|
1 <?php |
|
2 /* vim: set expandtab tabstop=4 shiftwidth=4 softtabstop=4: */ |
|
3 /** |
|
4 * Converts to and from JSON format. |
|
5 * |
|
6 * JSON (JavaScript Object Notation) is a lightweight data-interchange |
|
7 * format. It is easy for humans to read and write. It is easy for machines |
|
8 * to parse and generate. It is based on a subset of the JavaScript |
|
9 * Programming Language, Standard ECMA-262 3rd Edition - December 1999. |
|
10 * This feature can also be found in Python. JSON is a text format that is |
|
11 * completely language independent but uses conventions that are familiar |
|
12 * to programmers of the C-family of languages, including C, C++, C#, Java, |
|
13 * JavaScript, Perl, TCL, and many others. These properties make JSON an |
|
14 * ideal data-interchange language. |
|
15 * |
|
16 * This package provides a simple encoder and decoder for JSON notation. It |
|
17 * is intended for use with client-side Javascript applications that make |
|
18 * use of HTTPRequest to perform server communication functions - data can |
|
19 * be encoded into JSON notation for use in a client-side javascript, or |
|
20 * decoded from incoming Javascript requests. JSON format is native to |
|
21 * Javascript, and can be directly eval()'ed with no further parsing |
|
22 * overhead |
|
23 * |
|
24 * All strings should be in ASCII or UTF-8 format! |
|
25 * |
|
26 * LICENSE: Redistribution and use in source and binary forms, with or |
|
27 * without modification, are permitted provided that the following |
|
28 * conditions are met: Redistributions of source code must retain the |
|
29 * above copyright notice, this list of conditions and the following |
|
30 * disclaimer. Redistributions in binary form must reproduce the above |
|
31 * copyright notice, this list of conditions and the following disclaimer |
|
32 * in the documentation and/or other materials provided with the |
|
33 * distribution. |
|
34 * |
|
35 * THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESS OR IMPLIED |
|
36 * WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF |
|
37 * MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN |
|
38 * NO EVENT SHALL CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, |
|
39 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, |
|
40 * BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS |
|
41 * OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND |
|
42 * ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR |
|
43 * TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE |
|
44 * USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH |
|
45 * DAMAGE. |
|
46 * |
|
47 * @category |
|
48 * @package Services_JSON |
|
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: JSON.php 288200 2009-09-09 15:41:29Z alan_k $ |
|
54 * @license http://www.opensource.org/licenses/bsd-license.php |
|
55 * @link 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 class Services_JSON |
|
115 { |
|
116 /** |
|
117 * constructs a new JSON instance |
|
118 * |
|
119 * @param int $use object behavior flags; combine with boolean-OR |
|
120 * |
|
121 * possible values: |
|
122 * - SERVICES_JSON_LOOSE_TYPE: loose typing. |
|
123 * "{...}" syntax creates associative arrays |
|
124 * instead of objects in decode(). |
|
125 * - SERVICES_JSON_SUPPRESS_ERRORS: error suppression. |
|
126 * Values which can't be encoded (e.g. resources) |
|
127 * appear as NULL instead of throwing errors. |
|
128 * By default, a deeply-nested resource will |
|
129 * bubble up with an error, so all return values |
|
130 * from encode() should be checked with isError() |
|
131 */ |
|
132 function Services_JSON($use = 0) |
|
133 { |
|
134 $this->use = $use; |
|
135 } |
|
136 |
