| 117 | | /** |
| 118 | | * constructs a new JSON instance |
| 119 | | * |
| 120 | | * @param int $use object behavior flags; combine with boolean-OR |
| 121 | | * |
| 122 | | * possible values: |
| 123 | | * - SERVICES_JSON_LOOSE_TYPE: loose typing. |
| 124 | | * "{...}" syntax creates associative arrays |
| 125 | | * instead of objects in decode(). |
| 126 | | * - SERVICES_JSON_SUPPRESS_ERRORS: error suppression. |
| 127 | | * Values which can't be encoded (e.g. resources) |
| 128 | | * appear as NULL instead of throwing errors. |
| 129 | | * By default, a deeply-nested resource will |
| 130 | | * bubble up with an error, so all return values |
| 131 | | * from encode() should be checked with isError() |
| 132 | | */ |
| 133 | | function Services_JSON($use = 0) |
| 134 | | { |
| 135 | | $this->use = $use; |
| 136 | | } |
| 137 | | |
| 138 | | /** |
| 139 | | * convert a string from one UTF-16 char to one UTF-8 char |
| 140 | | * |
| 141 | | * Normally should be handled by mb_convert_encoding, but |
| 142 | | * provides a slower PHP-only method for installations |
| 143 | | * that lack the multibye string extension. |
| 144 | | * |
| 145 | | * @param string $utf16 UTF-16 character |
| 146 | | * @return string UTF-8 character |
| 147 | | * @access private |
| 148 | | */ |
| 149 | | function utf162utf8($utf16) |
| 150 | | { |
| 151 | | // oh please oh please oh please oh please oh please |
| 152 | | if(function_exists('mb_convert_encoding')) { |
| 153 | | return mb_convert_encoding($utf16, 'UTF-8', 'UTF-16'); |
| 154 | | } |
| 155 | | |
| 156 | | $bytes = (ord($utf16[0]) << 8) | ord($utf16[1]); |
| 157 | | |
| 158 | | switch(true) { |
| 159 | | case ((0x7F & $bytes) == $bytes): |
| 160 | | // this case should never be reached, because we are in ASCII range |
| 161 | | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 162 | | return chr(0x7F & $bytes); |
| 163 | | |
| 164 | | case (0x07FF & $bytes) == $bytes: |
| 165 | | // return a 2-byte UTF-8 character |
| 166 | | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 167 | | return chr(0xC0 | (($bytes >> 6) & 0x1F)) |
| 168 | | . chr(0x80 | ($bytes & 0x3F)); |
| 169 | | |
| 170 | | case (0xFFFF & $bytes) == $bytes: |
| 171 | | // return a 3-byte UTF-8 character |
| 172 | | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 173 | | return chr(0xE0 | (($bytes >> 12) & 0x0F)) |
| 174 | | . chr(0x80 | (($bytes >> 6) & 0x3F)) |
| 175 | | . chr(0x80 | ($bytes & 0x3F)); |
| 176 | | } |
| 177 | | |
| 178 | | // ignoring UTF-32 for now, sorry |
| 179 | | return ''; |
| 180 | | } |
| 181 | | |
| 182 | | /** |
| 183 | | * convert a string from one UTF-8 char to one UTF-16 char |
| 184 | | * |
| 185 | | * Normally should be handled by mb_convert_encoding, but |
| 186 | | * provides a slower PHP-only method for installations |
| 187 | | * that lack the multibye string extension. |
| 188 | | * |
| 189 | | * @param string $utf8 UTF-8 character |
| 190 | | * @return string UTF-16 character |
| 191 | | * @access private |
| 192 | | */ |
| 193 | | function utf82utf16($utf8) |
| 194 | | { |
| 195 | | // oh please oh please oh please oh please oh please |
| 196 | | if(function_exists('mb_convert_encoding')) { |
| 197 | | return mb_convert_encoding($utf8, 'UTF-16', 'UTF-8'); |
| 198 | | } |
| 199 | | |
| 200 | | switch(strlen($utf8)) { |
| 201 | | case 1: |
| 202 | | // this case should never be reached, because we are in ASCII range |
| 203 | | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 204 | | return $utf8; |
| 205 | | |
| 206 | | case 2: |
| 207 | | // return a UTF-16 character from a 2-byte UTF-8 char |
| 208 | | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 209 | | return chr(0x07 & (ord($utf8[0]) >> 2)) |
| 210 | | . chr((0xC0 & (ord($utf8[0]) << 6)) |
| 211 | | | (0x3F & ord($utf8[1]))); |
| 212 | | |
| 213 | | case 3: |
| 214 | | // return a UTF-16 character from a 3-byte UTF-8 char |
| 215 | | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 216 | | return chr((0xF0 & (ord($utf8[0]) << 4)) |
| 217 | | | (0x0F & (ord($utf8[1]) >> 2))) |
| 218 | | . chr((0xC0 & (ord($utf8[1]) << 6)) |
| 219 | | | (0x7F & ord($utf8[2]))); |
| 220 | | } |
| 221 | | |
| 222 | | // ignoring UTF-32 for now, sorry |
| 223 | | return ''; |
| 224 | | } |
| 225 | | |
| 226 | | /** |
| 227 | | * encodes an arbitrary variable into JSON format (and sends JSON Header) |
| 228 | | * |
| 229 | | * @param mixed $var any number, boolean, string, array, or object to be encoded. |
| 230 | | * see argument 1 to Services_JSON() above for array-parsing behavior. |
| 231 | | * if var is a strng, note that encode() always expects it |
| 232 | | * to be in ASCII or UTF-8 format! |
| 233 | | * |
| 234 | | * @return mixed JSON string representation of input var or an error if a problem occurs |
| 235 | | * @access public |
| 236 | | */ |
| 237 | | function encode($var) |
| 238 | | { |
| 239 | | header('Content-type: application/json'); |
| 240 | | return $this->_encode($var); |
| 241 | | } |
| 242 | | /** |
| 243 | | * encodes an arbitrary variable into JSON format without JSON Header - warning - may allow CSS!!!!) |
| 244 | | * |
| 245 | | * @param mixed $var any number, boolean, string, array, or object to be encoded. |
| 246 | | * see argument 1 to Services_JSON() above for array-parsing behavior. |
| 247 | | * if var is a strng, note that encode() always expects it |
| 248 | | * to be in ASCII or UTF-8 format! |
| 249 | | * |
| 250 | | * @return mixed JSON string representation of input var or an error if a problem occurs |
| 251 | | * @access public |
| 252 | | */ |
| 253 | | function encodeUnsafe($var) |
| 254 | | { |
| 255 | | return $this->_encode($var); |
| 256 | | } |
| 257 | | /** |
| 258 | | * PRIVATE CODE that does the work of encodes an arbitrary variable into JSON format |
| 259 | | * |
| 260 | | * @param mixed $var any number, boolean, string, array, or object to be encoded. |
| 261 | | * see argument 1 to Services_JSON() above for array-parsing behavior. |
| 262 | | * if var is a strng, note that encode() always expects it |
| 263 | | * to be in ASCII or UTF-8 format! |
| 264 | | * |
| 265 | | * @return mixed JSON string representation of input var or an error if a problem occurs |
| 266 | | * @access public |
| 267 | | */ |
| 268 | | function _encode($var) |
| 269 | | { |
| 270 | | |
| 271 | | switch (gettype($var)) { |
| 272 | | case 'boolean': |
| 273 | | return $var ? 'true' : 'false'; |
| 274 | | |
| 275 | | case 'NULL': |
| 276 | | return 'null'; |
| 277 | | |
| 278 | | case 'integer': |
| 279 | | return (int) $var; |
| 280 | | |
| 281 | | case 'double': |
| 282 | | case 'float': |
| 283 | | return (float) $var; |
| 284 | | |
| 285 | | case 'string': |
| 286 | | // STRINGS ARE EXPECTED TO BE IN ASCII OR UTF-8 FORMAT |
| 287 | | $ascii = ''; |
| 288 | | $strlen_var = strlen($var); |
| 289 | | |
| 290 | | /* |
| 291 | | * Iterate over every character in the string, |
| 292 | | * escaping with a slash or encoding to UTF-8 where necessary |
| 293 | | */ |
| 294 | | for ($c = 0; $c < $strlen_var; ++$c) { |
| 295 | | |
