Context Navigation

wbxmldecoder.php @ 7589

Revision 7589, 20.3 KB checked in by douglas, 11 years ago (diff)
Ticket #3209 - Integrar módulo de sincronização Z-push ao Expresso

Line
1	<?php
2	/***********************************************
3	* File : wbxmldecoder.php
4	* Project : Z-Push
5	* Descr : WBXMLDecoder decodes from Wap Binary XML
6	*
7	* Created : 01.10.2007
8	*
9	* Copyright 2007 - 2012 Zarafa Deutschland GmbH
10	*
11	* This program is free software: you can redistribute it and/or modify
12	* it under the terms of the GNU Affero General Public License, version 3,
13	* as published by the Free Software Foundation with the following additional
14	* term according to sec. 7:
15	*
16	* According to sec. 7 of the GNU Affero General Public License, version 3,
17	* the terms of the AGPL are supplemented with the following terms:
18	*
19	* "Zarafa" is a registered trademark of Zarafa B.V.
20	* "Z-Push" is a registered trademark of Zarafa Deutschland GmbH
21	* The licensing of the Program under the AGPL does not imply a trademark license.
22	* Therefore any rights, title and interest in our trademarks remain entirely with us.
23	*
24	* However, if you propagate an unmodified version of the Program you are
25	* allowed to use the term "Z-Push" to indicate that you distribute the Program.
26	* Furthermore you may use our trademarks where it is necessary to indicate
27	* the intended purpose of a product or service provided you use it in accordance
28	* with honest practices in industrial or commercial matters.
29	* If you want to propagate modified versions of the Program under the name "Z-Push",
30	* you may only do so if you have a written permission by Zarafa Deutschland GmbH
31	* (to acquire a permission please contact Zarafa at trademark@zarafa.com).
32	*
33	* This program is distributed in the hope that it will be useful,
34	* but WITHOUT ANY WARRANTY; without even the implied warranty of
35	* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
36	* GNU Affero General Public License for more details.
37	*
38	* You should have received a copy of the GNU Affero General Public License
39	* along with this program. If not, see <http://www.gnu.org/licenses/>.
40	*
41	* Consult LICENSE file for details
42	************************************************/
43
44
45	class WBXMLDecoder extends WBXMLDefs {
46	private $in;
47
48	private $version;
49	private $publicid;
50	private $publicstringid;
51	private $charsetid;
52	private $stringtable;
53
54	private $tagcp = 0;
55	private $attrcp = 0;
56
57	private $ungetbuffer;
58
59	private $logStack = array();
60
61	private $inputBuffer = "";
62	private $isWBXML = true;
63
64	const VERSION = 0x03;
65
66	/**
67	* WBXML Decode Constructor
68	*
69	* @param stream $input the incoming data stream
70	*
71	* @access public
72	*/
73	public function WBXMLDecoder($input) {
74	// make sure WBXML_DEBUG is defined. It should be at this point
75	if (!defined('WBXML_DEBUG')) define('WBXML_DEBUG', false);
76
77	$this->in = $input;
78
79	$this->readVersion();
80	if (isset($this->version) && $this->version != self::VERSION) {
81	$this->isWBXML = false;
82	return;
83	}
84
85	$this->publicid = $this->getMBUInt();
86	if($this->publicid == 0) {
87	$this->publicstringid = $this->getMBUInt();
88	}
89
90	$this->charsetid = $this->getMBUInt();
91	$this->stringtable = $this->getStringTable();
92	}
93
94	/**
95	* Returns either start, content or end, and auto-concatenates successive content
96	*
97	* @access public
98	* @return element/value
99	*/
100	public function getElement() {
101	$element = $this->getToken();