|
137 /** |
|
138 * convert a string from one UTF-16 char to one UTF-8 char |
|
139 * |
|
140 * Normally should be handled by mb_convert_encoding, but |
|
141 * provides a slower PHP-only method for installations |
|
142 * that lack the multibye string extension. |
|
143 * |
|
144 * @param string $utf16 UTF-16 character |
|
145 * @return string UTF-8 character |
|
146 * @access private |
|
147 */ |
|
148 function utf162utf8($utf16) |
|
149 { |
|
150 // oh please oh please oh please oh please oh please |
|
151 if(function_exists('mb_convert_encoding')) { |
|
152 return mb_convert_encoding($utf16, 'UTF-8', 'UTF-16'); |
|
153 } |
|
154 |
|
155 $bytes = (ord($utf16{0}) << 8) | ord($utf16{1}); |
|
156 |
|
157 switch(true) { |
|
158 case ((0x7F & $bytes) == $bytes): |
|
159 // this case should never be reached, because we are in ASCII range |
|
160 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
161 return chr(0x7F & $bytes); |
|
162 |
|
163 case (0x07FF & $bytes) == $bytes: |
|
164 // return a 2-byte UTF-8 character |
|
165 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
166 return chr(0xC0 | (($bytes >> 6) & 0x1F)) |
|
167 . chr(0x80 | ($bytes & 0x3F)); |
|
168 |
|
169 case (0xFFFF & $bytes) == $bytes: |
|
170 // return a 3-byte UTF-8 character |
|
171 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
172 return chr(0xE0 | (($bytes >> 12) & 0x0F)) |
|
173 . chr(0x80 | (($bytes >> 6) & 0x3F)) |
|
174 . chr(0x80 | ($bytes & 0x3F)); |
|
175 } |
|
176 |
|
177 // ignoring UTF-32 for now, sorry |
|
178 return ''; |
|
179 } |
|
180 |
|
181 /** |
|
182 * convert a string from one UTF-8 char to one UTF-16 char |
|
183 * |
|
184 * Normally should be handled by mb_convert_encoding, but |
|
185 * provides a slower PHP-only method for installations |
|
186 * that lack the multibye string extension. |
|
187 * |
|
188 * @param string $utf8 UTF-8 character |
|
189 * @return string UTF-16 character |
|
190 * @access private |
|
191 */ |
|
192 function utf82utf16($utf8) |
|
193 { |
|
194 // oh please oh please oh please oh please oh please |
|
195 if(function_exists('mb_convert_encoding')) { |
|
196 return mb_convert_encoding($utf8, 'UTF-16', 'UTF-8'); |
|
197 } |
|
198 |
|
199 switch(strlen($utf8)) { |
|
200 case 1: |
|
201 // this case should never be reached, because we are in ASCII range |
|
202 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
203 return $utf8; |
|
204 |
|
205 case 2: |
|
206 // return a UTF-16 character from a 2-byte UTF-8 char |
|
207 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
208 return chr(0x07 & (ord($utf8{0}) >> 2)) |
|
209 . chr((0xC0 & (ord($utf8{0}) << 6)) |
|
210 | (0x3F & ord($utf8{1}))); |
|
211 |
|
212 case 3: |
|
213 // return a UTF-16 character from a 3-byte UTF-8 char |
|
214 // see: http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
215 return chr((0xF0 & (ord($utf8{0}) << 4)) |
|
216 | (0x0F & (ord($utf8{1}) >> 2))) |
|
217 . chr((0xC0 & (ord($utf8{1}) << 6)) |
|
218 | (0x7F & ord($utf8{2}))); |
|
219 } |
|
220 |
|
221 // ignoring UTF-32 for now, sorry |
|
222 return ''; |
|
223 } |
|
224 |
|
225 /** |
|
226 * encodes an arbitrary variable into JSON format (and sends JSON Header) |
|
227 * |
|
228 * @param mixed $var any number, boolean, string, array, or object to be encoded. |
|
229 * see argument 1 to Services_JSON() above for array-parsing behavior. |
|
230 * if var is a strng, note that encode() always expects it |
|
231 * to be in ASCII or UTF-8 format! |
|
232 * |
|
233 * @return mixed JSON string representation of input var or an error if a problem occurs |
|
234 * @access public |
|
235 */ |
|
236 function encode($var) |
|
237 { |
|
238 header('Content-type: application/json'); |
|
239 return $this->_encode($var); |
