QpEncoder.php 8.7KB

123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293949596979899100101102103104105106107108109110111112113114115116117118119120121122123124125126127128129130131132133134135136137138139140141142143144145146147148149150151152153154155156157158159160161162163164165166167168169170171172173174175176177178179180181182183184185186187188189190191192193194195196197198199200201202203204205206207208209210211212213214215216217218219220221222223224225226227228229230231232233234235236237238239240241242243244245246247248249250251252253254255256257258259
  1. <?php
  2. /*
  3. * This file is part of SwiftMailer.
  4. * (c) 2004-2009 Chris Corbyn
  5. *
  6. * For the full copyright and license information, please view the LICENSE
  7. * file that was distributed with this source code.
  8. */
  9. /**
  10. * Handles Quoted Printable (QP) Encoding in Swift Mailer.
  11. * Possibly the most accurate RFC 2045 QP implementation found in PHP.
  12. * @package Swift
  13. * @subpackage Encoder
  14. * @author Chris Corbyn
  15. */
  16. class Swift_Encoder_QpEncoder implements Swift_Encoder
  17. {
  18. /**
  19. * The CharacterStream used for reading characters (as opposed to bytes).
  20. * @var Swift_CharacterStream
  21. * @access protected
  22. */
  23. protected $_charStream;
  24. /**
  25. * A filter used if input should be canonicalized.
  26. * @var Swift_StreamFilter
  27. * @access protected
  28. */
  29. protected $_filter;
  30. /**
  31. * Pre-computed QP for HUGE optmization.
  32. * @var string[]
  33. * @access protected
  34. */
  35. protected $_qpMap = array(
  36. 0 => '=00', 1 => '=01', 2 => '=02', 3 => '=03', 4 => '=04',
  37. 5 => '=05', 6 => '=06', 7 => '=07', 8 => '=08', 9 => '=09',
  38. 10 => '=0A', 11 => '=0B', 12 => '=0C', 13 => '=0D', 14 => '=0E',
  39. 15 => '=0F', 16 => '=10', 17 => '=11', 18 => '=12', 19 => '=13',
  40. 20 => '=14', 21 => '=15', 22 => '=16', 23 => '=17', 24 => '=18',
  41. 25 => '=19', 26 => '=1A', 27 => '=1B', 28 => '=1C', 29 => '=1D',
  42. 30 => '=1E', 31 => '=1F', 32 => '=20', 33 => '=21', 34 => '=22',
  43. 35 => '=23', 36 => '=24', 37 => '=25', 38 => '=26', 39 => '=27',
  44. 40 => '=28', 41 => '=29', 42 => '=2A', 43 => '=2B', 44 => '=2C',
  45. 45 => '=2D', 46 => '=2E', 47 => '=2F', 48 => '=30', 49 => '=31',
  46. 50 => '=32', 51 => '=33', 52 => '=34', 53 => '=35', 54 => '=36',
  47. 55 => '=37', 56 => '=38', 57 => '=39', 58 => '=3A', 59 => '=3B',
  48. 60 => '=3C', 61 => '=3D', 62 => '=3E', 63 => '=3F', 64 => '=40',
  49. 65 => '=41', 66 => '=42', 67 => '=43', 68 => '=44', 69 => '=45',
  50. 70 => '=46', 71 => '=47', 72 => '=48', 73 => '=49', 74 => '=4A',
  51. 75 => '=4B', 76 => '=4C', 77 => '=4D', 78 => '=4E', 79 => '=4F',
  52. 80 => '=50', 81 => '=51', 82 => '=52', 83 => '=53', 84 => '=54',
  53. 85 => '=55', 86 => '=56', 87 => '=57', 88 => '=58', 89 => '=59',
  54. 90 => '=5A', 91 => '=5B', 92 => '=5C', 93 => '=5D', 94 => '=5E',
  55. 95 => '=5F', 96 => '=60', 97 => '=61', 98 => '=62', 99 => '=63',
  56. 100 => '=64', 101 => '=65', 102 => '=66', 103 => '=67', 104 => '=68',
