includes/api/ApiFormatJson_json.php

   1 <?php
   2 /* vim: set expandtab tabstop=4 shiftwidth=4 softtabstop=4: */
   3
   4 /**
   5 * Converts to and from JSON format.
   6 *
   7 * JSON (JavaScript Object Notation) is a lightweight data-interchange
   8 * format. It is easy for humans to read and write. It is easy for machines
   9 * to parse and generate. It is based on a subset of the JavaScript
  10 * Programming Language, Standard ECMA-262 3rd Edition - December 1999.
  11 * This feature can also be found in  Python. JSON is a text format that is
  12 * completely language independent but uses conventions that are familiar
  13 * to programmers of the C-family of languages, including C, C++, C#, Java,
  14 * JavaScript, Perl, TCL, and many others. These properties make JSON an
  15 * ideal data-interchange language.
  16 *
  17 * This package provides a simple encoder and decoder for JSON notation. It
  18 * is intended for use with client-side Javascript applications that make
  19 * use of HTTPRequest to perform server communication functions - data can
  20 * be encoded into JSON notation for use in a client-side javascript, or
  21 * decoded from incoming Javascript requests. JSON format is native to
  22 * Javascript, and can be directly eval()'ed with no further parsing
  23 * overhead
  24 *
  25 * All strings should be in ASCII or UTF-8 format!
  26 *
  27 * LICENSE: Redistribution and use in source and binary forms, with or
  28 * without modification, are permitted provided that the following
  29 * conditions are met: Redistributions of source code must retain the
  30 * above copyright notice, this list of conditions and the following
  31 * disclaimer. Redistributions in binary form must reproduce the above
  32 * copyright notice, this list of conditions and the following disclaimer
  33 * in the documentation and/or other materials provided with the
  34 * distribution.
  35 *
  36 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED
  37 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF
  38 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN
  39 * NO EVENT SHALL CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT,
  40 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING,
  41 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS
  42 * OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND
  43 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR
  44 * TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE
  45 * USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH
  46 * DAMAGE.
  47 *
  48 * @ingroup     API
  49 * @author      Michal Migurski <mike-json@teczno.com>
  50 * @author      Matt Knapp <mdknapp[at]gmail[dot]com>
  51 * @author      Brett Stimmerman <brettstimmerman[at]gmail[dot]com>
  52 * @copyright   2005 Michal Migurski
  53 * @version     CVS: $Id$
  54 * @license     http://www.opensource.org/licenses/bsd-license.php
  55 * @see         http://pear.php.net/pepr/pepr-proposal-show.php?id=198
  56 */
  57
  58 /**
  59 * Marker constant for Services_JSON::decode(), used to flag stack state
  60 */
  61 define('SERVICES_JSON_SLICE',   1);
  62
  63 /**
  64 * Marker constant for Services_JSON::decode(), used to flag stack state
  65 */
  66 define('SERVICES_JSON_IN_STR',  2);
  67
  68 /**
  69 * Marker constant for Services_JSON::decode(), used to flag stack state
  70 */
  71 define('SERVICES_JSON_IN_ARR',  3);
  72
  73 /**
  74 * Marker constant for Services_JSON::decode(), used to flag stack state
  75 */
  76 define('SERVICES_JSON_IN_OBJ',  4);
  77
  78 /**
  79 * Marker constant for Services_JSON::decode(), used to flag stack state
  80 */
  81 define('SERVICES_JSON_IN_CMT', 5);
  82
  83 /**
  84 * Behavior switch for Services_JSON::decode()
  85 */
  86 define('SERVICES_JSON_LOOSE_TYPE', 16);
  87
  88 /**
  89 * Behavior switch for Services_JSON::decode()
  90 */
  91 define('SERVICES_JSON_SUPPRESS_ERRORS', 32);
  92
  93 /**
  94  * Converts to and from JSON format.