|
240 } |
|
241 /** |
|
242 * encodes an arbitrary variable into JSON format without JSON Header - warning - may allow CSS!!!!) |
|
243 * |
|
244 * @param mixed $var any number, boolean, string, array, or object to be encoded. |
|
245 * see argument 1 to Services_JSON() above for array-parsing behavior. |
|
246 * if var is a strng, note that encode() always expects it |
|
247 * to be in ASCII or UTF-8 format! |
|
248 * |
|
249 * @return mixed JSON string representation of input var or an error if a problem occurs |
|
250 * @access public |
|
251 */ |
|
252 function encodeUnsafe($var) |
|
253 { |
|
254 return $this->_encode($var); |
|
255 } |
|
256 /** |
|
257 * PRIVATE CODE that does the work of encodes an arbitrary variable into JSON format |
|
258 * |
|
259 * @param mixed $var any number, boolean, string, array, or object to be encoded. |
|
260 * see argument 1 to Services_JSON() above for array-parsing behavior. |
|
261 * if var is a strng, note that encode() always expects it |
|
262 * to be in ASCII or UTF-8 format! |
|
263 * |
|
264 * @return mixed JSON string representation of input var or an error if a problem occurs |
|
265 * @access public |
|
266 */ |
|
267 function _encode($var) |
|
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 (($ord_var_c & 0xE0) == 0xC0): |
|
327 // characters U-00000080 - U-000007FF, mask 110XXXXX |
|
328 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
329 if ($c+1 >= $strlen_var) { |
|
330 $c += 1; |
|
331 $ascii .= '?'; |
|
332 break; |
|
333 } |
|
334 |
|
335 $char = pack('C*', $ord_var_c, ord($var{$c + 1})); |
|
336 $c += 1; |
|
337 $utf16 = $this->utf82utf16($char); |
|
338 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
339 break; |
|
340 |
|
341 case (($ord_var_c & 0xF0) == 0xE0): |
|
342 if ($c+2 >= $strlen_var) { |
|
343 $c += 2; |
|
344 $ascii .= '?'; |
|
345 break; |
|
346 } |
|
347 // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
|
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 $c += 2; |
|
353 $utf16 = $this->utf82utf16($char); |
|
354 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
355 break; |
|
356 |
|
357 case (($ord_var_c & 0xF8) == 0xF0): |
|
358 if ($c+3 >= $strlen_var) { |
|
359 $c += 3; |
|
360 $ascii .= '?'; |
|
361 break; |
|
362 } |
|
363 // characters U-00010000 - U-001FFFFF, mask 11110XXX |
|
364 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
365 $char = pack('C*', $ord_var_c, |
|
366 ord($var{$c + 1}), |
|
367 ord($var{$c + 2}), |
|
368 ord($var{$c + 3})); |
|
369 $c += 3; |
|
370 $utf16 = $this->utf82utf16($char); |
|
371 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
372 break; |
|
373 |
|
374 case (($ord_var_c & 0xFC) == 0xF8): |
|
375 // characters U-00200000 - U-03FFFFFF, mask 111110XX |
|
376 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
377 if ($c+4 >= $strlen_var) { |
|
378 $c += 4; |
|
379 $ascii .= '?'; |
|
380 break; |
|
381 } |
|
382 $char = pack('C*', $ord_var_c, |
|
383 ord($var{$c + 1}), |
|
384 ord($var{$c + 2}), |
|
385 ord($var{$c + 3}), |
|
386 ord($var{$c + 4})); |
|
387 $c += 4; |
|
388 $utf16 = $this->utf82utf16($char); |
|
389 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
390 break; |
|
391 |
|
392 case (($ord_var_c & 0xFE) == 0xFC): |
|
393 if ($c+5 >= $strlen_var) { |
|
394 $c += 5; |
|
395 $ascii .= '?'; |
|
396 break; |
|
397 } |
|
398 // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
|
399 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
400 $char = pack('C*', $ord_var_c, |
|
401 ord($var{$c + 1}), |
|
402 ord($var{$c + 2}), |
|
403 ord($var{$c + 3}), |
|
404 ord($var{$c + 4}), |
|
405 ord($var{$c + 5})); |
|
406 $c += 5; |
|
407 $utf16 = $this->utf82utf16($char); |
|
