QpEncoder.php 7.7 KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195
  1. <?php
  2. /*
  3. * This file is part of the Symfony package.
  4. *
  5. * (c) Fabien Potencier <fabien@symfony.com>
  6. *
  7. * For the full copyright and license information, please view the LICENSE
  8. * file that was distributed with this source code.
  9. */
  10. namespace Symfony\Component\Mime\Encoder;
  11. use Symfony\Component\Mime\CharacterStream;
  12. /**
  13. * @author Chris Corbyn
  14. */
  15. class QpEncoder implements EncoderInterface
  16. {
  17. /**
  18. * Pre-computed QP for HUGE optimization.
  19. */
  20. private const QP_MAP = [
  21. 0 => '=00', 1 => '=01', 2 => '=02', 3 => '=03', 4 => '=04',
  22. 5 => '=05', 6 => '=06', 7 => '=07', 8 => '=08', 9 => '=09',
  23. 10 => '=0A', 11 => '=0B', 12 => '=0C', 13 => '=0D', 14 => '=0E',
  24. 15 => '=0F', 16 => '=10', 17 => '=11', 18 => '=12', 19 => '=13',
  25. 20 => '=14', 21 => '=15', 22 => '=16', 23 => '=17', 24 => '=18',
  26. 25 => '=19', 26 => '=1A', 27 => '=1B', 28 => '=1C', 29 => '=1D',
  27. 30 => '=1E', 31 => '=1F', 32 => '=20', 33 => '=21', 34 => '=22',
  28. 35 => '=23', 36 => '=24', 37 => '=25', 38 => '=26', 39 => '=27',
  29. 40 => '=28', 41 => '=29', 42 => '=2A', 43 => '=2B', 44 => '=2C',
  30. 45 => '=2D', 46 => '=2E', 47 => '=2F', 48 => '=30', 49 => '=31',
  31. 50 => '=32', 51 => '=33', 52 => '=34', 53 => '=35', 54 => '=36',
  32. 55 => '=37', 56 => '=38', 57 => '=39', 58 => '=3A', 59 => '=3B',
  33. 60 => '=3C', 61 => '=3D', 62 => '=3E', 63 => '=3F', 64 => '=40',
  34. 65 => '=41', 66 => '=42', 67 => '=43', 68 => '=44', 69 => '=45',
  35. 70 => '=46', 71 => '=47', 72 => '=48', 73 => '=49', 74 => '=4A',
  36. 75 => '=4B', 76 => '=4C', 77 => '=4D', 78 => '=4E', 79 => '=4F',
  37. 80 => '=50', 81 => '=51', 82 => '=52', 83 => '=53', 84 => '=54',
  38. 85 => '=55', 86 => '=56', 87 => '=57', 88 => '=58', 89 => '=59',
  39. 90 => '=5A', 91 => '=5B', 92 => '=5C', 93 => '=5D', 94 => '=5E',
  40. 95 => '=5F', 96 => '=60', 97 => '=61', 98 => '=62', 99 => '=63',
  41. 100 => '=64', 101 => '=65', 102 => '=66', 103 => '=67', 104 => '=68',
  42. 105 => '=69', 106 => '=6A', 107 => '=6B', 108 => '=6C', 109 => '=6D',
  43. 110 => '=6E', 111 => '=6F', 112 => '=70', 113 => '=71', 114 => '=72',
  44. 115 => '=73', 116 => '=74', 117 => '=75', 118 => '=76', 119 => '=77',
  45. 120 => '=78', 121 => '=79', 122 => '=7A', 123 => '=7B', 124 => '=7C',
  46. 125 => '=7D', 126 => '=7E', 127 => '=7F', 128 => '=80', 129 => '=81',
  47. 130 => '=82', 131 => '=83', 132 => '=84', 133 => '=85', 134 => '=86',
  48. 135 => '=87', 136 => '=88', 137 => '=89', 138 => '=8A', 139 => '=8B',
  49. 140 => '=8C', 141 => '=8D', 142 => '=8E', 143 => '=8F', 144 => '=90',
  50. 145 => '=91', 146 => '=92', 147 => '=93', 148 => '=94', 149 => '=95',
  51. 150 => '=96', 151 => '=97', 152 => '=98', 153 => '=99', 154 => '=9A',
  52. 155 => '=9B', 156 => '=9C', 157 => '=9D', 158 => '=9E', 159 => '=9F',
  53. 160 => '=A0', 161 => '=A1', 162 => '=A2', 163 => '=A3', 164 => '=A4',
  54. 165 => '=A5', 166 => '=A6', 167 => '=A7', 168 => '=A8', 169 => '=A9',
  55. 170 => '=AA', 171 => '=AB', 172 => '=AC', 173 => '=AD', 174 => '=AE',
  56. 175 => '=AF', 176 => '=B0', 177 => '=B1', 178 => '=B2', 179 => '=B3',
  57. 180 => '=B4', 181 => '=B5', 182 => '=B6', 183 => '=B7', 184 => '=B8',
  58. 185 => '=B9', 186 => '=BA', 187 => '=BB', 188 => '=BC', 189 => '=BD',
  59. 190 => '=BE', 191 => '=BF', 192 => '=C0', 193 => '=C1', 194 => '=C2',
  60. 195 => '=C3', 196 => '=C4', 197 => '=C5', 198 => '=C6', 199 => '=C7',
  61. 200 => '=C8', 201 => '=C9', 202 => '=CA', 203 => '=CB', 204 => '=CC',
  62. 205 => '=CD', 206 => '=CE', 207 => '=CF', 208 => '=D0', 209 => '=D1',
  63. 210 => '=D2', 211 => '=D3', 212 => '=D4', 213 => '=D5', 214 => '=D6',
  64. 215 => '=D7', 216 => '=D8', 217 => '=D9', 218 => '=DA', 219 => '=DB',
  65. 220 => '=DC', 221 => '=DD', 222 => '=DE', 223 => '=DF', 224 => '=E0',
  66. 225 => '=E1', 226 => '=E2', 227 => '=E3', 228 => '=E4', 229 => '=E5',
  67. 230 => '=E6', 231 => '=E7', 232 => '=E8', 233 => '=E9', 234 => '=EA',
  68. 235 => '=EB', 236 => '=EC', 237 => '=ED', 238 => '=EE', 239 => '=EF',
  69. 240 => '=F0', 241 => '=F1', 242 => '=F2', 243 => '=F3', 244 => '=F4',
  70. 245 => '=F5', 246 => '=F6', 247 => '=F7', 248 => '=F8', 249 => '=F9',
  71. 250 => '=FA', 251 => '=FB', 252 => '=FC', 253 => '=FD', 254 => '=FE',
  72. 255 => '=FF',
  73. ];
  74. private static $safeMapShare = [];
  75. /**
  76. * A map of non-encoded ascii characters.