  95  *
  96  * Brief example of use:
  97  *
  98  * <code>
  99  * // create a new instance of Services_JSON
 100  * $json = new Services_JSON();
 101  *
 102  * // convert a complexe value to JSON notation, and send it to the browser
 103  * $value = array('foo', 'bar', array(1, 2, 'baz'), array(3, array(4)));
 104  * $output = $json->encode($value);
 105  *
 106  * print($output);
 107  * // prints: ["foo","bar",[1,2,"baz"],[3,[4]]]
 108  *
 109  * // accept incoming POST data, assumed to be in JSON notation
 110  * $input = file_get_contents('php://input', 1000000);
 111  * $value = $json->decode($input);
 112  * </code>
 113  *
 114  * @ingroup API
 115  */
 116 class Services_JSON
 117 {
 118    /**
 119     * constructs a new JSON instance
 120     *
 121     * @param    int     $use    object behavior flags; combine with boolean-OR
 122     *
 123     *                           possible values:
 124     *                           - SERVICES_JSON_LOOSE_TYPE:  loose typing.
 125     *                                   "{...}" syntax creates associative arrays
 126     *                                   instead of objects in decode().
 127     *                           - SERVICES_JSON_SUPPRESS_ERRORS:  error suppression.
 128     *                                   Values which can't be encoded (e.g. resources)
 129     *                                   appear as NULL instead of throwing errors.
 130     *                                   By default, a deeply-nested resource will
 131     *                                   bubble up with an error, so all return values
 132     *                                   from encode() should be checked with isError()
 133     */
 134     function Services_JSON($use = 0)
 135     {
 136         $this->use = $use;
 137     }
 138
 139    /**
 140     * convert a string from one UTF-16 char to one UTF-8 char
 141     *
 142     * Normally should be handled by mb_convert_encoding, but
 143     * provides a slower PHP-only method for installations
 144     * that lack the multibye string extension.
 145     *
 146     * @param    string  $utf16  UTF-16 character
 147     * @return   string  UTF-8 character
 148     * @access   private
 149     */
 150     function utf162utf8($utf16)
 151     {
 152         // oh please oh please oh please oh please oh please
 153         if(function_exists('mb_convert_encoding')) {
 154             return mb_convert_encoding($utf16, 'UTF-8', 'UTF-16');
 155         }
 156
 157         $bytes = (ord($utf16{0}) << 8) | ord($utf16{1});
 158
 159         switch(true) {
 160             case ((0x7F & $bytes) == $bytes):
 161                 // this case should never be reached, because we are in ASCII range
 162                 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 163                 return chr(0x7F & $bytes);
 164
 165             case (0x07FF & $bytes) == $bytes:
 166                 // return a 2-byte UTF-8 character
 167                 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 168                 return chr(0xC0 | (($bytes >> 6) & 0x1F))
 169                      . chr(0x80 | ($bytes & 0x3F));
 170
 171             case (0xFFFF & $bytes) == $bytes:
 172                 // return a 3-byte UTF-8 character
 173                 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 174                 return chr(0xE0 | (($bytes >> 12) & 0x0F))
 175                      . chr(0x80 | (($bytes >> 6) & 0x3F))
 176                      . chr(0x80 | ($bytes & 0x3F));
 177         }
 178
 179         // ignoring UTF-32 for now, sorry
 180         return '';
 181     }
 182
 183    /**
 184     * convert a string from one UTF-8 char to one UTF-16 char
 185     *
 186     * Normally should be handled by mb_convert_encoding, but
 187     * provides a slower PHP-only method for installations
 188     * that lack the multibye string extension.