| 296 | | $ord_var_c = ord($var[$c]); |
| 297 | | |
| 298 | | switch (true) { |
| 299 | | case $ord_var_c == 0x08: |
| 300 | | $ascii .= '\b'; |
| 301 | | break; |
| 302 | | case $ord_var_c == 0x09: |
| 303 | | $ascii .= '\t'; |
| 304 | | break; |
| 305 | | case $ord_var_c == 0x0A: |
| 306 | | $ascii .= '\n'; |
| 307 | | break; |
| 308 | | case $ord_var_c == 0x0C: |
| 309 | | $ascii .= '\f'; |
| 310 | | break; |
| 311 | | case $ord_var_c == 0x0D: |
| 312 | | $ascii .= '\r'; |
| 313 | | break; |
| 314 | | |
| 315 | | case $ord_var_c == 0x22: |
| 316 | | case $ord_var_c == 0x2F: |
| 317 | | case $ord_var_c == 0x5C: |
| 318 | | // double quote, slash, slosh |
| 319 | | $ascii .= '\\'.$var[$c]; |
| 320 | | break; |
| 321 | | |
| 322 | | case (($ord_var_c >= 0x20) && ($ord_var_c <= 0x7F)): |
| 323 | | // characters U-00000000 - U-0000007F (same as ASCII) |
| 324 | | $ascii .= $var[$c]; |
| 325 | | break; |
| 326 | | |
| 327 | | case (($ord_var_c & 0xE0) == 0xC0): |
| 328 | | // characters U-00000080 - U-000007FF, mask 110XXXXX |
| 329 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 330 | | if ($c+1 >= $strlen_var) { |
| 331 | | $c += 1; |
| 332 | | $ascii .= '?'; |
| 333 | | break; |
| 334 | | } |
| 335 | | |
| 336 | | $char = pack('C*', $ord_var_c, ord($var[$c + 1])); |
| 337 | | $c += 1; |
| 338 | | $utf16 = $this->utf82utf16($char); |
| 339 | | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| 340 | | break; |
| 341 | | |
| 342 | | case (($ord_var_c & 0xF0) == 0xE0): |
| 343 | | if ($c+2 >= $strlen_var) { |
| 344 | | $c += 2; |
| 345 | | $ascii .= '?'; |
| 346 | | break; |
| 347 | | } |
| 348 | | // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
| 349 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 350 | | $char = pack('C*', $ord_var_c, |
| 351 | | @ord($var[$c + 1]), |
| 352 | | @ord($var[$c + 2])); |
| 353 | | $c += 2; |
| 354 | | $utf16 = $this->utf82utf16($char); |
| 355 | | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| 356 | | break; |
| 357 | | |
| 358 | | case (($ord_var_c & 0xF8) == 0xF0): |
| 359 | | if ($c+3 >= $strlen_var) { |
| 360 | | $c += 3; |
| 361 | | $ascii .= '?'; |
| 362 | | break; |
| 363 | | } |
| 364 | | // characters U-00010000 - U-001FFFFF, mask 11110XXX |
| 365 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 366 | | $char = pack('C*', $ord_var_c, |
| 367 | | ord($var[$c + 1]), |
| 368 | | ord($var[$c + 2]), |
| 369 | | ord($var[$c + 3])); |
| 370 | | $c += 3; |
| 371 | | $utf16 = $this->utf82utf16($char); |
| 372 | | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| 373 | | break; |
| 374 | | |
| 375 | | case (($ord_var_c & 0xFC) == 0xF8): |
| 376 | | // characters U-00200000 - U-03FFFFFF, mask 111110XX |
| 377 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 378 | | if ($c+4 >= $strlen_var) { |
| 379 | | $c += 4; |
| 380 | | $ascii .= '?'; |
| 381 | | break; |
| 382 | | } |
| 383 | | $char = pack('C*', $ord_var_c, |
| 384 | | ord($var[$c + 1]), |
| 385 | | ord($var[$c + 2]), |
| 386 | | ord($var[$c + 3]), |
| 387 | | ord($var[$c + 4])); |
| 388 | | $c += 4; |
| 389 | | $utf16 = $this->utf82utf16($char); |
| 390 | | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| 391 | | break; |
| 392 | | |
| 393 | | case (($ord_var_c & 0xFE) == 0xFC): |
| 394 | | if ($c+5 >= $strlen_var) { |
| 395 | | $c += 5; |
| 396 | | $ascii .= '?'; |
| 397 | | break; |
| 398 | | } |
| 399 | | // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
| 400 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 401 | | $char = pack('C*', $ord_var_c, |
| 402 | | ord($var[$c + 1]), |
| 403 | | ord($var[$c + 2]), |
| 404 | | ord($var[$c + 3]), |
| 405 | | ord($var[$c + 4]), |
| 406 | | ord($var[$c + 5])); |
| 407 | | $c += 5; |
| 408 | | $utf16 = $this->utf82utf16($char); |
| 409 | | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| 410 | | break; |
| 411 | | } |
| 412 | | } |
| 413 | | return '"'.$ascii.'"'; |
| 414 | | |
| 415 | | case 'array': |
| 416 | | /* |
| 417 | | * As per JSON spec if any array key is not an integer |
| 418 | | * we must treat the the whole array as an object. We |
| 419 | | * also try to catch a sparsely populated associative |
| 420 | | * array with numeric keys here because some JS engines |
| 421 | | * will create an array with empty indexes up to |
| 422 | | * max_index which can cause memory issues and because |
| 423 | | * the keys, which may be relevant, will be remapped |
| 424 | | * otherwise. |
| 425 | | * |
| 426 | | * As per the ECMA and JSON specification an object may |
| 427 | | * have any string as a property. Unfortunately due to |
| 428 | | * a hole in the ECMA specification if the key is a |
| 429 | | * ECMA reserved word or starts with a digit the |
| 430 | | * parameter is only accessible using ECMAScript's |
| 431 | | * bracket notation. |
| 432 | | */ |
| 433 | | |
| 434 | | // treat as a JSON object |
| 435 | | if (is_array($var) && count($var) && (array_keys($var) !== range(0, sizeof($var) - 1))) { |
| 436 | | $properties = array_map(array($this, 'name_value'), |
| 437 | | array_keys($var), |
| 438 | | array_values($var)); |
| 439 | | |
| 440 | | foreach($properties as $property) { |
| 441 | | if(Services_JSON::isError($property)) { |
| 442 | | return $property; |
| 443 | | } |
| 444 | | } |
| 445 | | |
| 446 | | return '{' . join(',', $properties) . '}'; |
| 447 | | } |
| 448 | | |
| 449 | | // treat it like a regular array |
| 450 | | $elements = array_map(array($this, '_encode'), $var); |
| 451 | | |
| 452 | | foreach($elements as $element) { |
| 453 | | if(Services_JSON::isError($element)) { |
| 454 | | return $element; |
| 455 | | } |
| 456 | | } |
| 457 | | |
| 458 | | return '[' . join(',', $elements) . ']'; |
| 459 | | |
| 460 | | case 'object': |
| 461 | | $vars = get_object_vars($var); |
| 462 | | |
| 463 | | $properties = array_map(array($this, 'name_value'), |
| 464 | | array_keys($vars), |
| 465 | | array_values($vars)); |
| 466 | | |
| 467 | | foreach($properties as $property) { |
| 468 | | if(Services_JSON::isError($property)) { |
| 469 | | return $property; |
| 470 | | } |
| 471 | | } |
| 472 | | |
| 473 | | return '{' . join(',', $properties) . '}'; |
| 474 | | |
| 475 | | default: |
| 476 | | return ($this->use & SERVICES_JSON_SUPPRESS_ERRORS) |
| 477 | | ? 'null' |
| 478 | | : new Services_JSON_Error(gettype($var)." can not be encoded as JSON string"); |
| 479 | | } |
| 480 | | } |
| 481 | | |
| 482 | | /** |
| 483 | | * array-walking function for use in generating JSON-formatted name-value pairs |
| 484 | | * |
| 485 | | * @param string $name name of key to use |
| 486 | | * @param mixed $value reference to an array element to be encoded |
| 487 | | * |
| 488 | | * @return string JSON-formatted name-value pair, like '"name":value' |
| 489 | | * @access private |
| 490 | | */ |
| 491 | | function name_value($name, $value) |
| 492 | | { |
| 493 | | $encoded_value = $this->_encode($value); |
| 494 | | |
| 495 | | if(Services_JSON::isError($encoded_value)) { |
| 496 | | return $encoded_value; |
| 497 | | } |
| 498 | | |
