|
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 * @category |
|
49 * @package Services_JSON |
|
50 * @author Michal Migurski <mike-json@teczno.com> |
|
51 * @author Matt Knapp <mdknapp[at]gmail[dot]com> |
|
52 * @author Brett Stimmerman <brettstimmerman[at]gmail[dot]com> |
|
53 * @copyright 2005 Michal Migurski |
|
54 * @version CVS: $Id: JSON.php,v 1.31 2006/06/28 05:54:17 migurski Exp $ |
|
55 * @license http://www.opensource.org/licenses/bsd-license.php |
|
56 * @link http://pear.php.net/pepr/pepr-proposal-show.php?id=198 |
|
57 */ |
|
58 |
|
59 /** |
|
60 * Marker constant for Services_JSON::decode(), used to flag stack state |
|
61 */ |
|
62 define('SERVICES_JSON_SLICE', 1); |
|
63 |
|
64 /** |
|
65 * Marker constant for Services_JSON::decode(), used to flag stack state |
|
66 */ |
|
67 define('SERVICES_JSON_IN_STR', 2); |
|
68 |
|
69 /** |
|
70 * Marker constant for Services_JSON::decode(), used to flag stack state |
|
71 */ |
|
72 define('SERVICES_JSON_IN_ARR', 3); |
|
73 |
|
74 /** |
|
75 * Marker constant for Services_JSON::decode(), used to flag stack state |
|
76 */ |
|
77 define('SERVICES_JSON_IN_OBJ', 4); |
|
78 |
|
79 /** |
|
80 * Marker constant for Services_JSON::decode(), used to flag stack state |
|
81 */ |
|
82 define('SERVICES_JSON_IN_CMT', 5); |
|
83 |
|
84 /** |
|
85 * Behavior switch for Services_JSON::decode() |
|
86 */ |
|
87 define('SERVICES_JSON_LOOSE_TYPE', 16); |
|
88 |
|
89 /** |
|
90 * Behavior switch for Services_JSON::decode() |
|
91 */ |
|
92 define('SERVICES_JSON_SUPPRESS_ERRORS', 32); |
|
93 |
|
94 /** |
|
95 * Converts to and from JSON format. |
|
96 * |
|
97 * @example <code> |
|
98 // create a new instance of Services_JSON |
|
99 $json = new Services_JSON(); |
|
100 |
|
101 // convert a complexe value to JSON notation, and send it to the browser |
|
102 $value = array('foo', 'bar', array(1, 2, 'baz'), array(3, array(4))); |
|
103 $output = $json->encode($value); |
|
104 |
|
105 print($output); |
|
106 // prints: ["foo","bar",[1,2,"baz"],[3,[4]]] |
|
107 |
|
108 // accept incoming POST data, assumed to be in JSON notation |
|
109 $input = file_get_contents('php://input', 1000000); |
|
110 $value = $json->decode($input); |
|
111 </code> |
|
112 * |
|
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 |
|
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 switch (gettype($var)) { |
|
239 case 'boolean': |
|
240 return $var ? 'true' : 'false'; |
|
241 |
|
242 case 'NULL': |
|
243 return 'null'; |
|
244 |
|
245 case 'integer': |
|
246 return (int) $var; |
|
247 |
|
248 case 'double': |
|
249 case 'float': |
|
250 return (float) $var; |
|
251 |
|
252 case 'string': |
|
253 // STRINGS ARE EXPECTED TO BE IN ASCII OR UTF-8 FORMAT |
|
254 $ascii = ''; |
|
255 $strlen_var = strlen($var); |
|
256 |
|
257 /* |
|
258 * Iterate over every character in the string, |
|
259 * escaping with a slash or encoding to UTF-8 where necessary |
|
260 */ |
|
261 for ($c = 0; $c < $strlen_var; ++$c) { |
|
262 |
|
263 $ord_var_c = ord($var{$c}); |
|
264 |
|
265 switch (true) { |
|
266 case $ord_var_c == 0x08: |
|
267 $ascii .= '\b'; |
|
268 break; |
|
269 case $ord_var_c == 0x09: |
|
270 $ascii .= '\t'; |
|
271 break; |
|
272 case $ord_var_c == 0x0A: |
|
273 $ascii .= '\n'; |
|
274 break; |
|
275 case $ord_var_c == 0x0C: |
|
276 $ascii .= '\f'; |
|
277 break; |
|
278 case $ord_var_c == 0x0D: |
|
279 $ascii .= '\r'; |
|
280 break; |
|
281 |
|
282 case $ord_var_c == 0x22: |
|
283 case $ord_var_c == 0x2F: |