408 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
409 break; |
|
410 } |
|
411 } |
|
412 return '"'.$ascii.'"'; |
|
413 |
|
414 case 'array': |
|
415 /* |
|
416 * As per JSON spec if any array key is not an integer |
|
417 * we must treat the the whole array as an object. We |
|
418 * also try to catch a sparsely populated associative |
|
419 * array with numeric keys here because some JS engines |
|
420 * will create an array with empty indexes up to |
|
421 * max_index which can cause memory issues and because |
|
422 * the keys, which may be relevant, will be remapped |
|
423 * otherwise. |
|
424 * |
|
425 * As per the ECMA and JSON specification an object may |
|
426 * have any string as a property. Unfortunately due to |
|
427 * a hole in the ECMA specification if the key is a |
|
428 * ECMA reserved word or starts with a digit the |
|
429 * parameter is only accessible using ECMAScript's |
|
430 * bracket notation. |
|
431 */ |
|
432 |
|
433 // treat as a JSON object |
|
434 if (is_array($var) && count($var) && (array_keys($var) !== range(0, sizeof($var) - 1))) { |
|
435 $properties = array_map(array($this, 'name_value'), |
|
436 array_keys($var), |
|
437 array_values($var)); |
|
438 |
|
439 foreach($properties as $property) { |
|
440 if(Services_JSON::isError($property)) { |
|
441 return $property; |
|
442 } |
|
443 } |
|
444 |
|
445 return '{' . join(',', $properties) . '}'; |
|
446 } |
|
447 |
|
448 // treat it like a regular array |
|
449 $elements = array_map(array($this, '_encode'), $var); |
|
450 |
|
451 foreach($elements as $element) { |
|
452 if(Services_JSON::isError($element)) { |
|
453 return $element; |
|
454 } |
|
455 } |
|
456 |
|
457 return '[' . join(',', $elements) . ']'; |
|
458 |
|
459 case 'object': |
|
460 $vars = get_object_vars($var); |
|
461 |
|
462 $properties = array_map(array($this, 'name_value'), |
|
463 array_keys($vars), |
|
464 array_values($vars)); |
|
465 |
|
466 foreach($properties as $property) { |
|
467 if(Services_JSON::isError($property)) { |
|
468 return $property; |
|
469 } |
|
470 } |
|
471 |
|
472 return '{' . join(',', $properties) . '}'; |
|
473 |
|
474 default: |
|
475 return ($this->use & SERVICES_JSON_SUPPRESS_ERRORS) |
|
476 ? 'null' |
|
477 : new Services_JSON_Error(gettype($var)." can not be encoded as JSON string"); |
|
478 } |
|
479 } |
|
480 |
|
481 /** |
|
482 * array-walking function for use in generating JSON-formatted name-value pairs |
|
483 * |
|
484 * @param string $name name of key to use |
|
485 * @param mixed $value reference to an array element to be encoded |
|
486 * |
|
487 * @return string JSON-formatted name-value pair, like '"name":value' |
|
488 * @access private |
|
489 */ |
|
490 function name_value($name, $value) |
|
491 { |
|
492 $encoded_value = $this->_encode($value); |
|
493 |
|
494 if(Services_JSON::isError($encoded_value)) { |
|
495 return $encoded_value; |
|
496 } |
|
497 |
|
498 return $this->_encode(strval($name)) . ':' . $encoded_value; |
|
499 } |
|
500 |
|
501 /** |
|
502 * reduce a string by removing leading and trailing comments and whitespace |
|
503 * |
|
504 * @param $str string string value to strip of comments and whitespace |
|
505 * |
|
506 * @return string string value stripped of comments and whitespace |
|
507 * @access private |
|
508 */ |
|
509 function reduce_string($str) |
|
510 { |
|
511 $str = preg_replace(array( |
|
512 |
|
513 // eliminate single line comments in '// ...' form |
|
514 '#^\s*//(.+)$#m', |
|
515 |
|
516 // eliminate multi-line comments in '/* ... */' form, at start of string |
|
517 '#^\s*/\*(.+)\*/#Us', |
|
518 |
|
519 // eliminate multi-line comments in '/* ... */' form, at end of string |
|
520 '#/\*(.+)\*/\s*$#Us' |
|
521 |
|
522 ), '', $str); |
|
523 |
|