| 499 | | return $this->_encode(strval($name)) . ':' . $encoded_value; |
| 500 | | } |
| 501 | | |
| 502 | | /** |
| 503 | | * reduce a string by removing leading and trailing comments and whitespace |
| 504 | | * |
| 505 | | * @param $str string string value to strip of comments and whitespace |
| 506 | | * |
| 507 | | * @return string string value stripped of comments and whitespace |
| 508 | | * @access private |
| 509 | | */ |
| 510 | | function reduce_string($str) |
| 511 | | { |
| 512 | | $str = preg_replace(array( |
| 513 | | |
| 514 | | // eliminate single line comments in '// ...' form |
| 515 | | '#^\s*//(.+)$#m', |
| 516 | | |
| 517 | | // eliminate multi-line comments in '/* ... */' form, at start of string |
| 518 | | '#^\s*/\*(.+)\*/#Us', |
| 519 | | |
| 520 | | // eliminate multi-line comments in '/* ... */' form, at end of string |
| 521 | | '#/\*(.+)\*/\s*$#Us' |
| 522 | | |
| 523 | | ), '', $str); |
| 524 | | |
| 525 | | // eliminate extraneous space |
| 526 | | return trim($str); |
| 527 | | } |
| 528 | | |
| 529 | | /** |
| 530 | | * decodes a JSON string into appropriate variable |
| 531 | | * |
| 532 | | * @param string $str JSON-formatted string |
| 533 | | * |
| 534 | | * @return mixed number, boolean, string, array, or object |
| 535 | | * corresponding to given JSON input string. |
| 536 | | * See argument 1 to Services_JSON() above for object-output behavior. |
| 537 | | * Note that decode() always returns strings |
| 538 | | * in ASCII or UTF-8 format! |
| 539 | | * @access public |
| 540 | | */ |
| 541 | | function decode($str) |
| 542 | | { |
| 543 | | $str = $this->reduce_string($str); |
| 544 | | |
| 545 | | switch (strtolower($str)) { |
| 546 | | case 'true': |
| 547 | | return true; |
| 548 | | |
| 549 | | case 'false': |
| 550 | | return false; |
| 551 | | |
| 552 | | case 'null': |
| 553 | | return null; |
| 554 | | |
| 555 | | default: |
| 556 | | $m = array(); |
| 557 | | |
| 558 | | if (is_numeric($str)) { |
| 559 | | // Lookie-loo, it's a number |
| 560 | | |
| 561 | | // This would work on its own, but I'm trying to be |
| 562 | | // good about returning integers where appropriate: |
| 563 | | // return (float)$str; |
| 564 | | |
| 565 | | // Return float or int, as appropriate |
| 566 | | return ((float)$str == (integer)$str) |
| 567 | | ? (integer)$str |
| 568 | | : (float)$str; |
| 569 | | |
| 570 | | } elseif (preg_match('/^("|\').*(\1)$/s', $str, $m) && $m[1] == $m[2]) { |
| 571 | | // STRINGS RETURNED IN UTF-8 FORMAT |
| 572 | | $delim = substr($str, 0, 1); |
| 573 | | $chrs = substr($str, 1, -1); |
| 574 | | $utf8 = ''; |
| 575 | | $strlen_chrs = strlen($chrs); |
| 576 | | |
| 577 | | for ($c = 0; $c < $strlen_chrs; ++$c) { |
| 578 | | |
| 579 | | $substr_chrs_c_2 = substr($chrs, $c, 2); |
| 580 | | $ord_chrs_c = ord($chrs[$c]); |
| 581 | | |
| 582 | | switch (true) { |
| 583 | | case $substr_chrs_c_2 == '\b': |
| 584 | | $utf8 .= chr(0x08); |
| 585 | | ++$c; |
| 586 | | break; |
| 587 | | case $substr_chrs_c_2 == '\t': |
| 588 | | $utf8 .= chr(0x09); |
| 589 | | ++$c; |
| 590 | | break; |
| 591 | | case $substr_chrs_c_2 == '\n': |
| 592 | | $utf8 .= chr(0x0A); |
| 593 | | ++$c; |
| 594 | | break; |
| 595 | | case $substr_chrs_c_2 == '\f': |
| 596 | | $utf8 .= chr(0x0C); |
| 597 | | ++$c; |
| 598 | | break; |
| 599 | | case $substr_chrs_c_2 == '\r': |
| 600 | | $utf8 .= chr(0x0D); |
| 601 | | ++$c; |
| 602 | | break; |
| 603 | | |
| 604 | | case $substr_chrs_c_2 == '\\"': |
| 605 | | case $substr_chrs_c_2 == '\\\'': |
| 606 | | case $substr_chrs_c_2 == '\\\\': |
| 607 | | case $substr_chrs_c_2 == '\\/': |
| 608 | | if (($delim == '"' && $substr_chrs_c_2 != '\\\'') || |
| 609 | | ($delim == "'" && $substr_chrs_c_2 != '\\"')) { |
| 610 | | $utf8 .= $chrs[++$c]; |
| 611 | | } |
| 612 | | break; |
| 613 | | |
| 614 | | case preg_match('/\\\u[0-9A-F]{4}/i', substr($chrs, $c, 6)): |
| 615 | | // single, escaped unicode character |
| 616 | | $utf16 = chr(hexdec(substr($chrs, ($c + 2), 2))) |
| 617 | | . chr(hexdec(substr($chrs, ($c + 4), 2))); |
| 618 | | $utf8 .= $this->utf162utf8($utf16); |
| 619 | | $c += 5; |
| 620 | | break; |
| 621 | | |
| 622 | | case ($ord_chrs_c >= 0x20) && ($ord_chrs_c <= 0x7F): |
| 623 | | $utf8 .= $chrs[$c]; |
| 624 | | break; |
| 625 | | |
| 626 | | case ($ord_chrs_c & 0xE0) == 0xC0: |
| 627 | | // characters U-00000080 - U-000007FF, mask 110XXXXX |
| 628 | | //see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 629 | | $utf8 .= substr($chrs, $c, 2); |
| 630 | | ++$c; |
| 631 | | break; |
| 632 | | |
| 633 | | case ($ord_chrs_c & 0xF0) == 0xE0: |
| 634 | | // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
| 635 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 636 | | $utf8 .= substr($chrs, $c, 3); |
| 637 | | $c += 2; |
| 638 | | break; |
| 639 | | |
| 640 | | case ($ord_chrs_c & 0xF8) == 0xF0: |
| 641 | | // characters U-00010000 - U-001FFFFF, mask 11110XXX |
| 642 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 643 | | $utf8 .= substr($chrs, $c, 4); |
| 644 | | $c += 3; |
| 645 | | break; |
| 646 | | |
| 647 | | case ($ord_chrs_c & 0xFC) == 0xF8: |
| 648 | | // characters U-00200000 - U-03FFFFFF, mask 111110XX |
| 649 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 650 | | $utf8 .= substr($chrs, $c, 5); |
| 651 | | $c += 4; |
| 652 | | break; |
| 653 | | |
| 654 | | case ($ord_chrs_c & 0xFE) == 0xFC: |
| 655 | | // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
| 656 | | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| 657 | | $utf8 .= substr($chrs, $c, 6); |
| 658 | | $c += 5; |
| 659 | | break; |
| 660 | | |
| 661 | | } |
| 662 | | |
| 663 | | } |
| 664 | | |
| 665 | | return $utf8; |
| 666 | | |
| 667 | | } elseif (preg_match('/^\[.*\]$/s', $str) || preg_match('/^\{.*\}$/s', $str)) { |
| 668 | | // array, or object notation |
| 669 | | |
| 670 | | if ($str[0] == '[') { |
| 671 | | $stk = array(SERVICES_JSON_IN_ARR); |
| 672 | | $arr = array(); |
| 673 | | } else { |
| 674 | | if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
| 675 | | $stk = array(SERVICES_JSON_IN_OBJ); |
| 676 | | $obj = array(); |
| 677 | | } else { |
| 678 | | $stk = array(SERVICES_JSON_IN_OBJ); |
| 679 | | $obj = new stdClass(); |
| 680 | | } |
| 681 | | } |
| 682 | | |
| 683 | | array_push($stk, array('what' => SERVICES_JSON_SLICE, |
| 684 | | 'where' => 0, |
| 685 | | 'delim' => false)); |
| 686 | | |
| 687 | | $chrs = substr($str, 1, -1); |
| 688 | | $chrs = $this->reduce_string($chrs); |
| 689 | | |
| 690 | | if ($chrs == '') { |
| 691 | | if (reset($stk) == SERVICES_JSON_IN_ARR) { |
| 692 | | return $arr; |
| 693 | | |
| 694 | | } else { |
| 695 | | return $obj; |
| 696 | | |
| 697 | | } |
| 698 | | } |
| 699 | | |
| 700 | | //print("\nparsing {$chrs}\n"); |
| 701 | | |
| 702 | | $strlen_chrs = strlen($chrs); |
| 703 | | |
| 704 | | for ($c = 0; $c <= $strlen_chrs; ++$c) { |
| 705 | | |
| 706 | | $top = end($stk); |
| 707 | | $substr_chrs_c_2 = substr($chrs, $c, 2); |
| 708 | | |
| 709 | | if (($c == $strlen_chrs) || (($chrs[$c] == ',') && ($top['what'] == SERVICES_JSON_SLICE))) { |