|
284 case $ord_var_c == 0x5C: |
|
285 // double quote, slash, slosh |
|
286 $ascii .= '\\'.$var{$c}; |
|
287 break; |
|
288 |
|
289 case (($ord_var_c >= 0x20) && ($ord_var_c <= 0x7F)): |
|
290 // characters U-00000000 - U-0000007F (same as ASCII) |
|
291 $ascii .= $var{$c}; |
|
292 break; |
|
293 |
|
294 case (($ord_var_c & 0xE0) == 0xC0): |
|
295 // characters U-00000080 - U-000007FF, mask 110XXXXX |
|
296 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
297 $char = pack('C*', $ord_var_c, ord($var{$c + 1})); |
|
298 $c += 1; |
|
299 $utf16 = $this->utf82utf16($char); |
|
300 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
301 break; |
|
302 |
|
303 case (($ord_var_c & 0xF0) == 0xE0): |
|
304 // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
|
305 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
306 $char = pack('C*', $ord_var_c, |
|
307 ord($var{$c + 1}), |
|
308 ord($var{$c + 2})); |
|
309 $c += 2; |
|
310 $utf16 = $this->utf82utf16($char); |
|
311 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
312 break; |
|
313 |
|
314 case (($ord_var_c & 0xF8) == 0xF0): |
|
315 // characters U-00010000 - U-001FFFFF, mask 11110XXX |
|
316 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
317 $char = pack('C*', $ord_var_c, |
|
318 ord($var{$c + 1}), |
|
319 ord($var{$c + 2}), |
|
320 ord($var{$c + 3})); |
|
321 $c += 3; |
|
322 $utf16 = $this->utf82utf16($char); |
|
323 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
324 break; |
|
325 |
|
326 case (($ord_var_c & 0xFC) == 0xF8): |
|
327 // characters U-00200000 - U-03FFFFFF, mask 111110XX |
|
328 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
329 $char = pack('C*', $ord_var_c, |
|
330 ord($var{$c + 1}), |
|
331 ord($var{$c + 2}), |
|
332 ord($var{$c + 3}), |
|
333 ord($var{$c + 4})); |
|
334 $c += 4; |
|
335 $utf16 = $this->utf82utf16($char); |
|
336 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
337 break; |
|
338 |
|
339 case (($ord_var_c & 0xFE) == 0xFC): |
|
340 // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
|
341 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
342 $char = pack('C*', $ord_var_c, |
|
343 ord($var{$c + 1}), |
|
344 ord($var{$c + 2}), |
|
345 ord($var{$c + 3}), |
|
346 ord($var{$c + 4}), |
|
347 ord($var{$c + 5})); |
|
348 $c += 5; |
|
349 $utf16 = $this->utf82utf16($char); |
|
350 $ascii .= sprintf('\u%04s', bin2hex($utf16)); |
|
351 break; |
|
352 } |
|
353 } |
|
354 |
|
355 return '"'.$ascii.'"'; |
|
356 |
|
357 case 'array': |
|
358 /* |
|
359 * As per JSON spec if any array key is not an integer |
|
360 * we must treat the the whole array as an object. We |
|
361 * also try to catch a sparsely populated associative |
|
362 * array with numeric keys here because some JS engines |
|
363 * will create an array with empty indexes up to |
|
364 * max_index which can cause memory issues and because |
|
365 * the keys, which may be relevant, will be remapped |
|
366 * otherwise. |
|
367 * |
|
368 * As per the ECMA and JSON specification an object may |
|
369 * have any string as a property. Unfortunately due to |
|
370 * a hole in the ECMA specification if the key is a |
|
371 * ECMA reserved word or starts with a digit the |
|
372 * parameter is only accessible using ECMAScript's |
|
373 * bracket notation. |
|
374 */ |
|
375 |
|
376 // treat as a JSON object |
|
377 if (is_array($var) && count($var) && (array_keys($var) !== range(0, sizeof($var) - 1))) { |
|
378 $properties = array_map(array($this, 'name_value'), |