  57. 105 => '=69', 106 => '=6A', 107 => '=6B', 108 => '=6C', 109 => '=6D',
  58. 110 => '=6E', 111 => '=6F', 112 => '=70', 113 => '=71', 114 => '=72',
  59. 115 => '=73', 116 => '=74', 117 => '=75', 118 => '=76', 119 => '=77',
  60. 120 => '=78', 121 => '=79', 122 => '=7A', 123 => '=7B', 124 => '=7C',
  61. 125 => '=7D', 126 => '=7E', 127 => '=7F', 128 => '=80', 129 => '=81',
  62. 130 => '=82', 131 => '=83', 132 => '=84', 133 => '=85', 134 => '=86',
  63. 135 => '=87', 136 => '=88', 137 => '=89', 138 => '=8A', 139 => '=8B',
  64. 140 => '=8C', 141 => '=8D', 142 => '=8E', 143 => '=8F', 144 => '=90',
  65. 145 => '=91', 146 => '=92', 147 => '=93', 148 => '=94', 149 => '=95',
  66. 150 => '=96', 151 => '=97', 152 => '=98', 153 => '=99', 154 => '=9A',
  67. 155 => '=9B', 156 => '=9C', 157 => '=9D', 158 => '=9E', 159 => '=9F',
  68. 160 => '=A0', 161 => '=A1', 162 => '=A2', 163 => '=A3', 164 => '=A4',
  69. 165 => '=A5', 166 => '=A6', 167 => '=A7', 168 => '=A8', 169 => '=A9',
  70. 170 => '=AA', 171 => '=AB', 172 => '=AC', 173 => '=AD', 174 => '=AE',
  71. 175 => '=AF', 176 => '=B0', 177 => '=B1', 178 => '=B2', 179 => '=B3',
  72. 180 => '=B4', 181 => '=B5', 182 => '=B6', 183 => '=B7', 184 => '=B8',
  73. 185 => '=B9', 186 => '=BA', 187 => '=BB', 188 => '=BC', 189 => '=BD',
  74. 190 => '=BE', 191 => '=BF', 192 => '=C0', 193 => '=C1', 194 => '=C2',
  75. 195 => '=C3', 196 => '=C4', 197 => '=C5', 198 => '=C6', 199 => '=C7',
  76. 200 => '=C8', 201 => '=C9', 202 => '=CA', 203 => '=CB', 204 => '=CC',
  77. 205 => '=CD', 206 => '=CE', 207 => '=CF', 208 => '=D0', 209 => '=D1',
  78. 210 => '=D2', 211 => '=D3', 212 => '=D4', 213 => '=D5', 214 => '=D6',
  79. 215 => '=D7', 216 => '=D8', 217 => '=D9', 218 => '=DA', 219 => '=DB',
  80. 220 => '=DC', 221 => '=DD', 222 => '=DE', 223 => '=DF', 224 => '=E0',
  81. 225 => '=E1', 226 => '=E2', 227 => '=E3', 228 => '=E4', 229 => '=E5',
  82. 230 => '=E6', 231 => '=E7', 232 => '=E8', 233 => '=E9', 234 => '=EA',
  83. 235 => '=EB', 236 => '=EC', 237 => '=ED', 238 => '=EE', 239 => '=EF',
  84. 240 => '=F0', 241 => '=F1', 242 => '=F2', 243 => '=F3', 244 => '=F4',
  85. 245 => '=F5', 246 => '=F6', 247 => '=F7', 248 => '=F8', 249 => '=F9',
  86. 250 => '=FA', 251 => '=FB', 252 => '=FC', 253 => '=FD', 254 => '=FE',
  87. 255 => '=FF'
  88. );
  89. /**
  90. * A map of non-encoded ascii characters.
  91. * @var string[]
  92. * @access protected
  93. */
  94. protected $_safeMap = array();
  95. /**
  96. * Creates a new QpEncoder for the given CharacterStream.
  97. * @param Swift_CharacterStream $charStream to use for reading characters
  98. * @param Swift_StreamFilter $filter if input should be canonicalized
  99. */
  100. public function __construct(Swift_CharacterStream $charStream,
  101. Swift_StreamFilter $filter = null)
  102. {
  103. $this->_charStream = $charStream;
  104. foreach (array_merge(
  105. array(0x09, 0x20), range(0x21, 0x3C), range(0x3E, 0x7E)) as $byte)
  106. {
  107. $this->_safeMap[$byte] = chr($byte);
  108. }
  109. $this->_filter = $filter;
  110. }
  111. /**
  112. * Takes an unencoded string and produces a QP encoded string from it.
