if (codePoint === null) {
          for (var replacement = 0; replacement < byteSequenceLength; replacement++) result += FALLBACK_REPLACER;
          i++;
          continue;
        } else {
          decodedChar = fromCodePoint(codePoint);
        }
      }
    }

    result += decodedChar;
    i++;
  }

  return result;
};
/* eslint-enable max-depth -- ok */

// https://url.spec.whatwg.org/#string-percent-encode-after-encoding
// a lone surrogate is the only input `encodeURIComponent` throws on, and the UTF-8
// encoder replaces it - so the throw selects the slow path instead of a per-call scan
var encode = function (input) {
  try {
    return $encodeURIComponent(input);
  } catch (error) {
    return $encodeURIComponent(replace(input, SURROGATE, replaceLoneSurrogate));
  }
};

module.exports = {
  decode: decode,
  encode: encode
};