| 710 | | // found a comma that is not inside a string, array, etc., |
| 711 | | // OR we've reached the end of the character list |
| 712 | | $slice = substr($chrs, $top['where'], ($c - $top['where'])); |
| 713 | | array_push($stk, array('what' => SERVICES_JSON_SLICE, 'where' => ($c + 1), 'delim' => false)); |
| 714 | | //print("Found split at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| 715 | | |
| 716 | | if (reset($stk) == SERVICES_JSON_IN_ARR) { |
| 717 | | // we are in an array, so just push an element onto the stack |
| 718 | | array_push($arr, $this->decode($slice)); |
| 719 | | |
| 720 | | } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
| 721 | | // we are in an object, so figure |
| 722 | | // out the property name and set an |
| 723 | | // element in an associative array, |
| 724 | | // for now |
| 725 | | $parts = array(); |
| 726 | | |
| 727 | | if (preg_match('/^\s*(["\'].*[^\\\]["\'])\s*:\s*(\S.*),?$/Uis', $slice, $parts)) { |
| 728 | | // "name":value pair |
| 729 | | $key = $this->decode($parts[1]); |
| 730 | | $val = $this->decode($parts[2]); |
| 731 | | |
| 732 | | if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
| 733 | | $obj[$key] = $val; |
| 734 | | } else { |
| 735 | | $obj->$key = $val; |
| 736 | | } |
| 737 | | } elseif (preg_match('/^\s*(\w+)\s*:\s*(\S.*),?$/Uis', $slice, $parts)) { |
| 738 | | // name:value pair, where name is unquoted |
| 739 | | $key = $parts[1]; |
| 740 | | $val = $this->decode($parts[2]); |
| 741 | | |
| 742 | | if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
| 743 | | $obj[$key] = $val; |
| 744 | | } else { |
| 745 | | $obj->$key = $val; |
| 746 | | } |
| 747 | | } |
| 748 | | |
| 749 | | } |
| 750 | | |
| 751 | | } elseif ((($chrs[$c] == '"') || ($chrs[$c] == "'")) && ($top['what'] != SERVICES_JSON_IN_STR)) { |
| 752 | | // found a quote, and we are not inside a string |
| 753 | | array_push($stk, array('what' => SERVICES_JSON_IN_STR, 'where' => $c, 'delim' => $chrs[$c])); |
| 754 | | //print("Found start of string at {$c}\n"); |
| 755 | | |
| 756 | | } elseif (($chrs[$c] == $top['delim']) && |
| 757 | | ($top['what'] == SERVICES_JSON_IN_STR) && |
| 758 | | ((strlen(substr($chrs, 0, $c)) - strlen(rtrim(substr($chrs, 0, $c), '\\'))) % 2 != 1)) { |
| 759 | | // found a quote, we're in a string, and it's not escaped |
| 760 | | // we know that it's not escaped becase there is _not_ an |
| 761 | | // odd number of backslashes at the end of the string so far |
| 762 | | array_pop($stk); |
| 763 | | //print("Found end of string at {$c}: ".substr($chrs, $top['where'], (1 + 1 + $c - $top['where']))."\n"); |
| 764 | | |
| 765 | | } elseif (($chrs[$c] == '[') && |
| 766 | | in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
| 767 | | // found a left-bracket, and we are in an array, object, or slice |
| 768 | | array_push($stk, array('what' => SERVICES_JSON_IN_ARR, 'where' => $c, 'delim' => false)); |
| 769 | | //print("Found start of array at {$c}\n"); |
| 770 | | |
| 771 | | } elseif (($chrs[$c] == ']') && ($top['what'] == SERVICES_JSON_IN_ARR)) { |
| 772 | | // found a right-bracket, and we're in an array |
| 773 | | array_pop($stk); |
| 774 | | //print("Found end of array at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| 775 | | |
| 776 | | } elseif (($chrs[$c] == '{') && |
| 777 | | in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
| 778 | | // found a left-brace, and we are in an array, object, or slice |
| 779 | | array_push($stk, array('what' => SERVICES_JSON_IN_OBJ, 'where' => $c, 'delim' => false)); |
| 780 | | //print("Found start of object at {$c}\n"); |
| 781 | | |
| 782 | | } elseif (($chrs[$c] == '}') && ($top['what'] == SERVICES_JSON_IN_OBJ)) { |
| 783 | | // found a right-brace, and we're in an object |
| 784 | | array_pop($stk); |
| 785 | | //print("Found end of object at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| 786 | | |
| 787 | | } elseif (($substr_chrs_c_2 == '/*') && |
| 788 | | in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
| 789 | | // found a comment start, and we are in an array, object, or slice |
| 790 | | array_push($stk, array('what' => SERVICES_JSON_IN_CMT, 'where' => $c, 'delim' => false)); |
| 791 | | $c++; |
| 792 | | //print("Found start of comment at {$c}\n"); |
| 793 | | |
| 794 | | } elseif (($substr_chrs_c_2 == '*/') && ($top['what'] == SERVICES_JSON_IN_CMT)) { |
| 795 | | // found a comment end, and we're in one now |
| 796 | | array_pop($stk); |
| 797 | | $c++; |
| 798 | | |
| 799 | | for ($i = $top['where']; $i <= $c; ++$i) |
| 800 | | $chrs = substr_replace($chrs, ' ', $i, 1); |
| 801 | | |
| 802 | | //print("Found end of comment at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| 803 | | |
| 804 | | } |
| 805 | | |
| 806 | | } |
| 807 | | |
| 808 | | if (reset($stk) == SERVICES_JSON_IN_ARR) { |
| 809 | | return $arr; |
| 810 | | |
| 811 | | } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
| 812 | | return $obj; |
| 813 | | |
| 814 | | } |
| 815 | | |
| 816 | | } |
| 817 | | } |
| 818 | | } |
| 819 | | |
| 820 | | /** |
| 821 | | * @todo Ultimately, this should just call PEAR::isError() |
| 822 | | */ |
| 823 | | function isError($data, $code = null) |
| 824 | | { |
| 825 | | if (class_exists('pear')) { |
| 826 | | return PEAR::isError($data, $code); |
| 827 | | } elseif (is_object($data) && (get_class($data) == 'services_json_error' || |
| 828 | | is_subclass_of($data, 'services_json_error'))) { |
| 829 | | return true; |
| 830 | | } |
| 831 | | |
| 832 | | return false; |
| 833 | | } |
| | 122 | /** |
| | 123 | * constructs a new JSON instance |
| | 124 | * |
| | 125 | * @param int $use object behavior flags; combine with boolean-OR |
| | 126 | * |
| | 127 | * possible values: |
| | 128 | * - SERVICES_JSON_LOOSE_TYPE: loose typing. |
| | 129 | * "{...}" syntax creates associative arrays |
| | 130 | * instead of objects in decode(). |
| | 131 | * - SERVICES_JSON_SUPPRESS_ERRORS: error suppression. |
| | 132 | * Values which can't be encoded (e.g. resources) |
| | 133 | * appear as NULL instead of throwing errors. |
| | 134 | * By default, a deeply-nested resource will |
| | 135 | * bubble up with an error, so all return values |
| | 136 | * from encode() should be checked with isError() |
| | 137 | * - SERVICES_JSON_USE_TO_JSON: call toJSON when serializing objects |
| | 138 | * It serializes the return value from the toJSON call rather |
| | 139 | * than the object it'self, toJSON can return associative arrays, |
| | 140 | * strings or numbers, if you return an object, make sure it does |
| | 141 | * not have a toJSON method, otherwise an error will occur. |
| | 142 | */ |
| | 143 | function Services_JSON($use = 0) |
| | 144 | { |
| | 145 | $this->use = $use; |
| | 146 | $this->_mb_strlen = function_exists('mb_strlen'); |
| | 147 | $this->_mb_convert_encoding = function_exists('mb_convert_encoding'); |
| | 148 | $this->_mb_substr = function_exists('mb_substr'); |
| | 149 | } |
| | 150 | // private - cache the mbstring lookup results.. |
| | 151 | var $_mb_strlen = false; |