 189     *
 190     * @param    string  $utf8   UTF-8 character
 191     * @return   string  UTF-16 character
 192     * @access   private
 193     */
 194     function utf82utf16($utf8)
 195     {
 196         // oh please oh please oh please oh please oh please
 197         if(function_exists('mb_convert_encoding')) {
 198             return mb_convert_encoding($utf8, 'UTF-16', 'UTF-8');
 199         }
 200
 201         switch(strlen($utf8)) {
 202             case 1:
 203                 // this case should never be reached, because we are in ASCII range
 204                 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 205                 return $utf8;
 206
 207             case 2:
 208                 // return a UTF-16 character from a 2-byte UTF-8 char
 209                 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 210                 return chr(0x07 & (ord($utf8{0}) >> 2))
 211                      . chr((0xC0 & (ord($utf8{0}) << 6))
 212                          | (0x3F & ord($utf8{1})));
 213
 214             case 3:
 215                 // return a UTF-16 character from a 3-byte UTF-8 char
 216                 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 217                 return chr((0xF0 & (ord($utf8{0}) << 4))
 218                          | (0x0F & (ord($utf8{1}) >> 2)))
 219                      . chr((0xC0 & (ord($utf8{1}) << 6))
 220                          | (0x7F & ord($utf8{2})));
 221         }
 222
 223         // ignoring UTF-32 for now, sorry
 224         return '';
 225     }
 226
 227    /**
 228     * encodes an arbitrary variable into JSON format
 229     *
 230     * @param    mixed   $var    any number, boolean, string, array, or object to be encoded.
 231     *                           see argument 1 to Services_JSON() above for array-parsing behavior.
 232     *                           if var is a strng, note that encode() always expects it
 233     *                           to be in ASCII or UTF-8 format!
 234     * @param    bool    $pretty    pretty-print output with indents and newlines
 235     *
 236     * @return   mixed   JSON string representation of input var or an error if a problem occurs
 237     * @access   public
 238     */
 239     function encode($var, $pretty=false)
 240     {
 241         $this->indent = 0;
 242         $this->pretty = $pretty;
 243         $this->nameValSeparator = $pretty ? ': ' : ':';
 244         return $this->encode2($var);
 245     }
 246
 247    /**
 248     * encodes an arbitrary variable into JSON format
 249     *
 250     * @param    mixed   $var    any number, boolean, string, array, or object to be encoded.
 251     *                           see argument 1 to Services_JSON() above for array-parsing behavior.
 252     *                           if var is a strng, note that encode() always expects it
 253     *                           to be in ASCII or UTF-8 format!
 254     *
 255     * @return   mixed   JSON string representation of input var or an error if a problem occurs
 256     * @access   private
 257     */
 258     function encode2($var)
 259     {
 260         if ($this->pretty) {
 261             $close = "\n" . str_repeat("\t", $this->indent);
 262             $open = $close . "\t";
 263             $mid = ',' . $open;
 264         }
 265         else {
 266             $open = $close = '';
 267             $mid = ',';
 268         }
 269
 270         switch (gettype($var)) {
 271             case 'boolean':
 272                 return $var ? 'true' : 'false';
 273
 274             case 'NULL':
 275                 return 'null';
 276
 277             case 'integer':
 278                 return (int) $var;
 279
 280             case 'double':
 281             case 'float':
 282                 return (float) $var;
 283
 284             case 'string':
 285                 // STRINGS ARE EXPECTED TO BE IN ASCII OR UTF-8 FORMAT
 286                 $ascii = '';
 287                 $strlen_var = strlen($var);
 288
 289                /*
 290                 * Iterate over every character in the string,
 291                 * escaping with a slash or encoding to UTF-8 where necessary
 292                 */
 293                 for ($c = 0; $c < $strlen_var; ++$c) {
 294
 295                     $ord_var_c = ord($var{$c});
 296
 297                     switch (true) {
 298                         case $ord_var_c == 0x08:
 299                             $ascii .= '\b';
 300                             break;
 301                         case $ord_var_c == 0x09:
 302                             $ascii .= '\t';
 303                             break;
 304                         case $ord_var_c == 0x0A:
 305                             $ascii .= '\n';
 306                             break;
 307                         case $ord_var_c == 0x0C:
 308                             $ascii .= '\f';
 309                             break;
 310                         case $ord_var_c == 0x0D:
 311                             $ascii .= '\r';
 312                             break;
 313
 314                         case $ord_var_c == 0x22:
 315                         case $ord_var_c == 0x2F:
 316                         case $ord_var_c == 0x5C:
 317                             // double quote, slash, slosh
 318                             $ascii .= '\\'.$var{$c};
 319                             break;
 320
 321                         case (($ord_var_c >= 0x20) && ($ord_var_c <= 0x7F)):
 322                             // characters U-00000000 - U-0000007F (same as ASCII)
 323                             $ascii .= $var{$c};
 324                             break;
 325
 326                         case (($ord_var_c & 0xE0) == 0xC0):
 327                             // characters U-00000080 - U-000007FF, mask 110XXXXX
 328                             // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 329                             $char = pack('C*', $ord_var_c, ord($var{$c + 1}));
 330                             $c += 1;
 331                             $utf16 = $this->utf82utf16($char);
 332                             $ascii .= sprintf('\u%04s', bin2hex($utf16));
 333                             break;
 334
 335                         case (($ord_var_c & 0xF0) == 0xE0):
 336                             // characters U-00000800 - U-0000FFFF, mask 1110XXXX
 337                             // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 338                             $char = pack('C*', $ord_var_c,
 339                                          ord($var{$c + 1}),
 340                                          ord($var{$c + 2}));
 341                             $c += 2;
 342                             $utf16 = $this->utf82utf16($char);
 343                             $ascii .= sprintf('\u%04s', bin2hex($utf16));
 344                             break;
 345
 346                         case (($ord_var_c & 0xF8) == 0xF0):
 347                             // characters U-00010000 - U-001FFFFF, mask 11110XXX
 348                             // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 349                             $char = pack('C*', $ord_var_c,
 350                                          ord($var{$c + 1}),
 351                                          ord($var{$c + 2}),
 352                                          ord($var{$c + 3}));
 353                             $c += 3;
 354                             $utf16 = $this->utf82utf16($char);
 355                             $ascii .= sprintf('\u%04s', bin2hex($utf16));
 356                             break;
 357
 358                         case (($ord_var_c & 0xFC) == 0xF8):
 359                             // characters U-00200000 - U-03FFFFFF, mask 111110XX
 360                             // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 361                             $char = pack('C*', $ord_var_c,
 362                                          ord($var{$c + 1}),
 363                                          ord($var{$c + 2}),
 364                                          ord($var{$c + 3}),
 365                                          ord($var{$c + 4}));
 366                             $c += 4;
 367                             $utf16 = $this->utf82utf16($char);
 368                             $ascii .= sprintf('\u%04s', bin2hex($utf16));
 369                             break;
 370
 371                         case (($ord_var_c & 0xFE) == 0xFC):
 372                             // characters U-04000000 - U-7FFFFFFF, mask 1111110X
 373                             // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 374                             $char = pack('C*', $ord_var_c,
 375                                          ord($var{$c + 1}),
 376                                          ord($var{$c + 2}),
 377                                          ord($var{$c + 3}),
 378                                          ord($var{$c + 4}),
 379                                          ord($var{$c + 5}));
 380                             $c += 5;
 381                             $utf16 = $this->utf82utf16($char);
 382                             $ascii .= sprintf('\u%04s', bin2hex($utf16));
 383                             break;
 384                     }
 385                 }
 386
 387                 return '"'.$ascii.'"';
 388
 389             case 'array':
 390                /*
 391                 * As per JSON spec if any array key is not an integer
 392                 * we must treat the the whole array as an object. We
 393                 * also try to catch a sparsely populated associative
 394                 * array with numeric keys here because some JS engines
 395                 * will create an array with empty indexes up to
 396                 * max_index which can cause memory issues and because
 397                 * the keys, which may be relevant, will be remapped
 398                 * otherwise.
 399                 *
 400                 * As per the ECMA and JSON specification an object may
 401                 * have any string as a property. Unfortunately due to
 402                 * a hole in the ECMA specification if the key is a
 403                 * ECMA reserved word or starts with a digit the
 404                 * parameter is only accessible using ECMAScript's
 405                 * bracket notation.