|
379 array_keys($var), |
|
380 array_values($var)); |
|
381 |
|
382 foreach($properties as $property) { |
|
383 if(Services_JSON::isError($property)) { |
|
384 return $property; |
|
385 } |
|
386 } |
|
387 |
|
388 return '{' . join(',', $properties) . '}'; |
|
389 } |
|
390 |
|
391 // treat it like a regular array |
|
392 $elements = array_map(array($this, 'encode'), $var); |
|
393 |
|
394 foreach($elements as $element) { |
|
395 if(Services_JSON::isError($element)) { |
|
396 return $element; |
|
397 } |
|
398 } |
|
399 |
|
400 return '[' . join(',', $elements) . ']'; |
|
401 |
|
402 case 'object': |
|
403 $vars = get_object_vars($var); |
|
404 |
|
405 $properties = array_map(array($this, 'name_value'), |
|
406 array_keys($vars), |
|
407 array_values($vars)); |
|
408 |
|
409 foreach($properties as $property) { |
|
410 if(Services_JSON::isError($property)) { |
|
411 return $property; |
|
412 } |
|
413 } |
|
414 |
|
415 return '{' . join(',', $properties) . '}'; |
|
416 |
|
417 default: |
|
418 return ($this->use & SERVICES_JSON_SUPPRESS_ERRORS) |
|
419 ? 'null' |
|
420 : new Services_JSON_Error(gettype($var)." can not be encoded as JSON string"); |
|
421 } |
|
422 } |
|
423 |
|
424 /** |
|
425 * array-walking function for use in generating JSON-formatted name-value pairs |
|
426 * |
|
427 * @param string $name name of key to use |
|
428 * @param mixed $value reference to an array element to be encoded |
|
429 * |
|
430 * @return string JSON-formatted name-value pair, like '"name":value' |
|
431 * @access private |
|
432 */ |
|
433 function name_value($name, $value) |
|
434 { |
|
435 $encoded_value = $this->encode($value); |
|
436 |
|
437 if(Services_JSON::isError($encoded_value)) { |
|
438 return $encoded_value; |
|
439 } |
|
440 |
|
441 return $this->encode(strval($name)) . ':' . $encoded_value; |
|
442 } |
|
443 |
|
444 /** |
|
445 * reduce a string by removing leading and trailing comments and whitespace |
|
446 * |
|
447 * @param $str string string value to strip of comments and whitespace |
|
448 * |
|
449 * @return string string value stripped of comments and whitespace |
|
450 * @access private |
|
451 */ |
|
452 function reduce_string($str) |
|
453 { |
|
454 $str = preg_replace(array( |
|
455 |
|
456 // eliminate single line comments in '// ...' form |
|
457 '#^\s*//(.+)$#m', |
|
458 |
|
459 // eliminate multi-line comments in '/* ... */' form, at start of string |
|
460 '#^\s*/\*(.+)\*/#Us', |
|
461 |
|
462 // eliminate multi-line comments in '/* ... */' form, at end of string |
|
463 '#/\*(.+)\*/\s*$#Us' |
|
464 |
|
465 ), '', $str); |
|
466 |
|
467 // eliminate extraneous space |
|
468 return trim($str); |
|
469 } |
|
470 |
|
471 /** |
|
472 * decodes a JSON string into appropriate variable |
|
473 * |
|
474 * @param string $str JSON-formatted string |
|
475 * |
|
476 * @return mixed number, boolean, string, array, or object |
|
477 * corresponding to given JSON input string. |
|
478 * See argument 1 to Services_JSON() above for object-output behavior. |
|
479 * Note that decode() always returns strings |
|
480 * in ASCII or UTF-8 format! |
|
481 * @access public |
|
482 */ |
|
483 function decode($str) |
|
484 { |
|
485 $str = $this->reduce_string($str); |
|
486 |
|
487 switch (strtolower($str)) { |
|
488 case 'true': |
|
489 return true; |
|
490 |
|
491 case 'false': |
|
492 return false; |
|
493 |
|
494 case 'null': |
|
495 return null; |
|
496 |
|
497 default: |
|
498 $m = array(); |