| | 152 | var $_mb_substr = false; |
| | 153 | var $_mb_convert_encoding = false; |
| | 154 | |
| | 155 | /** |
| | 156 | * convert a string from one UTF-16 char to one UTF-8 char |
| | 157 | * |
| | 158 | * Normally should be handled by mb_convert_encoding, but |
| | 159 | * provides a slower PHP-only method for installations |
| | 160 | * that lack the multibye string extension. |
| | 161 | * |
| | 162 | * @param string $utf16 UTF-16 character |
| | 163 | * @return string UTF-8 character |
| | 164 | * @access private |
| | 165 | */ |
| | 166 | function utf162utf8($utf16) |
| | 167 | { |
| | 168 | // oh please oh please oh please oh please oh please |
| | 169 | if($this->_mb_convert_encoding) { |
| | 170 | return mb_convert_encoding($utf16, 'UTF-8', 'UTF-16'); |
| | 171 | } |
| | 172 | |
| | 173 | $bytes = (ord($utf16{0}) << 8) | ord($utf16{1}); |
| | 174 | |
| | 175 | switch(true) { |
| | 176 | case ((0x7F & $bytes) == $bytes): |
| | 177 | // this case should never be reached, because we are in ASCII range |
| | 178 | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 179 | return chr(0x7F & $bytes); |
| | 180 | |
| | 181 | case (0x07FF & $bytes) == $bytes: |
| | 182 | // return a 2-byte UTF-8 character |
| | 183 | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 184 | return chr(0xC0 | (($bytes >> 6) & 0x1F)) |
| | 185 | . chr(0x80 | ($bytes & 0x3F)); |
| | 186 | |
| | 187 | case (0xFFFF & $bytes) == $bytes: |
| | 188 | // return a 3-byte UTF-8 character |
| | 189 | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 190 | return chr(0xE0 | (($bytes >> 12) & 0x0F)) |
| | 191 | . chr(0x80 | (($bytes >> 6) & 0x3F)) |
| | 192 | . chr(0x80 | ($bytes & 0x3F)); |
| | 193 | } |
| | 194 | |
| | 195 | // ignoring UTF-32 for now, sorry |
| | 196 | return ''; |
| | 197 | } |
| | 198 | |
| | 199 | /** |
| | 200 | * convert a string from one UTF-8 char to one UTF-16 char |
| | 201 | * |
| | 202 | * Normally should be handled by mb_convert_encoding, but |
| | 203 | * provides a slower PHP-only method for installations |
| | 204 | * that lack the multibye string extension. |
| | 205 | * |
| | 206 | * @param string $utf8 UTF-8 character |
| | 207 | * @return string UTF-16 character |
| | 208 | * @access private |
| | 209 | */ |
| | 210 | function utf82utf16($utf8) |
| | 211 | { |
| | 212 | // oh please oh please oh please oh please oh please |
| | 213 | if($this->_mb_convert_encoding) { |
| | 214 | return mb_convert_encoding($utf8, 'UTF-16', 'UTF-8'); |
| | 215 | } |
| | 216 | |
| | 217 | switch($this->strlen8($utf8)) { |
| | 218 | case 1: |
| | 219 | // this case should never be reached, because we are in ASCII range |
| | 220 | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 221 | return $utf8; |
| | 222 | |
| | 223 | case 2: |
| | 224 | // return a UTF-16 character from a 2-byte UTF-8 char |
| | 225 | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 226 | return chr(0x07 & (ord($utf8{0}) >> 2)) |
| | 227 | . chr((0xC0 & (ord($utf8{0}) << 6)) |
| | 228 | | (0x3F & ord($utf8{1}))); |
| | 229 | |
| | 230 | case 3: |
| | 231 | // return a UTF-16 character from a 3-byte UTF-8 char |
| | 232 | // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 233 | return chr((0xF0 & (ord($utf8{0}) << 4)) |
| | 234 | | (0x0F & (ord($utf8{1}) >> 2))) |
| | 235 | . chr((0xC0 & (ord($utf8{1}) << 6)) |
| | 236 | | (0x7F & ord($utf8{2}))); |
| | 237 | } |
| | 238 | |
| | 239 | // ignoring UTF-32 for now, sorry |
| | 240 | return ''; |
| | 241 | } |
| | 242 | |
| | 243 | /** |
| | 244 | * encodes an arbitrary variable into JSON format (and sends JSON Header) |
| | 245 | * |
| | 246 | * @param mixed $var any number, boolean, string, array, or object to be encoded. |
| | 247 | * see argument 1 to Services_JSON() above for array-parsing behavior. |
| | 248 | * if var is a strng, note that encode() always expects it |
| | 249 | * to be in ASCII or UTF-8 format! |
| | 250 | * |
| | 251 | * @return mixed JSON string representation of input var or an error if a problem occurs |
| | 252 | * @access public |
| | 253 | */ |
| | 254 | function encode($var) |
| | 255 | { |
| | 256 | header('Content-type: application/json'); |
| | 257 | return $this->encodeUnsafe($var); |
| | 258 | } |
| | 259 | /** |
| | 260 | * encodes an arbitrary variable into JSON format without JSON Header - warning - may allow XSS!!!!) |
| | 261 | * |
| | 262 | * @param mixed $var any number, boolean, string, array, or object to be encoded. |
| | 263 | * see argument 1 to Services_JSON() above for array-parsing behavior. |
| | 264 | * if var is a strng, note that encode() always expects it |
| | 265 | * to be in ASCII or UTF-8 format! |
| | 266 | * |
| | 267 | * @return mixed JSON string representation of input var or an error if a problem occurs |
| | 268 | * @access public |
| | 269 | */ |
| | 270 | function encodeUnsafe($var) |
| | 271 | { |
| | 272 | // see bug #16908 - regarding numeric locale printing |
| | 273 | $lc = setlocale(LC_NUMERIC, 0); |
| | 274 | setlocale(LC_NUMERIC, 'C'); |
| | 275 | $ret = $this->_encode($var); |
| | 276 | setlocale(LC_NUMERIC, $lc); |
| | 277 | return $ret; |
| | 278 | |
| | 279 | } |
| | 280 | /** |
| | 281 | * PRIVATE CODE that does the work of encodes an arbitrary variable into JSON format |
| | 282 | * |
| | 283 | * @param mixed $var any number, boolean, string, array, or object to be encoded. |
| | 284 | * see argument 1 to Services_JSON() above for array-parsing behavior. |
| | 285 | * if var is a strng, note that encode() always expects it |
| | 286 | * to be in ASCII or UTF-8 format! |
| | 287 | * |
| | 288 | * @return mixed JSON string representation of input var or an error if a problem occurs |
| | 289 | * @access public |
| | 290 | */ |
| | 291 | function _encode($var) |
| | 292 | { |
| | 293 | |
| | 294 | switch (gettype($var)) { |
| | 295 | case 'boolean': |
| | 296 | return $var ? 'true' : 'false'; |
| | 297 | |
| | 298 | case 'NULL': |
| | 299 | return 'null'; |
| | 300 | |
| | 301 | case 'integer': |
| | 302 | return (int) $var; |
| | 303 | |
| | 304 | case 'double': |
| | 305 | case 'float': |
| | 306 | return (float) $var; |
| | 307 | |
| | 308 | case 'string': |
| | 309 | // STRINGS ARE EXPECTED TO BE IN ASCII OR UTF-8 FORMAT |
| | 310 | $ascii = ''; |
| | 311 | $strlen_var = $this->strlen8($var); |
| | 312 | |
| | 313 | /* |
| | 314 | * Iterate over every character in the string, |
| | 315 | * escaping with a slash or encoding to UTF-8 where necessary |
| | 316 | */ |
| | 317 | for ($c = 0; $c < $strlen_var; ++$c) { |
| | 318 | |
| | 319 | $ord_var_c = ord($var{$c}); |
| | 320 | |
| | 321 | switch (true) { |
| | 322 | case $ord_var_c == 0x08: |
| | 323 | $ascii .= '\b'; |
| | 324 | break; |
| | 325 | case $ord_var_c == 0x09: |
| | 326 | $ascii .= '\t'; |
| | 327 | break; |
| | 328 | case $ord_var_c == 0x0A: |
| | 329 | $ascii .= '\n'; |
| | 330 | break; |
| | 331 | case $ord_var_c == 0x0C: |
| | 332 | $ascii .= '\f'; |
| | 333 | break; |
| | 334 | case $ord_var_c == 0x0D: |
| | 335 | $ascii .= '\r'; |
| | 336 | break; |
| | 337 | |
| | 338 | case $ord_var_c == 0x22: |