524 // eliminate extraneous space |
|
525 return trim($str); |
|
526 } |
|
527 |
|
528 /** |
|
529 * decodes a JSON string into appropriate variable |
|
530 * |
|
531 * @param string $str JSON-formatted string |
|
532 * |
|
533 * @return mixed number, boolean, string, array, or object |
|
534 * corresponding to given JSON input string. |
|
535 * See argument 1 to Services_JSON() above for object-output behavior. |
|
536 * Note that decode() always returns strings |
|
537 * in ASCII or UTF-8 format! |
|
538 * @access public |
|
539 */ |
|
540 function decode($str) |
|
541 { |
|
542 $str = $this->reduce_string($str); |
|
543 |
|
544 switch (strtolower($str)) { |
|
545 case 'true': |
|
546 return true; |
|
547 |
|
548 case 'false': |
|
549 return false; |
|
550 |
|
551 case 'null': |
|
552 return null; |
|
553 |
|
554 default: |
|
555 $m = array(); |
|
556 |
|
557 if (is_numeric($str)) { |
|
558 // Lookie-loo, it's a number |
|
559 |
|
560 // This would work on its own, but I'm trying to be |
|
561 // good about returning integers where appropriate: |
|
562 // return (float)$str; |
|
563 |
|
564 // Return float or int, as appropriate |
|
565 return ((float)$str == (integer)$str) |
|
566 ? (integer)$str |
|
567 : (float)$str; |
|
568 |
|
569 } elseif (preg_match('/^("|\').*(\1)$/s', $str, $m) && $m[1] == $m[2]) { |
|
570 // STRINGS RETURNED IN UTF-8 FORMAT |
|
571 $delim = substr($str, 0, 1); |
|
572 $chrs = substr($str, 1, -1); |
|
573 $utf8 = ''; |
|
574 $strlen_chrs = strlen($chrs); |
|
575 |
|
576 for ($c = 0; $c < $strlen_chrs; ++$c) { |
|
577 |
|
578 $substr_chrs_c_2 = substr($chrs, $c, 2); |
|
579 $ord_chrs_c = ord($chrs{$c}); |
|
580 |
|
581 switch (true) { |
|
582 case $substr_chrs_c_2 == '\b': |
|
583 $utf8 .= chr(0x08); |
|
584 ++$c; |
|
585 break; |
|
586 case $substr_chrs_c_2 == '\t': |
|
587 $utf8 .= chr(0x09); |
|
588 ++$c; |
|
589 break; |
|
590 case $substr_chrs_c_2 == '\n': |
|
591 $utf8 .= chr(0x0A); |
|
592 ++$c; |
|
593 break; |
|
594 case $substr_chrs_c_2 == '\f': |
|
595 $utf8 .= chr(0x0C); |
|
596 ++$c; |
|
597 break; |
|
598 case $substr_chrs_c_2 == '\r': |
|
599 $utf8 .= chr(0x0D); |
|
600 ++$c; |
|
601 break; |
|
602 |
|
603 case $substr_chrs_c_2 == '\\"': |
|
604 case $substr_chrs_c_2 == '\\\'': |
|
605 case $substr_chrs_c_2 == '\\\\': |
|
606 case $substr_chrs_c_2 == '\\/': |
|
607 if (($delim == '"' && $substr_chrs_c_2 != '\\\'') || |
|
608 ($delim == "'" && $substr_chrs_c_2 != '\\"')) { |
|
609 $utf8 .= $chrs{++$c}; |
|
610 } |
|
611 break; |
|
612 |
|
613 case preg_match('/\\\u[0-9A-F]{4}/i', substr($chrs, $c, 6)): |
|
614 // single, escaped unicode character |
|
615 $utf16 = chr(hexdec(substr($chrs, ($c + 2), 2))) |
|
616 . chr(hexdec(substr($chrs, ($c + 4), 2))); |
|
617 $utf8 .= $this->utf162utf8($utf16); |
|
618 $c += 5; |
|
619 break; |
|
620 |
|
621 case ($ord_chrs_c >= 0x20) && ($ord_chrs_c <= 0x7F): |
|
622 $utf8 .= $chrs{$c}; |
|
623 break; |
|
624 |
|
625 case ($ord_chrs_c & 0xE0) == 0xC0: |
|
626 // characters U-00000080 - U-000007FF, mask 110XXXXX |
|
627 //see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
628 $utf8 .= substr($chrs, $c, 2); |
|
629 ++$c; |
|
630 break; |
|
631 |
|
632 case ($ord_chrs_c & 0xF0) == 0xE0: |
|
633 // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
|
634 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
635 $utf8 .= substr($chrs, $c, 3); |
|
636 $c += 2; |
|
637 break; |
|
638 |
|
639 case ($ord_chrs_c & 0xF8) == 0xF0: |
|
640 // characters U-00010000 - U-001FFFFF, mask 11110XXX |
|
641 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
642 $utf8 .= substr($chrs, $c, 4); |
|
643 $c += 3; |
|
644 break; |
|
645 |
|
646 case ($ord_chrs_c & 0xFC) == 0xF8: |
|
647 // characters U-00200000 - U-03FFFFFF, mask 111110XX |