|
499 |
|
500 if (is_numeric($str)) { |
|
501 // Lookie-loo, it's a number |
|
502 |
|
503 // This would work on its own, but I'm trying to be |
|
504 // good about returning integers where appropriate: |
|
505 // return (float)$str; |
|
506 |
|
507 // Return float or int, as appropriate |
|
508 return ((float)$str == (integer)$str) |
|
509 ? (integer)$str |
|
510 : (float)$str; |
|
511 |
|
512 } elseif (preg_match('/^("|\').*(\1)$/s', $str, $m) && $m[1] == $m[2]) { |
|
513 // STRINGS RETURNED IN UTF-8 FORMAT |
|
514 $delim = substr($str, 0, 1); |
|
515 $chrs = substr($str, 1, -1); |
|
516 $utf8 = ''; |
|
517 $strlen_chrs = strlen($chrs); |
|
518 |
|
519 for ($c = 0; $c < $strlen_chrs; ++$c) { |
|
520 |
|
521 $substr_chrs_c_2 = substr($chrs, $c, 2); |
|
522 $ord_chrs_c = ord($chrs{$c}); |
|
523 |
|
524 switch (true) { |
|
525 case $substr_chrs_c_2 == '\b': |
|
526 $utf8 .= chr(0x08); |
|
527 ++$c; |
|
528 break; |
|
529 case $substr_chrs_c_2 == '\t': |
|
530 $utf8 .= chr(0x09); |
|
531 ++$c; |
|
532 break; |
|
533 case $substr_chrs_c_2 == '\n': |
|
534 $utf8 .= chr(0x0A); |
|
535 ++$c; |
|
536 break; |
|
537 case $substr_chrs_c_2 == '\f': |
|
538 $utf8 .= chr(0x0C); |
|
539 ++$c; |
|
540 break; |
|
541 case $substr_chrs_c_2 == '\r': |
|
542 $utf8 .= chr(0x0D); |
|
543 ++$c; |
|
544 break; |
|
545 |
|
546 case $substr_chrs_c_2 == '\\"': |
|
547 case $substr_chrs_c_2 == '\\\'': |
|
548 case $substr_chrs_c_2 == '\\\\': |
|
549 case $substr_chrs_c_2 == '\\/': |
|
550 if (($delim == '"' && $substr_chrs_c_2 != '\\\'') || |
|
551 ($delim == "'" && $substr_chrs_c_2 != '\\"')) { |
|
552 $utf8 .= $chrs{++$c}; |
|
553 } |
|
554 break; |
|
555 |
|
556 case preg_match('/\\\u[0-9A-F]{4}/i', substr($chrs, $c, 6)): |
|
557 // single, escaped unicode character |
|
558 $utf16 = chr(hexdec(substr($chrs, ($c + 2), 2))) |
|
559 . chr(hexdec(substr($chrs, ($c + 4), 2))); |
|
560 $utf8 .= $this->utf162utf8($utf16); |
|
561 $c += 5; |
|
562 break; |
|
563 |
|
564 case ($ord_chrs_c >= 0x20) && ($ord_chrs_c <= 0x7F): |
|
565 $utf8 .= $chrs{$c}; |
|
566 break; |
|
567 |
|
568 case ($ord_chrs_c & 0xE0) == 0xC0: |
|
569 // characters U-00000080 - U-000007FF, mask 110XXXXX |
|
570 //see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
571 $utf8 .= substr($chrs, $c, 2); |
|
572 ++$c; |
|
573 break; |
|
574 |
|
575 case ($ord_chrs_c & 0xF0) == 0xE0: |
|
576 // characters U-00000800 - U-0000FFFF, mask 1110XXXX |
|
577 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
578 $utf8 .= substr($chrs, $c, 3); |
|
579 $c += 2; |
|
580 break; |
|
581 |
|
582 case ($ord_chrs_c & 0xF8) == 0xF0: |
|
583 // characters U-00010000 - U-001FFFFF, mask 11110XXX |
|
584 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
585 $utf8 .= substr($chrs, $c, 4); |
|
586 $c += 3; |
|
587 break; |
|
588 |
|
589 case ($ord_chrs_c & 0xFC) == 0xF8: |
|
590 // characters U-00200000 - U-03FFFFFF, mask 111110XX |
|
591 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
592 $utf8 .= substr($chrs, $c, 5); |
|
593 $c += 4; |
|
594 break; |
|
595 |
|
596 case ($ord_chrs_c & 0xFE) == 0xFC: |
|
597 // characters U-04000000 - U-7FFFFFFF, mask 1111110X |
|
598 // see http://www.cl.cam.ac.uk/~mgk25/unicode.html#utf-8 |
|
599 $utf8 .= substr($chrs, $c, 6); |
|
600 $c += 5; |
|
601 break; |
|
602 |
|
603 } |
|
604 |
|
605 } |
|
606 |
|
607 return $utf8; |
|
608 |
|
609 } elseif (preg_match('/^\[.*\]$/s', $str) || preg_match('/^\{.*\}$/s', $str)) { |
|
610 // array, or object notation |