| | 339 | case $ord_var_c == 0x2F: |
| | 340 | case $ord_var_c == 0x5C: |
| | 341 | // double quote, slash, slosh |
| | 342 | $ascii .= '\\'.$var{$c}; |
| | 343 | break; |
| | 344 | |
| | 345 | case (($ord_var_c >= 0x20) && ($ord_var_c <= 0x7F)): |
| | 346 | // characters U-00000000 - U-0000007F (same as ASCII) |
| | 347 | $ascii .= $var{$c}; |
| | 348 | break; |
| | 349 | |
| | 350 | case (($ord_var_c & 0xE0) == 0xC0): |
| | 351 | // characters U-00000080 - U-000007FF, mask 110XXXXX |
| | 352 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 353 | if ($c+1 >= $strlen_var) { |
| | 354 | $c += 1; |
| | 355 | $ascii .= '?'; |
| | 356 | break; |
| | 357 | } |
| | 358 | |
| | 359 | $char = pack('C*', $ord_var_c, ord($var{$c + 1})); |
| | 360 | $c += 1; |
| | 361 | $utf16 = $this->utf82utf16($char); |
| | 362 | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| | 363 | break; |
| | 364 | |
| | 365 | case (($ord_var_c & 0xF0) == 0xE0): |
| | 366 | if ($c+2 >= $strlen_var) { |
| | 367 | $c += 2; |
| | 368 | $ascii .= '?'; |
| | 369 | break; |
| | 370 | } |
| | 371 | // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
| | 372 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 373 | $char = pack('C*', $ord_var_c, |
| | 374 | @ord($var{$c + 1}), |
| | 375 | @ord($var{$c + 2})); |
| | 376 | $c += 2; |
| | 377 | $utf16 = $this->utf82utf16($char); |
| | 378 | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| | 379 | break; |
| | 380 | |
| | 381 | case (($ord_var_c & 0xF8) == 0xF0): |
| | 382 | if ($c+3 >= $strlen_var) { |
| | 383 | $c += 3; |
| | 384 | $ascii .= '?'; |
| | 385 | break; |
| | 386 | } |
| | 387 | // characters U-00010000 - U-001FFFFF, mask 11110XXX |
| | 388 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 389 | $char = pack('C*', $ord_var_c, |
| | 390 | ord($var{$c + 1}), |
| | 391 | ord($var{$c + 2}), |
| | 392 | ord($var{$c + 3})); |
| | 393 | $c += 3; |
| | 394 | $utf16 = $this->utf82utf16($char); |
| | 395 | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| | 396 | break; |
| | 397 | |
| | 398 | case (($ord_var_c & 0xFC) == 0xF8): |
| | 399 | // characters U-00200000 - U-03FFFFFF, mask 111110XX |
| | 400 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 401 | if ($c+4 >= $strlen_var) { |
| | 402 | $c += 4; |
| | 403 | $ascii .= '?'; |
| | 404 | break; |
| | 405 | } |
| | 406 | $char = pack('C*', $ord_var_c, |
| | 407 | ord($var{$c + 1}), |
| | 408 | ord($var{$c + 2}), |
| | 409 | ord($var{$c + 3}), |
| | 410 | ord($var{$c + 4})); |
| | 411 | $c += 4; |
| | 412 | $utf16 = $this->utf82utf16($char); |
| | 413 | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| | 414 | break; |
| | 415 | |
| | 416 | case (($ord_var_c & 0xFE) == 0xFC): |
| | 417 | if ($c+5 >= $strlen_var) { |
| | 418 | $c += 5; |
| | 419 | $ascii .= '?'; |
| | 420 | break; |
| | 421 | } |
| | 422 | // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
| | 423 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 424 | $char = pack('C*', $ord_var_c, |
| | 425 | ord($var{$c + 1}), |
| | 426 | ord($var{$c + 2}), |
| | 427 | ord($var{$c + 3}), |
| | 428 | ord($var{$c + 4}), |
| | 429 | ord($var{$c + 5})); |
| | 430 | $c += 5; |
| | 431 | $utf16 = $this->utf82utf16($char); |
| | 432 | $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
| | 433 | break; |
| | 434 | } |
| | 435 | } |
| | 436 | return '"'.$ascii.'"'; |
| | 437 | |
| | 438 | case 'array': |
| | 439 | /* |
| | 440 | * As per JSON spec if any array key is not an integer |
| | 441 | * we must treat the the whole array as an object. We |
| | 442 | * also try to catch a sparsely populated associative |
| | 443 | * array with numeric keys here because some JS engines |
| | 444 | * will create an array with empty indexes up to |
| | 445 | * max_index which can cause memory issues and because |
| | 446 | * the keys, which may be relevant, will be remapped |
| | 447 | * otherwise. |
| | 448 | * |
| | 449 | * As per the ECMA and JSON specification an object may |
| | 450 | * have any string as a property. Unfortunately due to |
| | 451 | * a hole in the ECMA specification if the key is a |
| | 452 | * ECMA reserved word or starts with a digit the |
| | 453 | * parameter is only accessible using ECMAScript's |
| | 454 | * bracket notation. |
| | 455 | */ |
| | 456 | |
| | 457 | // treat as a JSON object |
| | 458 | if (is_array($var) && count($var) && (array_keys($var) !== range(0, sizeof($var) - 1))) { |
| | 459 | $properties = array_map(array($this, 'name_value'), |
| | 460 | array_keys($var), |
| | 461 | array_values($var)); |
| | 462 | |
| | 463 | foreach($properties as $property) { |
| | 464 | if(Services_JSON::isError($property)) { |
| | 465 | return $property; |
| | 466 | } |
| | 467 | } |
| | 468 | |
| | 469 | return '{' . join(',', $properties) . '}'; |
| | 470 | } |
| | 471 | |
| | 472 | // treat it like a regular array |
| | 473 | $elements = array_map(array($this, '_encode'), $var); |
| | 474 | |
| | 475 | foreach($elements as $element) { |
| | 476 | if(Services_JSON::isError($element)) { |
| | 477 | return $element; |
| | 478 | } |
| | 479 | } |
| | 480 | |
| | 481 | return '[' . join(',', $elements) . ']'; |
| | 482 | |
| | 483 | case 'object': |
| | 484 | |
| | 485 | // support toJSON methods. |
| | 486 | if (($this->use & SERVICES_JSON_USE_TO_JSON) && method_exists($var, 'toJSON')) { |
| | 487 | // this may end up allowing unlimited recursion |
| | 488 | // so we check the return value to make sure it's not got the same method. |
| | 489 | $recode = $var->toJSON(); |
| | 490 | |
| | 491 | if (method_exists($recode, 'toJSON')) { |
| | 492 | |
| | 493 | return ($this->use & SERVICES_JSON_SUPPRESS_ERRORS) |
| | 494 | ? 'null' |
| | 495 | : new Services_JSON_Error(class_name($var). |
| | 496 | " toJSON returned an object with a toJSON method."); |
| | 497 | |
| | 498 | } |
| | 499 | |
| | 500 | return $this->_encode( $recode ); |
| | 501 | } |
| | 502 | |
| | 503 | $vars = get_object_vars($var); |
| | 504 | |
| | 505 | $properties = array_map(array($this, 'name_value'), |
| | 506 | array_keys($vars), |
| | 507 | array_values($vars)); |
| | 508 | |
| | 509 | foreach($properties as $property) { |
| | 510 | if(Services_JSON::isError($property)) { |
| | 511 | return $property; |
| | 512 | } |
| | 513 | } |
| | 514 | |
| | 515 | return '{' . join(',', $properties) . '}'; |
| | 516 | |
| | 517 | default: |
| | 518 | return ($this->use & SERVICES_JSON_SUPPRESS_ERRORS) |
| | 519 | ? 'null' |
| | 520 | : new Services_JSON_Error(gettype($var)." can not be encoded as JSON string"); |
| | 521 | } |
| | 522 | } |
| | 523 | |
| | 524 | /** |
| | 525 | * array-walking function for use in generating JSON-formatted name-value pairs |
| | 526 | * |
| | 527 | * @param string $name name of key to use |
| | 528 | * @param mixed $value reference to an array element to be encoded |
| | 529 | * |
| | 530 | * @return string JSON-formatted name-value pair, like '"name":value' |
| | 531 | * @access private |
| | 532 | */ |
| | 533 | function name_value($name, $value) |
| | 534 | { |