102
103	switch($element[EN_TYPE]) {
104	case EN_TYPE_STARTTAG:
105	return $element;
106	case EN_TYPE_ENDTAG:
107	return $element;
108	case EN_TYPE_CONTENT:
109	while(1) {
110	$next = $this->getToken();
111	if($next == false)
112	return false;
113	else if($next[EN_TYPE] == EN_CONTENT) {
114	$element[EN_CONTENT] .= $next[EN_CONTENT];
115	} else {
116	$this->ungetElement($next);
117	break;
118	}
119	}
120	return $element;
121	}
122
123	return false;
124	}
125
126	/**
127	* Get a peek at the next element
128	*
129	* @access public
130	* @return element
131	*/
132	public function peek() {
133	$element = $this->getElement();
134	$this->ungetElement($element);
135	return $element;
136	}
137
138	/**
139	* Get the element of a StartTag
140	*
141	* @param $tag
142	*
143	* @access public
144	* @return element/boolean returns false if not available
145	*/
146	public function getElementStartTag($tag) {
147	$element = $this->getToken();
148
149	if($element[EN_TYPE] == EN_TYPE_STARTTAG && $element[EN_TAG] == $tag)
150	return $element;
151	else {
152	ZLog::Write(LOGLEVEL_WBXMLSTACK, sprintf("WBXMLDecoder->getElementStartTag(): unmatched WBXML tag: '%s' matching '%s' type '%s' flags '%s'", $tag, ((isset($element[EN_TAG]))?$element[EN_TAG]:""), ((isset($element[EN_TYPE]))?$element[EN_TYPE]:""), ((isset($element[EN_FLAGS]))?$element[EN_FLAGS]:"")));
153	$this->ungetElement($element);
154	}
155
156	return false;
157	}
158
159	/**
160	* Get the element of a EndTag
161	*
162	* @access public
163	* @return element/boolean returns false if not available
164	*/
165	public function getElementEndTag() {
166	$element = $this->getToken();
167
168	if($element[EN_TYPE] == EN_TYPE_ENDTAG)
169	return $element;
170	else {
171	ZLog::Write(LOGLEVEL_WBXMLSTACK, sprintf("WBXMLDecoder->getElementEndTag(): unmatched WBXML tag: '%s' type '%s' flags '%s'", ((isset($element[EN_TAG]))?$element[EN_TAG]:""), ((isset($element[EN_TYPE]))?$element[EN_TYPE]:""), ((isset($element[EN_FLAGS]))?$element[EN_FLAGS]:"")));
172
173	$bt = debug_backtrace();
174	ZLog::Write(LOGLEVEL_ERROR, sprintf("WBXMLDecoder->getElementEndTag(): could not read end tag in '%s'. Please enable the LOGLEVEL_WBXML and send the log to the Z-Push dev team.", $bt[0]["file"] . ":" . $bt[0]["line"]));
175
176	// log the remaining wbxml content
177	$this->ungetElement($element);
178	while($el = $this->getElement());
179	}
180
181	return false;
182	}
183
184	/**
185	* Get the content of an element
186	*
187	* @access public
188	* @return string/boolean returns false if not available
189	*/
190	public function getElementContent() {
191	$element = $this->getToken();
192
193	if($element[EN_TYPE] == EN_TYPE_CONTENT) {
194	return $element[EN_CONTENT];
195	}
196	else {
197	ZLog::Write(LOGLEVEL_WBXMLSTACK, sprintf("WBXMLDecoder->getElementContent(): unmatched WBXML content: '%s' type '%s' flags '%s'", ((isset($element[EN_TAG]))?$element[EN_TAG]:""), ((isset($element[EN_TYPE]))?$element[EN_TYPE]:""), ((isset($element[EN_FLAGS]))?$element[EN_FLAGS]:"")));
198	$this->ungetElement($element);
199	}
200
201	return false;
202	}
203
204	/**
205	* 'Ungets' an element writing it into a buffer to be 'get' again
206	*
207	* @param element $element the element to get ungetten
208	*
209	* @access public
210	* @return
211	*/
212	public function ungetElement($element) {
213	if($this->ungetbuffer)
214	ZLog::Write(LOGLEVEL_ERROR,sprintf("WBXMLDecoder->ungetElement(): WBXML double unget on tag: '%s' type '%s' flags '%s'", ((isset($element[EN_TAG]))?$element[EN_TAG]:""), ((isset($element[EN_TYPE]))?$element[EN_TYPE]:""), ((isset($element[EN_FLAGS]))?$element[EN_FLAGS]:"")));