|
611 |
|
612 if ($str{0} == '[') { |
|
613 $stk = array(SERVICES_JSON_IN_ARR); |
|
614 $arr = array(); |
|
615 } else { |
|
616 if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
|
617 $stk = array(SERVICES_JSON_IN_OBJ); |
|
618 $obj = array(); |
|
619 } else { |
|
620 $stk = array(SERVICES_JSON_IN_OBJ); |
|
621 $obj = new stdClass(); |
|
622 } |
|
623 } |
|
624 |
|
625 array_push($stk, array('what' => SERVICES_JSON_SLICE, |
|
626 'where' => 0, |
|
627 'delim' => false)); |
|
628 |
|
629 $chrs = substr($str, 1, -1); |
|
630 $chrs = $this->reduce_string($chrs); |
|
631 |
|
632 if ($chrs == '') { |
|
633 if (reset($stk) == SERVICES_JSON_IN_ARR) { |
|
634 return $arr; |
|
635 |
|
636 } else { |
|
637 return $obj; |
|
638 |
|
639 } |
|
640 } |
|
641 |
|
642 //print("\nparsing {$chrs}\n"); |
|
643 |
|
644 $strlen_chrs = strlen($chrs); |
|
645 |
|
646 for ($c = 0; $c <= $strlen_chrs; ++$c) { |
|
647 |
|
648 $top = end($stk); |
|
649 $substr_chrs_c_2 = substr($chrs, $c, 2); |
|
650 |
|
651 if (($c == $strlen_chrs) || (($chrs{$c} == ',') && ($top['what'] == SERVICES_JSON_SLICE))) { |
|
652 // found a comma that is not inside a string, array, etc., |
|
653 // OR we've reached the end of the character list |
|
654 $slice = substr($chrs, $top['where'], ($c - $top['where'])); |
|
655 array_push($stk, array('what' => SERVICES_JSON_SLICE, 'where' => ($c + 1), 'delim' => false)); |
|
656 //print("Found split at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
657 |
|
658 if (reset($stk) == SERVICES_JSON_IN_ARR) { |
|
659 // we are in an array, so just push an element onto the stack |
|
660 array_push($arr, $this->decode($slice)); |
|
661 |
|
662 } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
|
663 // we are in an object, so figure |
|
664 // out the property name and set an |
|
665 // element in an associative array, |
|
666 // for now |
|
667 $parts = array(); |
|
668 |
|
669 if (preg_match('/^\s*(["\'].*[^\\\]["\'])\s*:\s*(\S.*),?$/Uis', $slice, $parts)) { |
|
670 // "name":value pair |
|
671 $key = $this->decode($parts[1]); |
|
672 $val = $this->decode($parts[2]); |
|
673 |
|
674 if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
|
675 $obj[$key] = $val; |
|
676 } else { |
|
677 $obj->$key = $val; |
|
678 } |
|
679 } elseif (preg_match('/^\s*(\w+)\s*:\s*(\S.*),?$/Uis', $slice, $parts)) { |
|
680 // name:value pair, where name is unquoted |
|
681 $key = $parts[1]; |
|
682 $val = $this->decode($parts[2]); |
|
683 |
|
684 if ($this->use & SERVICES_JSON_LOOSE_TYPE) { |
|
685 $obj[$key] = $val; |
|
686 } else { |
|
687 $obj->$key = $val; |
|
688 } |
|
689 } |
|
690 |
|
691 } |
|
692 |
|
693 } elseif ((($chrs{$c} == '"') || ($chrs{$c} == "'")) && ($top['what'] != SERVICES_JSON_IN_STR)) { |
|
694 // found a quote, and we are not inside a string |
|
695 array_push($stk, array('what' => SERVICES_JSON_IN_STR, 'where' => $c, 'delim' => $chrs{$c})); |
|
696 //print("Found start of string at {$c}\n"); |
|
697 |
|
698 } elseif (($chrs{$c} == $top['delim']) && |
|
699 ($top['what'] == SERVICES_JSON_IN_STR) && |
|
700 ((strlen(substr($chrs, 0, $c)) - strlen(rtrim(substr($chrs, 0, $c), '\\'))) % 2 != 1)) { |
|
701 // found a quote, we're in a string, and it's not escaped |
|
702 // we know that it's not escaped becase there is _not_ an |
|
703 // odd number of backslashes at the end of the string so far |
|
704 array_pop($stk); |
|
705 //print("Found end of string at {$c}: ".substr($chrs, $top['where'], (1 + 1 + $c - $top['where']))."\n"); |