| | 535 | $encoded_value = $this->_encode($value); |
| | 536 | |
| | 537 | if(Services_JSON::isError($encoded_value)) { |
| | 538 | return $encoded_value; |
| | 539 | } |
| | 540 | |
| | 541 | return $this->_encode(strval($name)) . ':' . $encoded_value; |
| | 542 | } |
| | 543 | |
| | 544 | /** |
| | 545 | * reduce a string by removing leading and trailing comments and whitespace |
| | 546 | * |
| | 547 | * @param $str string string value to strip of comments and whitespace |
| | 548 | * |
| | 549 | * @return string string value stripped of comments and whitespace |
| | 550 | * @access private |
| | 551 | */ |
| | 552 | function reduce_string($str) |
| | 553 | { |
| | 554 | $str = preg_replace(array( |
| | 555 | |
| | 556 | // eliminate single line comments in '// ...' form |
| | 557 | '#^\s*//(.+)$#m', |
| | 558 | |
| | 559 | // eliminate multi-line comments in '/* ... */' form, at start of string |
| | 560 | '#^\s*/\*(.+)\*/#Us', |
| | 561 | |
| | 562 | // eliminate multi-line comments in '/* ... */' form, at end of string |
| | 563 | '#/\*(.+)\*/\s*$#Us' |
| | 564 | |
| | 565 | ), '', $str); |
| | 566 | |
| | 567 | // eliminate extraneous space |
| | 568 | return trim($str); |
| | 569 | } |
| | 570 | |
| | 571 | /** |
| | 572 | * decodes a JSON string into appropriate variable |
| | 573 | * |
| | 574 | * @param string $str JSON-formatted string |
| | 575 | * |
| | 576 | * @return mixed number, boolean, string, array, or object |
| | 577 | * corresponding to given JSON input string. |
| | 578 | * See argument 1 to Services_JSON() above for object-output behavior. |
| | 579 | * Note that decode() always returns strings |
| | 580 | * in ASCII or UTF-8 format! |
| | 581 | * @access public |
| | 582 | */ |
| | 583 | function decode($str) |
| | 584 | { |
| | 585 | $str = $this->reduce_string($str); |
| | 586 | |
| | 587 | switch (strtolower($str)) { |
| | 588 | case 'true': |
| | 589 | return true; |
| | 590 | |
| | 591 | case 'false': |
| | 592 | return false; |
| | 593 | |
| | 594 | case 'null': |
| | 595 | return null; |
| | 596 | |
| | 597 | default: |
| | 598 | $m = array(); |
| | 599 | |
| | 600 | if (is_numeric($str)) { |
| | 601 | // Lookie-loo, it's a number |
| | 602 | |
| | 603 | // This would work on its own, but I'm trying to be |
| | 604 | // good about returning integers where appropriate: |
| | 605 | // return (float)$str; |
| | 606 | |
| | 607 | // Return float or int, as appropriate |
| | 608 | return ((float)$str == (integer)$str) |
| | 609 | ? (integer)$str |
| | 610 | : (float)$str; |
| | 611 | |
| | 612 | } elseif (preg_match('/^("|\').*(\1)$/s', $str, $m) && $m[1] == $m[2]) { |
| | 613 | // STRINGS RETURNED IN UTF-8 FORMAT |
| | 614 | $delim = $this->substr8($str, 0, 1); |
| | 615 | $chrs = $this->substr8($str, 1, -1); |
| | 616 | $utf8 = ''; |
| | 617 | $strlen_chrs = $this->strlen8($chrs); |
| | 618 | |
| | 619 | for ($c = 0; $c < $strlen_chrs; ++$c) { |
| | 620 | |
| | 621 | $substr_chrs_c_2 = $this->substr8($chrs, $c, 2); |
| | 622 | $ord_chrs_c = ord($chrs{$c}); |
| | 623 | |
| | 624 | switch (true) { |
| | 625 | case $substr_chrs_c_2 == '\b': |
| | 626 | $utf8 .= chr(0x08); |
| | 627 | ++$c; |
| | 628 | break; |
| | 629 | case $substr_chrs_c_2 == '\t': |
| | 630 | $utf8 .= chr(0x09); |
| | 631 | ++$c; |
| | 632 | break; |
| | 633 | case $substr_chrs_c_2 == '\n': |
| | 634 | $utf8 .= chr(0x0A); |
| | 635 | ++$c; |
| | 636 | break; |
| | 637 | case $substr_chrs_c_2 == '\f': |
| | 638 | $utf8 .= chr(0x0C); |
| | 639 | ++$c; |
| | 640 | break; |
| | 641 | case $substr_chrs_c_2 == '\r': |
| | 642 | $utf8 .= chr(0x0D); |
| | 643 | ++$c; |
| | 644 | break; |
| | 645 | |
| | 646 | case $substr_chrs_c_2 == '\\"': |
| | 647 | case $substr_chrs_c_2 == '\\\'': |
| | 648 | case $substr_chrs_c_2 == '\\\\': |
| | 649 | case $substr_chrs_c_2 == '\\/': |
| | 650 | if (($delim == '"' && $substr_chrs_c_2 != '\\\'') || |
| | 651 | ($delim == "'" && $substr_chrs_c_2 != '\\"')) { |
| | 652 | $utf8 .= $chrs{++$c}; |
| | 653 | } |
| | 654 | break; |
| | 655 | |
| | 656 | case preg_match('/\\\u[0-9A-F]{4}/i', $this->substr8($chrs, $c, 6)): |
| | 657 | // single, escaped unicode character |
| | 658 | $utf16 = chr(hexdec($this->substr8($chrs, ($c + 2), 2))) |
| | 659 | . chr(hexdec($this->substr8($chrs, ($c + 4), 2))); |
| | 660 | $utf8 .= $this->utf162utf8($utf16); |
| | 661 | $c += 5; |
| | 662 | break; |
| | 663 | |
| | 664 | case ($ord_chrs_c >= 0x20) && ($ord_chrs_c <= 0x7F): |
| | 665 | $utf8 .= $chrs{$c}; |
| | 666 | break; |
| | 667 | |
| | 668 | case ($ord_chrs_c & 0xE0) == 0xC0: |
| | 669 | // characters U-00000080 - U-000007FF, mask 110XXXXX |
| | 670 | //see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 671 | $utf8 .= $this->substr8($chrs, $c, 2); |
| | 672 | ++$c; |
| | 673 | break; |
| | 674 | |
| | 675 | case ($ord_chrs_c & 0xF0) == 0xE0: |
| | 676 | // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
| | 677 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 678 | $utf8 .= $this->substr8($chrs, $c, 3); |
| | 679 | $c += 2; |
| | 680 | break; |
| | 681 | |
| | 682 | case ($ord_chrs_c & 0xF8) == 0xF0: |
| | 683 | // characters U-00010000 - U-001FFFFF, mask 11110XXX |
| | 684 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 685 | $utf8 .= $this->substr8($chrs, $c, 4); |
| | 686 | $c += 3; |
| | 687 | break; |
| | 688 | |
| | 689 | case ($ord_chrs_c & 0xFC) == 0xF8: |
| | 690 | // characters U-00200000 - U-03FFFFFF, mask 111110XX |
| | 691 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 692 | $utf8 .= $this->substr8($chrs, $c, 5); |
| | 693 | $c += 4; |
| | 694 | break; |
| | 695 | |
| | 696 | case ($ord_chrs_c & 0xFE) == 0xFC: |
| | 697 | // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
| | 698 | // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
| | 699 | $utf8 .= $this->substr8($chrs, $c, 6); |
| | 700 | $c += 5; |
| | 701 | break; |
| | 702 | |
| | 703 | } |
| | 704 | |
| | 705 | } |
| | 706 | |
| | 707 | return $utf8; |
| | 708 | |
| | 709 | } elseif (preg_match('/^\[.*\]$/s', $str) || preg_match('/^\{.*\}$/s', $str)) { |
| | 710 | // array, or object notation |
| | 711 | |
| | 712 | if ($str{0} == '[') { |
| | 713 | $stk = array(SERVICES_JSON_IN_ARR); |
| | 714 | $arr = array(); |
| | 715 | } else { |
| | 716 | if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
| | 717 | $stk = array(SERVICES_JSON_IN_OBJ); |
| | 718 | $obj = array(); |
| | 719 | } else { |
| | 720 | $stk = array(SERVICES_JSON_IN_OBJ); |
| | 721 | $obj = new stdClass(); |
| | 722 | } |
| | 723 | } |
| | 724 | |
| | 725 | array_push($stk, array('what' => SERVICES_JSON_SLICE, |
| | 726 | 'where' => 0, |
| | 727 | 'delim' => false)); |
| | 728 | |
| | 729 | $chrs = $this->substr8($str, 1, -1); |
| | 730 | $chrs = $this->reduce_string($chrs); |
| | 731 | |
| | 732 | if ($chrs == '') { |
| | 733 | if (reset($stk) == SERVICES_JSON_IN_ARR) { |
| | 734 | return $arr; |
| | 735 | |
| | 736 | } else { |
| | 737 | return $obj; |
| | 738 | |
| | 739 | } |
| | 740 | } |
| | 741 | |
| | 742 | //print("\nparsing {$chrs}\n"); |
| | 743 | |
| | 744 | $strlen_chrs = $this->strlen8($chrs); |