 406                 */
 407
 408                 // treat as a JSON object
 409                 if (is_array($var) && count($var) && (array_keys($var) !== range(0, sizeof($var) - 1))) {
 410                     $this->indent++;
 411                     $properties = array_map(array($this, 'name_value'),
 412                                             array_keys($var),
 413                                             array_values($var));
 414                     $this->indent--;
 415
 416                     foreach($properties as $property) {
 417                         if(Services_JSON::isError($property)) {
 418                             return $property;
 419                         }
 420                     }
 421
 422                     return '{' . $open . join($mid, $properties) . $close . '}';
 423                 }
 424
 425                 // treat it like a regular array
 426                 $this->indent++;
 427                 $elements = array_map(array($this, 'encode2'), $var);
 428                 $this->indent--;
 429
 430                 foreach($elements as $element) {
 431                     if(Services_JSON::isError($element)) {
 432                         return $element;
 433                     }
 434                 }
 435
 436                 return '[' . $open . join($mid, $elements) . $close . ']';
 437
 438             case 'object':
 439                 $vars = get_object_vars($var);
 440
 441                 $this->indent++;
 442                 $properties = array_map(array($this, 'name_value'),
 443                                         array_keys($vars),
 444                                         array_values($vars));
 445                 $this->indent--;
 446
 447                 foreach($properties as $property) {
 448                     if(Services_JSON::isError($property)) {
 449                         return $property;
 450                     }
 451                 }
 452
 453                 return '{' . $open . join($mid, $properties) . $close . '}';
 454
 455             default:
 456                 return ($this->use & SERVICES_JSON_SUPPRESS_ERRORS)
 457                     ? 'null'
 458                     : new Services_JSON_Error(gettype($var)." can not be encoded as JSON string");
 459         }
 460     }
 461
 462    /**
 463     * array-walking function for use in generating JSON-formatted name-value pairs
 464     *
 465     * @param    string  $name   name of key to use
 466     * @param    mixed   $value  reference to an array element to be encoded
 467     *
 468     * @return   string  JSON-formatted name-value pair, like '"name":value'
 469     * @access   private
 470     */
 471     function name_value($name, $value)
 472     {
 473         $encoded_value = $this->encode2($value);
 474
 475         if(Services_JSON::isError($encoded_value)) {
 476             return $encoded_value;
 477         }
 478
 479         return $this->encode2(strval($name)) . $this->nameValSeparator . $encoded_value;
 480     }
 481
 482    /**
 483     * reduce a string by removing leading and trailing comments and whitespace
 484     *
 485     * @param    $str    string      string value to strip of comments and whitespace
 486     *
 487     * @return   string  string value stripped of comments and whitespace
 488     * @access   private
 489     */
 490     function reduce_string($str)
 491     {
 492         $str = preg_replace(array(
 493
 494                 // eliminate single line comments in '// ...' form
 495                 '#^\s*//(.+)$#m',
 496
 497                 // eliminate multi-line comments in '/* ... */' form, at start of string
 498                 '#^\s*/\*(.+)\*/#Us',
 499
 500                 // eliminate multi-line comments in '/* ... */' form, at end of string
 501                 '#/\*(.+)\*/\s*$#Us'
 502
 503             ), '', $str);
 504
 505         // eliminate extraneous space
 506         return trim($str);
 507     }
 508
 509    /**
 510     * decodes a JSON string into appropriate variable
 511     *
 512     * @param    string  $str    JSON-formatted string
 513     *
 514     * @return   mixed   number, boolean, string, array, or object
 515     *                   corresponding to given JSON input string.
 516     *                   See argument 1 to Services_JSON() above for object-output behavior.
 517     *                   Note that decode() always returns strings
 518     *                   in ASCII or UTF-8 format!