  77. *
  78. * @var string[]
  79. *
  80. * @internal
  81. */
  82. protected $safeMap = [];
  83. public function __construct()
  84. {
  85. $id = static::class;
  86. if (!isset(self::$safeMapShare[$id])) {
  87. $this->initSafeMap();
  88. self::$safeMapShare[$id] = $this->safeMap;
  89. } else {
  90. $this->safeMap = self::$safeMapShare[$id];
  91. }
  92. }
  93. protected function initSafeMap(): void
  94. {
  95. foreach (array_merge([0x09, 0x20], range(0x21, 0x3C), range(0x3E, 0x7E)) as $byte) {
  96. $this->safeMap[$byte] = \chr($byte);
  97. }
  98. }
  99. /**
  100. * {@inheritdoc}
  101. *
  102. * Takes an unencoded string and produces a QP encoded string from it.
  103. *
  104. * QP encoded strings have a maximum line length of 76 characters.
  105. * If the first line needs to be shorter, indicate the difference with
  106. * $firstLineOffset.
  107. */
  108. public function encodeString(string $string, ?string $charset = 'utf-8', int $firstLineOffset = 0, int $maxLineLength = 0): string
  109. {
  110. if ($maxLineLength > 76 || $maxLineLength <= 0) {
  111. $maxLineLength = 76;
  112. }
  113. $thisLineLength = $maxLineLength - $firstLineOffset;
  114. $lines = [];
  115. $lNo = 0;
  116. $lines[$lNo] = '';
  117. $currentLine = &$lines[$lNo++];
  118. $size = $lineLen = 0;
  119. $charStream = new CharacterStream($string, $charset);
  120. // Fetching more than 4 chars at one is slower, as is fetching fewer bytes
  121. // Conveniently 4 chars is the UTF-8 safe number since UTF-8 has up to 6
  122. // bytes per char and (6 * 4 * 3 = 72 chars per line) * =NN is 3 bytes
  123. while (null !== $bytes = $charStream->readBytes(4)) {
  124. $enc = $this->encodeByteSequence($bytes, $size);
  125. $i = strpos($enc, '=0D=0A');
  126. $newLineLength = $lineLen + (false === $i ? $size : $i);
  127. if ($currentLine && $newLineLength >= $thisLineLength) {
  128. $lines[$lNo] = '';
  129. $currentLine = &$lines[$lNo++];
  130. $thisLineLength = $maxLineLength;
  131. $lineLen = 0;
  132. }
  133. $currentLine .= $enc;
  134. if (false === $i) {
  135. $lineLen += $size;
  136. } else {
  137. // 6 is the length of '=0D=0A'.
  138. $lineLen = $size - strrpos($enc, '=0D=0A') - 6;
  139. }
  140. }
  141. return $this->standardize(implode("=\r\n", $lines));
  142. }
  143. /**
  144. * Encode the given byte array into a verbatim QP form.
  145. */
  146. private function encodeByteSequence(array $bytes, int &$size): string
  147. {
  148. $ret = '';
  149. $size = 0;
  150. foreach ($bytes as $b) {
  151. if (isset($this->safeMap[$b])) {
  152. $ret .= $this->safeMap[$b];
  153. ++$size;
  154. } else {
  155. $ret .= self::QP_MAP[$b];
  156. $size += 3;
  157. }
  158. }
  159. return $ret;
  160. }
  161. /**
  162. * Make sure CRLF is correct and HT/SPACE are in valid places.
  163. */
  164. private function standardize(string $string): string
  165. {
  166. $string = str_replace(["\t=0D=0A", ' =0D=0A', '=0D=0A'], ["=09\r\n", "=20\r\n", "\r\n"], $string);
  167. switch ($end = \ord(substr($string, -1))) {
  168. case 0x09:
  169. case 0x20:
  170. $string = substr_replace($string, self::QP_MAP[$end], -1);
  171. }
  172. return $string;
  173. }
  174. }