|
648 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
649 $utf8 .= substr($chrs, $c, 5); |
|
650 $c += 4; |
|
651 break; |
|
652 |
|
653 case ($ord_chrs_c & 0xFE) == 0xFC: |
|
654 // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
|
655 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
656 $utf8 .= substr($chrs, $c, 6); |
|
657 $c += 5; |
|
658 break; |
|
659 |
|
660 } |
|
661 |
|
662 } |
|
663 |
|
664 return $utf8; |
|
665 |
|
666 } elseif (preg_match('/^\[.*\]$/s', $str) || preg_match('/^\{.*\}$/s', $str)) { |
|
667 // array, or object notation |
|
668 |
|
669 if ($str{0} == '[') { |
|
670 $stk = array(SERVICES_JSON_IN_ARR); |
|
671 $arr = array(); |
|
672 } else { |
|
673 if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
|
674 $stk = array(SERVICES_JSON_IN_OBJ); |
|
675 $obj = array(); |
|
676 } else { |
|
677 $stk = array(SERVICES_JSON_IN_OBJ); |
|
678 $obj = new stdClass(); |
|
679 } |
|
680 } |
|
681 |
|
682 array_push($stk, array('what' => SERVICES_JSON_SLICE, |
|
683 'where' => 0, |
|
684 'delim' => false)); |
|
685 |
|
686 $chrs = substr($str, 1, -1); |
|
687 $chrs = $this->reduce_string($chrs); |
|
688 |
|
689 if ($chrs == '') { |
|
690 if (reset($stk) == SERVICES_JSON_IN_ARR) { |
|
691 return $arr; |
|
692 |
|
693 } else { |
|
694 return $obj; |
|
695 |
|
696 } |
|
697 } |
|
698 |
|
699 //print("\nparsing {$chrs}\n"); |
|
700 |
|
701 $strlen_chrs = strlen($chrs); |
|
702 |
|
703 for ($c = 0; $c <= $strlen_chrs; ++$c) { |
|
704 |
|
705 $top = end($stk); |
|
706 $substr_chrs_c_2 = substr($chrs, $c, 2); |
|
707 |
|
708 if (($c == $strlen_chrs) || (($chrs{$c} == ',') && ($top['what'] == SERVICES_JSON_SLICE))) { |
|
709 // found a comma that is not inside a string, array, etc., |
|
710 // OR we've reached the end of the character list |
|
711 $slice = substr($chrs, $top['where'], ($c - $top['where'])); |
|
712 array_push($stk, array('what' => SERVICES_JSON_SLICE, 'where' => ($c + 1), 'delim' => false)); |
|
713 //print("Found split at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
714 |
|
715 if (reset($stk) == SERVICES_JSON_IN_ARR) { |
|
716 // we are in an array, so just push an element onto the stack |
|
717 array_push($arr, $this->decode($slice)); |
|
718 |
|
719 } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
|
720 // we are in an object, so figure |
|
721 // out the property name and set an |
|
722 // element in an associative array, |
|
723 // for now |
|
724 $parts = array(); |
|
725 |
|
726 if (preg_match('/^\s*(["\'].*[^\\\]["\'])\s*:\s*(\S.*),?$/Uis', $slice, $parts)) { |
|
727 // "name":value pair |
|
728 $key = $this->decode($parts[1]); |
|
729 $val = $this->decode($parts[2]); |
|
730 |
|
731 if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
|
732 $obj[$key] = $val; |
|
733 } else { |
|
734 $obj->$key = $val; |
|
735 } |
|
736 } elseif (preg_match('/^\s*(\w+)\s*:\s*(\S.*),?$/Uis', $slice, $parts)) { |
|
737 // name:value pair, where name is unquoted |
|
738 $key = $parts[1]; |
|
739 $val = $this->decode($parts[2]); |
|
740 |
|
741 if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
|
742 $obj[$key] = $val; |
|
743 } else { |
|
744 $obj->$key = $val; |
|
745 } |
|
746 } |
|
747 |
|
748 } |
|
749 |
|
750 } elseif ((($chrs{$c} == '"') || ($chrs{$c} == "'")) && ($top['what'] != SERVICES_JSON_IN_STR)) { |
|
751 // found a quote, and we are not inside a string |
|
752 array_push($stk, array('what' => SERVICES_JSON_IN_STR, 'where' => $c, 'delim' => $chrs{$c})); |
|
753 //print("Found start of string at {$c}\n"); |
|
754 |
|
755 } elseif (($chrs{$c} == $top['delim']) && |
|
756 ($top['what'] == SERVICES_JSON_IN_STR) && |
|
757 ((strlen(substr($chrs, 0, $c)) - strlen(rtrim(substr($chrs, 0, $c), '\\'))) % 2 != 1)) { |