 519     * @access   public
 520     */
 521     function decode($str)
 522     {
 523         $str = $this->reduce_string($str);
 524
 525         switch (strtolower($str)) {
 526             case 'true':
 527                 return true;
 528
 529             case 'false':
 530                 return false;
 531
 532             case 'null':
 533                 return null;
 534
 535             default:
 536                 $m = array();
 537
 538                 if (is_numeric($str)) {
 539                     // Lookie-loo, it's a number
 540
 541                     // This would work on its own, but I'm trying to be
 542                     // good about returning integers where appropriate:
 543                     // return (float)$str;
 544
 545                     // Return float or int, as appropriate
 546                     return ((float)$str == (integer)$str)
 547                         ? (integer)$str
 548                         : (float)$str;
 549
 550                 } elseif (preg_match('/^("|\').*(\1)$/s', $str, $m) && $m[1] == $m[2]) {
 551                     // STRINGS RETURNED IN UTF-8 FORMAT
 552                     $delim = substr($str, 0, 1);
 553                     $chrs = substr($str, 1, -1);
 554                     $utf8 = '';
 555                     $strlen_chrs = strlen($chrs);
 556
 557                     for ($c = 0; $c < $strlen_chrs; ++$c) {
 558
 559                         $substr_chrs_c_2 = substr($chrs, $c, 2);
 560                         $ord_chrs_c = ord($chrs{$c});
 561
 562                         switch (true) {
 563                             case $substr_chrs_c_2 == '\b':
 564                                 $utf8 .= chr(0x08);
 565                                 ++$c;
 566                                 break;
 567                             case $substr_chrs_c_2 == '\t':
 568                                 $utf8 .= chr(0x09);
 569                                 ++$c;
 570                                 break;
 571                             case $substr_chrs_c_2 == '\n':
 572                                 $utf8 .= chr(0x0A);
 573                                 ++$c;
 574                                 break;
 575                             case $substr_chrs_c_2 == '\f':
 576                                 $utf8 .= chr(0x0C);
 577                                 ++$c;
 578                                 break;
 579                             case $substr_chrs_c_2 == '\r':
 580                                 $utf8 .= chr(0x0D);
 581                                 ++$c;
 582                                 break;
 583
 584                             case $substr_chrs_c_2 == '\\"':
 585                             case $substr_chrs_c_2 == '\\\'':
 586                             case $substr_chrs_c_2 == '\\\\':
 587                             case $substr_chrs_c_2 == '\\/':
 588                                 if (($delim == '"' && $substr_chrs_c_2 != '\\\'') ||
 589                                    ($delim == "'" && $substr_chrs_c_2 != '\\"')) {
 590                                     $utf8 .= $chrs{++$c};
 591                                 }
 592                                 break;
 593
 594                             case preg_match('/\\\u[0-9A-F]{4}/i', substr($chrs, $c, 6)):
 595                                 // single, escaped unicode character
 596                                 $utf16 = chr(hexdec(substr($chrs, ($c + 2), 2)))
 597                                        . chr(hexdec(substr($chrs, ($c + 4), 2)));
 598                                 $utf8 .= $this->utf162utf8($utf16);
 599                                 $c += 5;
 600                                 break;
 601
 602                             case ($ord_chrs_c >= 0x20) && ($ord_chrs_c <= 0x7F):
 603                                 $utf8 .= $chrs{$c};
 604                                 break;
 605
 606                             case ($ord_chrs_c & 0xE0) == 0xC0:
 607                                 // characters U-00000080 - U-000007FF, mask 110XXXXX
 608                                 //see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 609                                 $utf8 .= substr($chrs, $c, 2);
 610                                 ++$c;
 611                                 break;
 612
 613                             case ($ord_chrs_c & 0xF0) == 0xE0:
 614                                 // characters U-00000800 - U-0000FFFF, mask 1110XXXX
 615                                 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 616                                 $utf8 .= substr($chrs, $c, 3);
 617                                 $c += 2;
 618                                 break;
 619
 620                             case ($ord_chrs_c & 0xF8) == 0xF0:
 621                                 // characters U-00010000 - U-001FFFFF, mask 11110XXX
 622                                 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 623                                 $utf8 .= substr($chrs, $c, 4);
 624                                 $c += 3;
 625                                 break;
 626
 627                             case ($ord_chrs_c & 0xFC) == 0xF8:
 628                                 // characters U-00200000 - U-03FFFFFF, mask 111110XX