| | 745 | |
| | 746 | for ($c = 0; $c <= $strlen_chrs; ++$c) { |
| | 747 | |
| | 748 | $top = end($stk); |
| | 749 | $substr_chrs_c_2 = $this->substr8($chrs, $c, 2); |
| | 750 | |
| | 751 | if (($c == $strlen_chrs) || (($chrs{$c} == ',') && ($top['what'] == SERVICES_JSON_SLICE))) { |
| | 752 | // found a comma that is not inside a string, array, etc., |
| | 753 | // OR we've reached the end of the character list |
| | 754 | $slice = $this->substr8($chrs, $top['where'], ($c - $top['where'])); |
| | 755 | array_push($stk, array('what' => SERVICES_JSON_SLICE, 'where' => ($c + 1), 'delim' => false)); |
| | 756 | //print("Found split at {$c}: ".$this->substr8($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| | 757 | |
| | 758 | if (reset($stk) == SERVICES_JSON_IN_ARR) { |
| | 759 | // we are in an array, so just push an element onto the stack |
| | 760 | array_push($arr, $this->decode($slice)); |
| | 761 | |
| | 762 | } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
| | 763 | // we are in an object, so figure |
| | 764 | // out the property name and set an |
| | 765 | // element in an associative array, |
| | 766 | // for now |
| | 767 | $parts = array(); |
| | 768 | |
| | 769 | if (preg_match('/^\s*(["\'].*[^\\\]["\'])\s*:/Uis', $slice, $parts)) { |
| | 770 | // "name":value pair |
| | 771 | $key = $this->decode($parts[1]); |
| | 772 | $val = $this->decode(trim(substr($slice, strlen($parts[0])), ", \t\n\r\0\x0B")); |
| | 773 | if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
| | 774 | $obj[$key] = $val; |
| | 775 | } else { |
| | 776 | $obj->$key = $val; |
| | 777 | } |
| | 778 | } elseif (preg_match('/^\s*(\w+)\s*:/Uis', $slice, $parts)) { |
| | 779 | // name:value pair, where name is unquoted |
| | 780 | $key = $parts[1]; |
| | 781 | $val = $this->decode(trim(substr($slice, strlen($parts[0])), ", \t\n\r\0\x0B")); |
| | 782 | |
| | 783 | if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
| | 784 | $obj[$key] = $val; |
| | 785 | } else { |
| | 786 | $obj->$key = $val; |
| | 787 | } |
| | 788 | } |
| | 789 | |
| | 790 | } |
| | 791 | |
| | 792 | } elseif ((($chrs{$c} == '"') || ($chrs{$c} == "'")) && ($top['what'] != SERVICES_JSON_IN_STR)) { |
| | 793 | // found a quote, and we are not inside a string |
| | 794 | array_push($stk, array('what' => SERVICES_JSON_IN_STR, 'where' => $c, 'delim' => $chrs{$c})); |
| | 795 | //print("Found start of string at {$c}\n"); |
| | 796 | |
| | 797 | } elseif (($chrs{$c} == $top['delim']) && |
| | 798 | ($top['what'] == SERVICES_JSON_IN_STR) && |
| | 799 | (($this->strlen8($this->substr8($chrs, 0, $c)) - $this->strlen8(rtrim($this->substr8($chrs, 0, $c), '\\'))) % 2 != 1)) { |
| | 800 | // found a quote, we're in a string, and it's not escaped |
| | 801 | // we know that it's not escaped becase there is _not_ an |
| | 802 | // odd number of backslashes at the end of the string so far |
| | 803 | array_pop($stk); |
| | 804 | //print("Found end of string at {$c}: ".$this->substr8($chrs, $top['where'], (1 + 1 + $c - $top['where']))."\n"); |
| | 805 | |
| | 806 | } elseif (($chrs{$c} == '[') && |
| | 807 | in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
| | 808 | // found a left-bracket, and we are in an array, object, or slice |
| | 809 | array_push($stk, array('what' => SERVICES_JSON_IN_ARR, 'where' => $c, 'delim' => false)); |
| | 810 | //print("Found start of array at {$c}\n"); |
| | 811 | |
| | 812 | } elseif (($chrs{$c} == ']') && ($top['what'] == SERVICES_JSON_IN_ARR)) { |
| | 813 | // found a right-bracket, and we're in an array |
| | 814 | array_pop($stk); |
| | 815 | //print("Found end of array at {$c}: ".$this->substr8($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| | 816 | |
| | 817 | } elseif (($chrs{$c} == '{') && |
| | 818 | in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
| | 819 | // found a left-brace, and we are in an array, object, or slice |
| | 820 | array_push($stk, array('what' => SERVICES_JSON_IN_OBJ, 'where' => $c, 'delim' => false)); |
| | 821 | //print("Found start of object at {$c}\n"); |
| | 822 | |
| | 823 | } elseif (($chrs{$c} == '}') && ($top['what'] == SERVICES_JSON_IN_OBJ)) { |
| | 824 | // found a right-brace, and we're in an object |
| | 825 | array_pop($stk); |
| | 826 | //print("Found end of object at {$c}: ".$this->substr8($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| | 827 | |
| | 828 | } elseif (($substr_chrs_c_2 == '/*') && |
| | 829 | in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
| | 830 | // found a comment start, and we are in an array, object, or slice |
| | 831 | array_push($stk, array('what' => SERVICES_JSON_IN_CMT, 'where' => $c, 'delim' => false)); |
| | 832 | $c++; |
| | 833 | //print("Found start of comment at {$c}\n"); |
| | 834 | |
| | 835 | } elseif (($substr_chrs_c_2 == '*/') && ($top['what'] == SERVICES_JSON_IN_CMT)) { |
| | 836 | // found a comment end, and we're in one now |
| | 837 | array_pop($stk); |
| | 838 | $c++; |
| | 839 | |
| | 840 | for ($i = $top['where']; $i <= $c; ++$i) |
| | 841 | $chrs = substr_replace($chrs, ' ', $i, 1); |
| | 842 | |
| | 843 | //print("Found end of comment at {$c}: ".$this->substr8($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
| | 844 | |
| | 845 | } |
| | 846 | |
| | 847 | } |
| | 848 | |
| | 849 | if (reset($stk) == SERVICES_JSON_IN_ARR) { |
| | 850 | return $arr; |
| | 851 | |
| | 852 | } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
| | 853 | return $obj; |
| | 854 | |
| | 855 | } |
| | 856 | |
| | 857 | } |
| | 858 | } |
| | 859 | } |
| | 860 | |
| | 861 | /** |
| | 862 | * @todo Ultimately, this should just call PEAR::isError() |
| | 863 | */ |
| | 864 | function isError($data, $code = null) |
| | 865 | { |
| | 866 | if (class_exists('pear')) { |
| | 867 | return PEAR::isError($data, $code); |
| | 868 | } elseif (is_object($data) && (get_class($data) == 'services_json_error' || |
| | 869 | is_subclass_of($data, 'services_json_error'))) { |
| | 870 | return true; |
| | 871 | } |
| | 872 | |
| | 873 | return false; |
| | 874 | } |
| | 875 | |
| | 876 | /** |
| | 877 | * Calculates length of string in bytes |
| | 878 | * @param string |
| | 879 | * @return integer length |
| | 880 | */ |
| | 881 | function strlen8( $str ) |
| | 882 | { |
| | 883 | if ( $this->_mb_strlen ) { |
| | 884 | return mb_strlen( $str, "8bit" ); |
| | 885 | } |
| | 886 | return strlen( $str ); |
| | 887 | } |
| | 888 | |
| | 889 | /** |
| | 890 | * Returns part of a string, interpreting $start and $length as number of bytes. |
| | 891 | * @param string |
| | 892 | * @param integer start |
| | 893 | * @param integer length |
| | 894 | * @return integer length |
| | 895 | */ |
| | 896 | function substr8( $string, $start, $length=false ) |
| | 897 | { |
| | 898 | if ( $length === false ) { |
| | 899 | $length = $this->strlen8( $string ) - $start; |
| | 900 | } |
| | 901 | if ( $this->_mb_substr ) { |
| | 902 | return mb_substr( $string, $start, $length, "8bit" ); |
| | 903 | } |
| | 904 | return substr( $string, $start, $length ); |
| | 905 | } |
| | 906 | |