Source: lib/util/string_utils.js

  1. /**
  2.  * @license
  3.  * Copyright 2016 Google Inc.
  4.  *
  5.  * Licensed under the Apache License, Version 2.0 (the "License");
  6.  * you may not use this file except in compliance with the License.
  7.  * You may obtain a copy of the License at
  8.  *
  9.  *     http://www.apache.org/licenses/LICENSE-2.0
  10.  *
  11.  * Unless required by applicable law or agreed to in writing, software
  12.  * distributed under the License is distributed on an "AS IS" BASIS,
  13.  * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
  14.  * See the License for the specific language governing permissions and
  15.  * limitations under the License.
  16.  */
  18. goog.provide('shaka.util.StringUtils');
  20. goog.require('shaka.log');
  21. goog.require('shaka.util.Error');
  24. /**
  25.  * @namespace shaka.util.StringUtils
  26.  * @summary A set of string utility functions.
  27.  * @exportDoc
  28.  */
  31. /**
  32.  * Creates a string from the given buffer as UTF-8 encoding.
  33.  *
  34.  * @param {?BufferSource} data
  35.  * @return {string}
  36.  * @throws {shaka.util.Error}
  37.  * @export
  38.  */
  39. shaka.util.StringUtils.fromUTF8 = function(data) {
  40.   if (!data) return '';
  42.   var uint8 = new Uint8Array(data);
  43.   // If present, strip off the UTF-8 BOM.
  44.   if (uint8[0] == 0xef && uint8[1] == 0xbb && uint8[2] == 0xbf) {
  45.     uint8 = uint8.subarray(3);
  46.   }
  48.   // http://stackoverflow.com/a/13691499
  49.   var utf8 = shaka.util.StringUtils.fromCharCode_(uint8);
  50.   // This converts each character in the string to an escape sequence.  If the
  51.   // character is in the ASCII range, it is not converted; otherwise it is
  52.   // converted to a URI escape sequence.
  53.   // Example: '\x67\x35\xe3\x82\xac' -> 'g#%E3%82%AC'
  54.   var escaped = escape(utf8);
  55.   // Decode the escaped sequence.  This will interpret UTF-8 sequences into the
  56.   // correct character.
  57.   // Example: 'g#%E3%82%AC' -> 'g#€'
  58.   try {
  59.     return decodeURIComponent(escaped);
  60.   } catch (e) {
  61.     throw new shaka.util.Error(
  62.         shaka.util.Error.Severity.CRITICAL, shaka.util.Error.Category.TEXT,
  63.         shaka.util.Error.Code.BAD_ENCODING);
  64.   }
  65. };
  68. /**
  69.  * Creates a string from the given buffer as UTF-16 encoding.
  70.  *
  71.  * @param {?BufferSource} data
  72.  * @param {boolean} littleEndian true to read little endian, false to read big.
  73.  * @return {string}
  74.  * @throws {shaka.util.Error}
  75.  * @export
  76.  */
  77. shaka.util.StringUtils.fromUTF16 = function(data, littleEndian) {
  78.   if (!data) return '';
  80.   if (data.byteLength % 2 != 0) {
  81.     shaka.log.error('Data has an incorrect length, must be even.');
  82.     throw new shaka.util.Error(
  83.         shaka.util.Error.Severity.CRITICAL, shaka.util.Error.Category.TEXT,
  84.         shaka.util.Error.Code.BAD_ENCODING);
  85.   }
  87.   /** @type {ArrayBuffer} */
  88.   var buffer;
  89.   if (data instanceof ArrayBuffer) {
  90.     buffer = data;
  91.   } else {
  92.     // Have to create a new buffer because the argument may be a smaller
  93.     // view on a larger ArrayBuffer.  We cannot use an ArrayBufferView in
  94.     // a DataView.
  95.     var temp = new Uint8Array(data.byteLength);
  96.     temp.set(new Uint8Array(data));
  97.     buffer = temp.buffer;
  98.   }
  100.   // Use a DataView to ensure correct endianness.