 629                                 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 630                                 $utf8 .= substr($chrs, $c, 5);
 631                                 $c += 4;
 632                                 break;
 633
 634                             case ($ord_chrs_c & 0xFE) == 0xFC:
 635                                 // characters U-04000000 - U-7FFFFFFF, mask 1111110X
 636                                 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8
 637                                 $utf8 .= substr($chrs, $c, 6);
 638                                 $c += 5;
 639                                 break;
 640
 641                         }
 642
 643                     }
 644
 645                     return $utf8;
 646
 647                 } elseif (preg_match('/^\[.*\]$/s', $str) || preg_match('/^\{.*\}$/s', $str)) {
 648                     // array, or object notation
 649
 650                     if ($str{0} == '[') {
 651                         $stk = array(SERVICES_JSON_IN_ARR);
 652                         $arr = array();
 653                     } else {
 654                         if ($this->use & SERVICES_JSON_LOOSE_TYPE) {
 655                             $stk = array(SERVICES_JSON_IN_OBJ);
 656                             $obj = array();
 657                         } else {
 658                             $stk = array(SERVICES_JSON_IN_OBJ);
 659                             $obj = new stdClass();
 660                         }
 661                     }
 662
 663                     array_push($stk, array('what'  => SERVICES_JSON_SLICE,
 664                                            'where' => 0,
 665                                            'delim' => false));
 666
 667                     $chrs = substr($str, 1, -1);
 668                     $chrs = $this->reduce_string($chrs);
 669
 670                     if ($chrs == '') {
 671                         if (reset($stk) == SERVICES_JSON_IN_ARR) {
 672                             return $arr;
 673
 674                         } else {
 675                             return $obj;
 676
 677                         }
 678                     }
 679
 680                     //print("\nparsing {$chrs}\n");
 681
 682                     $strlen_chrs = strlen($chrs);
 683
 684                     for ($c = 0; $c <= $strlen_chrs; ++$c) {
 685
 686                         $top = end($stk);
 687                         $substr_chrs_c_2 = substr($chrs, $c, 2);
 688
 689                         if (($c == $strlen_chrs) || (($chrs{$c} == ',') && ($top['what'] == SERVICES_JSON_SLICE))) {
 690                             // found a comma that is not inside a string, array, etc.,
 691                             // OR we've reached the end of the character list
 692                             $slice = substr($chrs, $top['where'], ($c - $top['where']));
 693                             array_push($stk, array('what' => SERVICES_JSON_SLICE, 'where' => ($c + 1), 'delim' => false));
 694                             //print("Found split at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n");
 695
 696                             if (reset($stk) == SERVICES_JSON_IN_ARR) {
 697                                 // we are in an array, so just push an element onto the stack
 698                                 array_push($arr, $this->decode($slice));
 699
 700                             } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) {
 701                                 // we are in an object, so figure
 702                                 // out the property name and set an
 703                                 // element in an associative array,
 704                                 // for now
 705                                 $parts = array();
 706
 707                                 if (preg_match('/^\s*(["\'].*[^\\\]["\'])\s*:\s*(\S.*),?$/Uis', $slice, $parts)) {
 708                                     // "name":value pair
 709                                     $key = $this->decode($parts[1]);
 710                                     $val = $this->decode($parts[2]);
 711
 712                                     if ($this->use & SERVICES_JSON_LOOSE_TYPE) {
 713                                         $obj[$key] = $val;
 714                                     } else {
 715                                         $obj->$key = $val;
 716                                     }
 717                                 } elseif (preg_match('/^\s*(\w+)\s*:\s*(\S.*),?$/Uis', $slice, $parts)) {
 718                                     // name:value pair, where name is unquoted
 719                                     $key = $parts[1];
 720                                     $val = $this->decode($parts[2]);
 721
 722                                     if ($this->use & SERVICES_JSON_LOOSE_TYPE) {
 723                                         $obj[$key] = $val;
 724                                     } else {
 725                                         $obj->$key = $val;
 726                                     }
 727                                 }
 728
 729                             }
 730
 731                         } elseif ((($chrs{$c} == '"') || ($chrs{$c} == "'")) && ($top['what'] != SERVICES_JSON_IN_STR)) {
 732                             // found a quote, and we are not inside a string
