rbxts-pako/lib/utils/strings.js

// String encode/decode helpers
'use strict';


var utils = require('./common');


// Quick check if we can use fast array to bin string conversion
var STR_APPLY_OK = true;
try { String.fromCharCode.apply(null, [0]); } catch(__) { STR_APPLY_OK = false; }


// Table with utf8 lengths (calculated by first byte of sequence)
// Note, that 5 & 6-byte values and some 4-byte values can not be represented in JS,
// because max possible codepoiny is 0x10ffff
var _utf8len = new utils.Buf8(256);
for (var i=0; i<256; i++) {
  _utf8len[i] = (i >= 252 ? 6 : i >= 248 ? 5 : i >= 240 ? 4 : i >= 224 ? 3 : i >= 192 ? 2 : 1);
}
_utf8len[254]=_utf8len[254]=1; // Invalid sequence start


// convert string to array (typed, when possible)
// src: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Base64_encoding_and_decoding
exports.string2buf = function (str) {
  var buf, c, c2, m_pos, i, str_len = str.length, buf_len = 0;

  /* mapping... */

  for (m_pos = 0; m_pos < str_len; m_pos++) {
    c = str.charCodeAt(m_pos);
    if ((c & 0xfc00) === 0xd800 && (m_pos+1 < str_len)) {
      c2 = str.charCodeAt(m_pos+1);
      if ((c2 & 0xfc00) === 0xdc00) {
        c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00);
        m_pos++;
      }
    }
    buf_len += c < 0x80 ? 1 : c < 0x800 ? 2 : c < 0x10000 ? 3 : 4;
  }

  buf = new utils.Buf8(buf_len);

  /* transcription... */

  for (i=0, m_pos = 0; i < buf_len; m_pos++) {
    c = str.charCodeAt(m_pos);
    if ((c & 0xfc00) === 0xd800 && (m_pos+1 < str_len)) {
      c2 = str.charCodeAt(m_pos+1);
      if ((c2 & 0xfc00) === 0xdc00) {
        c = 0x10000 + ((c - 0xd800) << 10) + (c2 - 0xdc00);
        m_pos++;
      }
    }
    if (c < 0x80) {
      /* one byte */
      buf[i++] = c;
    } else if (c < 0x800) {
      /* two bytes */
      buf[i++] = 0xC0 | (c >>> 6);
      buf[i++] = 0x80 | (c & 0x3f);
    } else if (c < 0x10000) {
      /* three bytes */
      buf[i++] = 0xE0 | (c >>> 12);
      buf[i++] = 0x80 | (c >>> 6 & 0x3f);
      buf[i++] = 0x80 | (c & 0x3f);
    } else {
      /* four bytes */
      buf[i++] = 0xf0 | (c >>> 18);
      buf[i++] = 0x80 | (c >>> 12 & 0x3f);
      buf[i++] = 0x80 | (c >>> 6 & 0x3f);
      buf[i++] = 0x80 | (c & 0x3f);
    }
  }

  return buf;
};


// Convert byte array to binary string
exports.buf2binstring = function(buf) {
  // use fallback for big arrays to avoid stack overflow
  if (STR_APPLY_OK && buf.length < 65537) {
    return String.fromCharCode.apply(null, buf);
  }

  var result = '';
  for(var i=0, len=buf.length; i < len; i++) {
    result += String.fromCharCode(buf[i]);
  }
  return result;
};


// Convert binary string (typed, when possible)
exports.binstring2buf = function(str) {
  var buf = new utils.Buf8(str.length);
  for(var i=0, len=buf.length; i < len; i++) {
    buf[i] = str.charCodeAt(i);
  }
  return buf;
};


// convert array to string
// src: https://developer.mozilla.org/en-US/docs/Web/JavaScript/Base64_encoding_and_decoding
exports.buf2string = function (buf, max) {
  /*jshint nonstandard:true*/
  // That's not as fast as via String.fromCharCode.apply
  /*return decodeURIComponent(escape(exports.buf2binstring(
    (buf.length === max) ?
      buf
    :
      buf.subarray ? buf.subarray(0, max) : buf.slice(0, max)
  )));*/

  var str, i, out, part, char_len;
  var len = max || buf.length;

  // Reserve max possible length (2 words per char)
  var utf16buf = new utils.Buf16(len*2);

  for (out=0, i=0; i<len; i++) {
    part = buf[i];
    char_len = _utf8len[part];

    // edge case - broken sequence
    if (i + char_len > len) {
      utf16buf[out++] = 0xfffd;
      break; // end of string reached, stop
    }
    switch (char_len) {
      case 1:
        utf16buf[out++] = part;
        break;
      case 2:
        utf16buf[out++] = ((part & 0x1f) << 6) | (buf[++i] & 0x3f);
        break;
      case 3:
        utf16buf[out++] = ((part & 0x0f) << 12) | ((buf[++i] & 0x3f) << 6) | (buf[++i] & 0x3f);
        break;
      case 4:
        // surrogate pair
        part = ((part & 0x07) << 18) | ((buf[++i] & 0x3f) << 12) | ((buf[++i] & 0x3f) << 6) + (buf[++i] & 0x3f);
        part -= 0x10000;
        utf16buf[out++] = 0xd800 | ((part >> 10) & 0x3ff);
        utf16buf[out++] = 0xdc00 | (part & 0x3ff);
        break;
      // 5 & 6 bytes uticodes not supported in UTF16 (JS),
      // so fill with dummy symbol & update scan position.
      case 5:
        i += 4;
        utf16buf[out++] = 0xfffd;
        break;
      case 6:
        i += 5;
        utf16buf[out++] = 0xfffd;
        break;
    }
  }

  if (STR_APPLY_OK) {
    return String.fromCharCode.apply(null, utils.shrinkBuf(utf16buf, out));
  }

  // Fallback, when String.fromCharCode.apply not available
  str = '';
  for (i=0, len=out; i<len; i++) {
    str += String.fromCharCode(utf16buf[i]);
  }
  return str;
};


// Calculate max possible position in utf8 buffer,
// that will not break sequence. If that's not possible
// - (very small limits) return max size as is.
//
// buf[] - utf8 bytes array
// max   - length limit (mandatory);
exports.utf8border = function(buf, max) {
  var pos;

  max = max || buf.length;
  if (max > buf.length) { max = buf.length; }

  // go back from last position, until start of sequence found
  pos = max-1;
  while (pos >= 0 && (buf[pos] & 0xC0) === 0x80) { pos--; }

  // Fuckup - very small and broken sequence,
  // return max, because we should return something anyway.
  if (pos < 0) { return max; }

  // If we came to start of buffer - that means vuffer is too small,
  // return max too.
  if (pos === 0) { return max; }

  return (pos + _utf8len[buf[pos]] > max) ? pos : max;
};