  101.   var length = data.byteLength / 2;
  102.   var arr = new Uint16Array(length);
  103.   var dataView = new DataView(buffer);
  104.   for (var i = 0; i < length; i++) {
  105.     arr[i] = dataView.getUint16(i * 2, littleEndian);
  106.   }
  107.   return shaka.util.StringUtils.fromCharCode_(arr);
  108. };
  111. /**
  112.  * Creates a string from the given buffer, auto-detecting the encoding that is
  113.  * being used.  If it cannot detect the encoding, it will throw an exception.
  114.  *
  115.  * @param {?BufferSource} data
  116.  * @return {string}
  117.  * @throws {shaka.util.Error}
  118.  * @export
  119.  */
  120. shaka.util.StringUtils.fromBytesAutoDetect = function(data) {
  121.   var StringUtils = shaka.util.StringUtils;
  123.   var uint8 = new Uint8Array(data);
  124.   if (uint8[0] == 0xef && uint8[1] == 0xbb && uint8[2] == 0xbf)
  125.     return StringUtils.fromUTF8(uint8);
  126.   else if (uint8[0] == 0xfe && uint8[1] == 0xff)
  127.     return StringUtils.fromUTF16(uint8.subarray(2), false /* littleEndian */);
  128.   else if (uint8[0] == 0xff && uint8[1] == 0xfe)
  129.     return StringUtils.fromUTF16(uint8.subarray(2), true /* littleEndian */);
  131.   var isAscii = (function(arr, i) {
  132.     // arr[i] >= ' ' && arr[i] <= '~';
  133.     return arr.byteLength <= i || (arr[i] >= 0x20 && arr[i] <= 0x7e);
  134.   }.bind(null, uint8));
  136.   shaka.log.debug('Unable to find byte-order-mark, making an educated guess.');
  137.   if (uint8[0] == 0 && uint8[2] == 0)
  138.     return StringUtils.fromUTF16(data, false /* littleEndian */);
  139.   else if (uint8[1] == 0 && uint8[3] == 0)
  140.     return StringUtils.fromUTF16(data, true /* littleEndian */);
  141.   else if (isAscii(0) && isAscii(1) && isAscii(2) && isAscii(3))
  142.     return StringUtils.fromUTF8(data);
  144.   throw new shaka.util.Error(
  145.       shaka.util.Error.Severity.CRITICAL,
  146.       shaka.util.Error.Category.TEXT,
  147.       shaka.util.Error.Code.UNABLE_TO_DETECT_ENCODING);
  148. };
  151. /**
  152.  * Creates a ArrayBuffer from the given string, converting to UTF-8 encoding.
  153.  *
  154.  * @param {string} str
  155.  * @return {!ArrayBuffer}
  156.  * @export
  157.  */
  158. shaka.util.StringUtils.toUTF8 = function(str) {
  159.   // http://stackoverflow.com/a/13691499
  160.   // Converts the given string to a URI encoded string.  If a character falls
  161.   // in the ASCII range, it is not converted; otherwise it will be converted to
  162.   // a series of URI escape sequences according to UTF-8.
  163.   // Example: 'g#€' -> 'g#%E3%82%AC'
  164.   var encoded = encodeURIComponent(str);
  165.   // Convert each escape sequence individually into a character.  Each escape
  166.   // sequence is interpreted as a code-point, so if an escape sequence happens
  167.   // to be part of a multi-byte sequence, each byte will be converted to a
  168.   // single character.
  169.   // Example: 'g#%E3%82%AC' -> '\x67\x35\xe3\x82\xac'
  170.   var utf8 = unescape(encoded);
  172.   var result = new Uint8Array(utf8.length);
  173.   for (var i = 0; i < utf8.length; ++i) {
  174.     result[i] = utf8.charCodeAt(i);
  175.   }
  176.   return result.buffer;
  177. };
  180. /**
  181.  * Creates a new string from the given array of char codes.
  182.  *
  183.  * @param {!TypedArray} args
  184.  * @return {string}
  185.  * @private
  186.  */
  187. shaka.util.StringUtils.fromCharCode_ = function(args) {
  188.   var max = 16000;
  189.   var ret = '';
  190.   for (var i = 0; i < args.length; i += max) {
  191.     var subArray = args.subarray(i, i + max);
  192.     ret += String.fromCharCode.apply(null, subArray);
  193.   }
  195.   return ret;
  196. };