  113. * QP encoded strings have a maximum line length of 76 characters.
  114. * If the first line needs to be shorter, indicate the difference with
  115. * $firstLineOffset.
  116. * @param string $string to encode
  117. * @param int $firstLineOffset, optional
  118. * @param int $maxLineLength, optional, 0 indicates the default of 76 chars
  119. * @return string
  120. */
  121. public function encodeString($string, $firstLineOffset = 0,
  122. $maxLineLength = 0)
  123. {
  124. if ($maxLineLength > 76 || $maxLineLength <= 0)
  125. {
  126. $maxLineLength = 76;
  127. }
  128. $thisLineLength = $maxLineLength - $firstLineOffset;
  129. $lines = array();
  130. $lNo = 0;
  131. $lines[$lNo] = '';
  132. $currentLine =& $lines[$lNo++];
  133. $size=$lineLen=0;
  134. $this->_charStream->flushContents();
  135. $this->_charStream->importString($string);
  136. //Fetching more than 4 chars at one is slower, as is fetching fewer bytes
  137. // Conveniently 4 chars is the UTF-8 safe number since UTF-8 has up to 6
  138. // bytes per char and (6 * 4 * 3 = 72 chars per line) * =NN is 3 bytes
  139. while (false !== $bytes = $this->_nextSequence())
  140. {
  141. //If we're filtering the input
  142. if (isset($this->_filter))
  143. {
  144. //If we can't filter because we need more bytes
  145. while ($this->_filter->shouldBuffer($bytes))
  146. {
  147. //Then collect bytes into the buffer
  148. if (false === $moreBytes = $this->_nextSequence(1))
  149. {
  150. break;
  151. }
  152. foreach ($moreBytes as $b)
  153. {
  154. $bytes[] = $b;
  155. }
  156. }
  157. //And filter them
  158. $bytes = $this->_filter->filter($bytes);
  159. }
  160. $enc = $this->_encodeByteSequence($bytes, $size);
  161. if ($currentLine && $lineLen+$size >= $thisLineLength)
  162. {
  163. $lines[$lNo] = '';
  164. $currentLine =& $lines[$lNo++];
  165. $thisLineLength = $maxLineLength;
  166. $lineLen=0;
  167. }
  168. $lineLen+=$size;
  169. $currentLine .= $enc;
  170. }
  171. return $this->_standardize(implode("=\r\n", $lines));
  172. }
  173. /**
  174. * Updates the charset used.
  175. * @param string $charset
  176. */
  177. public function charsetChanged($charset)
  178. {
  179. $this->_charStream->setCharacterSet($charset);
  180. }
  181. // -- Protected methods
  182. /**
  183. * Encode the given byte array into a verbatim QP form.
  184. * @param int[] $bytes
  185. * @return string
  186. * @access protected
  187. */
  188. protected function _encodeByteSequence(array $bytes, &$size)
  189. {
  190. $ret = '';
  191. $size=0;
  192. foreach ($bytes as $b)
  193. {
  194. if (isset($this->_safeMap[$b]))
  195. {
  196. $ret .= $this->_safeMap[$b];
  197. ++$size;
  198. }
  199. else
  200. {
  201. $ret .= $this->_qpMap[$b];
  202. $size+=3;
  203. }
  204. }
  205. return $ret;
  206. }
  207. /**
  208. * Get the next sequence of bytes to read from the char stream.
  209. * @param int $size number of bytes to read
  210. * @return int[]
  211. * @access protected
  212. */
  213. protected function _nextSequence($size = 4)
  214. {
  215. return $this->_charStream->readBytes($size);
  216. }
  217. /**
  218. * Make sure CRLF is correct and HT/SPACE are in valid places.
  219. * @param string $string
  220. * @return string
  221. * @access protected
  222. */
  223. protected function _standardize($string)
  224. {
  225. $string = str_replace(array("\t=0D=0A", " =0D=0A", "=0D=0A"),
  226. array("=09\r\n", "=20\r\n", "\r\n"), $string
  227. );
  228. switch ($end = ord(substr($string, -1)))
  229. {
  230. case 0x09:
  231. case 0x20:
  232. $string = substr_replace($string, $this->_qpMap[$end], -1);
  233. }
  234. return $string;
  235. }
  236. }