|
706 |
|
707 } elseif (($chrs{$c} == '[') && |
|
708 in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
|
709 // found a left-bracket, and we are in an array, object, or slice |
|
710 array_push($stk, array('what' => SERVICES_JSON_IN_ARR, 'where' => $c, 'delim' => false)); |
|
711 //print("Found start of array at {$c}\n"); |
|
712 |
|
713 } elseif (($chrs{$c} == ']') && ($top['what'] == SERVICES_JSON_IN_ARR)) { |
|
714 // found a right-bracket, and we're in an array |
|
715 array_pop($stk); |
|
716 //print("Found end of array at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
717 |
|
718 } elseif (($chrs{$c} == '{') && |
|
719 in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
|
720 // found a left-brace, and we are in an array, object, or slice |
|
721 array_push($stk, array('what' => SERVICES_JSON_IN_OBJ, 'where' => $c, 'delim' => false)); |
|
722 //print("Found start of object at {$c}\n"); |
|
723 |
|
724 } elseif (($chrs{$c} == '}') && ($top['what'] == SERVICES_JSON_IN_OBJ)) { |
|
725 // found a right-brace, and we're in an object |
|
726 array_pop($stk); |
|
727 //print("Found end of object at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
728 |
|
729 } elseif (($substr_chrs_c_2 == '/*') && |
|
730 in_array($top['what'], array(SERVICES_JSON_SLICE, SERVICES_JSON_IN_ARR, SERVICES_JSON_IN_OBJ))) { |
|
731 // found a comment start, and we are in an array, object, or slice |
|
732 array_push($stk, array('what' => SERVICES_JSON_IN_CMT, 'where' => $c, 'delim' => false)); |
|
733 $c++; |
|
734 //print("Found start of comment at {$c}\n"); |
|
735 |
|
736 } elseif (($substr_chrs_c_2 == '*/') && ($top['what'] == SERVICES_JSON_IN_CMT)) { |
|
737 // found a comment end, and we're in one now |
|
738 array_pop($stk); |
|
739 $c++; |
|
740 |
|
741 for ($i = $top['where']; $i <= $c; ++$i) |
|
742 $chrs = substr_replace($chrs, ' ', $i, 1); |
|
743 |
|
744 //print("Found end of comment at {$c}: ".substr($chrs, $top['where'], (1 + $c - $top['where']))."\n"); |
|
745 |
|
746 } |
|
747 |
|
748 } |
|
749 |
|
750 if (reset($stk) == SERVICES_JSON_IN_ARR) { |
|
751 return $arr; |
|
752 |
|
753 } elseif (reset($stk) == SERVICES_JSON_IN_OBJ) { |
|
754 return $obj; |
|
755 |
|
756 } |
|
757 |
|
758 } |
|
759 } |
|
760 } |
|
761 |
|
762 /** |
|
763 * @todo Ultimately, this should just call PEAR::isError() |
|
764 */ |
|
765 function isError($data, $code = null) |
|
766 { |
|
767 if (class_exists('pear')) { |
|
768 return PEAR::isError($data, $code); |
|
769 } elseif (is_object($data) && (get_class($data) == 'services_json_error' || |
|
770 is_subclass_of($data, 'services_json_error'))) { |
|
771 return true; |
|
772 } |
|
773 |
|
774 return false; |
|
775 } |
|
776 } |
|
777 |
|
778 if (class_exists('PEAR_Error')) { |
|
779 |
|
780 class Services_JSON_Error extends PEAR_Error |
|
781 { |
|
782 function Services_JSON_Error($message = 'unknown error', $code = null, |
|
783 $mode = null, $options = null, $userinfo = null) |
|
784 { |
|
785 parent::PEAR_Error($message, $code, $mode, $options, $userinfo); |
|
786 } |
|
787 } |
|
788 |
|
789 } else { |
|
790 |
|
791 /** |
|
792 * @todo Ultimately, this class shall be descended from PEAR_Error |
|
793 */ |
|
794 class Services_JSON_Error |
|
795 { |
|
796 function Services_JSON_Error($message = 'unknown error', $code = null, |
|
797 $mode = null, $options = null, $userinfo = null) |
|
798 { |
|
799 |
|
800 } |
|
801 } |
|
802 |
|
803 } |
|
804 |
|
805 ?> |