|
758 // found a quote, we're in a string, and it's not escaped |
|
759 // we know that it's not escaped becase there is _not_ an |
|
760 // odd number of backslashes at the end of the string so far |
|
761 array_pop($stk); |
|
762 //print("Found end of string at {$c}: ".substr($chrs, $top['where'], (1 + 1 + $c - $top['where']))."\n"); |
|
763 |
|
764 } elseif (($chrs{$c} == '[') && |
|
765 in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
|
766 // found a left-bracket, and we are in an array, object, or slice |
|
767 array_push($stk, array('what' => SERVICES_JSON_IN_ARR, 'where' => $c, 'delim' => false)); |
|
768 //print("Found start of array at {$c}\n"); |
|
769 |
|
770 } elseif (($chrs{$c} == ']') && ($top['what'] == SERVICES_JSON_IN_ARR)) { |
|
771 // found a right-bracket, and we're in an array |
|
772 array_pop($stk); |
|
773 //print("Found end of array at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
774 |
|
775 } elseif (($chrs{$c} == '{') && |
|
776 in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
|
777 // found a left-brace, and we are in an array, object, or slice |
|
778 array_push($stk, array('what' => SERVICES_JSON_IN_OBJ, 'where' => $c, 'delim' => false)); |
|
779 //print("Found start of object at {$c}\n"); |
|
780 |
|
781 } elseif (($chrs{$c} == '}') && ($top['what'] == SERVICES_JSON_IN_OBJ)) { |
|
782 // found a right-brace, and we're in an object |
|
783 array_pop($stk); |
|
784 //print("Found end of object at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
785 |
|
786 } elseif (($substr_chrs_c_2 == '/*') && |
|
787 in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
|
788 // found a comment start, and we are in an array, object, or slice |
|
789 array_push($stk, array('what' => SERVICES_JSON_IN_CMT, 'where' => $c, 'delim' => false)); |
|
790 $c++; |
|
791 //print("Found start of comment at {$c}\n"); |
|
792 |
|
793 } elseif (($substr_chrs_c_2 == '*/') && ($top['what'] == SERVICES_JSON_IN_CMT)) { |
|
794 // found a comment end, and we're in one now |
|
795 array_pop($stk); |
|
796 $c++; |
|
797 |
|
798 for ($i = $top['where']; $i <= $c; ++$i) |
|
799 $chrs = substr_replace($chrs, ' ', $i, 1); |
|
800 |
|
801 //print("Found end of comment at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
802 |
|
803 } |
|
804 |
|
805 } |
|
806 |
|
807 if (reset($stk) == SERVICES_JSON_IN_ARR) { |
|
808 return $arr; |
|
809 |
|
810 } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
|
811 return $obj; |
|
812 |
|
813 } |
|
814 |
|
815 } |
|
816 } |
|
817 } |
|
818 |
|
819 /** |
|
820 * @todo Ultimately, this should just call PEAR::isError() |
|
821 */ |
|
822 function isError($data, $code = null) |
|
823 { |
|
824 if (class_exists('pear')) { |
|
825 return PEAR::isError($data, $code); |
|
826 } elseif (is_object($data) && (get_class($data) == 'services_json_error' || |
|
827 is_subclass_of($data, 'services_json_error'))) { |
|
828 return true; |
|
829 } |
|
830 |
|
831 return false; |
|
832 } |
|
833 } |
|
834 |
|
835 if (class_exists('PEAR_Error')) { |
|
836 |
|
837 class Services_JSON_Error extends PEAR_Error |
|
838 { |
|
839 function Services_JSON_Error($message = 'unknown error', $code = null, |
|
840 $mode = null, $options = null, $userinfo = null) |
|
841 { |
|
842 parent::PEAR_Error($message, $code, $mode, $options, $userinfo); |
|
843 } |
|
844 } |
|
845 |
|
846 } else { |
|
847 |
|
848 /** |
|
849 * @todo Ultimately, this class shall be descended from PEAR_Error |
|
850 */ |
|
851 class Services_JSON_Error |
|
852 { |
|
853 function Services_JSON_Error($message = 'unknown error', $code = null, |
|
854 $mode = null, $options = null, $userinfo = null) |
|
855 { |
|
856 |
|
857 } |
|
858 } |
|
859 |
|
860 } |