 733                             array_push($stk, array('what' => SERVICES_JSON_IN_STR, 'where' => $c, 'delim' => $chrs{$c}));
 734                             //print("Found start of string at {$c}\n");
 735
 736                         } elseif (($chrs{$c} == $top['delim']) &&
 737                                  ($top['what'] == SERVICES_JSON_IN_STR) &&
 738                                  (($chrs{$c - 1} != '\\') ||
 739                                  ($chrs{$c - 1} == '\\' && $chrs{$c - 2} == '\\'))) {
 740                             // found a quote, we're in a string, and it's not escaped
 741                             array_pop($stk);
 742                             //print("Found end of string at {$c}: ".substr($chrs, $top['where'], (1 + 1 + $c - $top['where']))."\n");
 743
 744                         } elseif (($chrs{$c} == '[') &&
 745                                  in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) {
 746                             // found a left-bracket, and we are in an array, object, or slice
 747                             array_push($stk, array('what' => SERVICES_JSON_IN_ARR, 'where' => $c, 'delim' => false));
 748                             //print("Found start of array at {$c}\n");
 749
 750                         } elseif (($chrs{$c} == ']') && ($top['what'] == SERVICES_JSON_IN_ARR)) {
 751                             // found a right-bracket, and we're in an array
 752                             array_pop($stk);
 753                             //print("Found end of array at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n");
 754
 755                         } elseif (($chrs{$c} == '{') &&
 756                                  in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) {
 757                             // found a left-brace, and we are in an array, object, or slice
 758                             array_push($stk, array('what' => SERVICES_JSON_IN_OBJ, 'where' => $c, 'delim' => false));
 759                             //print("Found start of object at {$c}\n");
 760
 761                         } elseif (($chrs{$c} == '}') && ($top['what'] == SERVICES_JSON_IN_OBJ)) {
 762                             // found a right-brace, and we're in an object
 763                             array_pop($stk);
 764                             //print("Found end of object at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n");
 765
 766                         } elseif (($substr_chrs_c_2 == '/*') &&
 767                                  in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) {
 768                             // found a comment start, and we are in an array, object, or slice
 769                             array_push($stk, array('what' => SERVICES_JSON_IN_CMT, 'where' => $c, 'delim' => false));
 770                             $c++;
 771                             //print("Found start of comment at {$c}\n");
 772
 773                         } elseif (($substr_chrs_c_2 == '*/') && ($top['what'] == SERVICES_JSON_IN_CMT)) {
 774                             // found a comment end, and we're in one now
 775                             array_pop($stk);
 776                             $c++;
 777
 778                             for ($i = $top['where']; $i <= $c; ++$i)
 779                                 $chrs = substr_replace($chrs, ' ', $i, 1);
 780
 781                             //print("Found end of comment at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n");
 782
 783                         }
 784
 785                     }
 786
 787                     if (reset($stk) == SERVICES_JSON_IN_ARR) {
 788                         return $arr;
 789
 790                     } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) {
 791                         return $obj;
 792
 793                     }
 794
 795                 }
 796         }
 797     }
 798
 799     /**
 800      * @todo Ultimately, this should just call PEAR::isError()
 801      */
 802     function isError($data, $code = null)
 803     {
 804         if (class_exists('pear')) {
 805             return PEAR::isError($data, $code);
 806         } elseif (is_object($data) && (get_class($data) == 'services_json_error' ||
 807                                  is_subclass_of($data, 'services_json_error'))) {
 808             return true;
 809         }
 810
 811         return false;
 812     }
 813 }
 814
 815
 816 // Hide the PEAR_Error variant from Doxygen
 817 /// @cond
 818 if (class_exists('PEAR_Error')) {
 819
 820     /**
 821      * @ingroup API
 822      */
 823     class Services_JSON_Error extends PEAR_Error
 824     {
 825         function Services_JSON_Error($message = 'unknown error', $code = null,
 826                                      $mode = null, $options = null, $userinfo = null)
 827         {
 828             parent::PEAR_Error($message, $code, $mode, $options, $userinfo);
 829         }
 830     }
 831
 832 } else {
 833 /// @endcond
 834
 835     /**
 836      * @todo Ultimately, this class shall be descended from PEAR_Error
 837      * @ingroup API
 838      */
 839     class Services_JSON_Error
 840     {
 841         function Services_JSON_Error($message = 'unknown error', $code = null,
 842                                      $mode = null, $options = null, $userinfo = null)
 843         {
 844
 845         }
 846     }
 847 }