215
216	$this->ungetbuffer = $element;
217	}
218
219	/**
220	* Returns the plain input stream
221	*
222	* @access public
223	* @return string
224	*/
225	public function GetPlainInputStream() {
226	$plain = $this->inputBuffer;
227	while($data = fread($this->in, 4096))
228	$plain .= $data;
229
230	return $plain;
231	}
232
233	/**
234	* Returns if the input is WBXML
235	*
236	* @access public
237	* @return boolean
238	*/
239	public function IsWBXML() {
240	return $this->isWBXML;
241	}
242
243
244
245	/**----------------------------------------------------------------------------------------------------------
246	* Private WBXMLDecoder stuff
247	*/
248
249	/**
250	* Returns the next token
251	*
252	* @access private
253	* @return token
254	*/
255	private function getToken() {
256	// See if there's something in the ungetBuffer
257	if($this->ungetbuffer) {
258	$element = $this->ungetbuffer;
259	$this->ungetbuffer = false;
260	return $element;
261	}
262
263	$el = $this->_getToken();
264	$this->logToken($el);
265
266	return $el;
267	}
268
269	/**
270	* Log the a token to ZLog
271	*
272	* @param string $el token
273	*
274	* @access private
275	* @return
276	*/
277	private function logToken($el) {
278	if(!WBXML_DEBUG)
279	return;
280
281	$spaces = str_repeat(" ", count($this->logStack));
282
283	switch($el[EN_TYPE]) {
284	case EN_TYPE_STARTTAG:
285	if($el[EN_FLAGS] & EN_FLAGS_CONTENT) {
286	ZLog::Write(LOGLEVEL_WBXML,"I " . $spaces . " <". $el[EN_TAG] . ">");
287	array_push($this->logStack, $el[EN_TAG]);
288	} else
289	ZLog::Write(LOGLEVEL_WBXML,"I " . $spaces . " <" . $el[EN_TAG] . "/>");
290
291	break;
292	case EN_TYPE_ENDTAG:
293	$tag = array_pop($this->logStack);
294	ZLog::Write(LOGLEVEL_WBXML,"I " . $spaces . "</" . $tag . ">");
295	break;
296	case EN_TYPE_CONTENT:
297	ZLog::Write(LOGLEVEL_WBXML,"I " . $spaces . " " . $el[EN_CONTENT]);
298	break;
299	}
300	}
301
302	/**
303	* Returns either a start tag, content or end tag
304	*
305	* @access private
306	* @return
307	*/
308	private function _getToken() {
309	// Get the data from the input stream
310	$element = array();
311
312	while(1) {
313	$byte = $this->getByte();
314
315	if(!isset($byte))
316	break;
317
318	switch($byte) {
319	case WBXML_SWITCH_PAGE:
320	$this->tagcp = $this->getByte();
321	continue;
322
323	case WBXML_END:
324	$element[EN_TYPE] = EN_TYPE_ENDTAG;
325	return $element;
326
327	case WBXML_ENTITY:
328	$entity = $this->getMBUInt();
329	$element[EN_TYPE] = EN_TYPE_CONTENT;
330	$element[EN_CONTENT] = $this->entityToCharset($entity);
331	return $element;
332
333	case WBXML_STR_I:
334	$element[EN_TYPE] = EN_TYPE_CONTENT;
335	$element[EN_CONTENT] = $this->getTermStr();
336	return $element;
337
338	case WBXML_LITERAL:
339	$element[EN_TYPE] = EN_TYPE_STARTTAG;
340	$element[EN_TAG] = $this->getStringTableEntry($this->getMBUInt());
341	$element[EN_FLAGS] = 0;
342	return $element;
343
344	case WBXML_EXT_I_0:
345	case WBXML_EXT_I_1:
346	case WBXML_EXT_I_2:
347	$this->getTermStr();
348	// Ignore extensions
349	continue;
350
351	case WBXML_PI:
352	// Ignore PI
353	$this->getAttributes();
354	continue;
355
356	case WBXML_LITERAL_C:
357	$element[EN_TYPE] = EN_TYPE_STARTTAG;
358	$element[EN_TAG] = $this->getStringTableEntry($this->getMBUInt());
359	$element[EN_FLAGS] = EN_FLAGS_CONTENT;
360	return $element;
361
362	case WBXML_EXT_T_0:
363	case WBXML_EXT_T_1:
364	case WBXML_EXT_T_2:
365	$this->getMBUInt();
366	// Ingore extensions;
367	continue;
368
369	case WBXML_STR_T:
370	$element[EN_TYPE] = EN_TYPE_CONTENT;
371	$element[EN_CONTENT] = $this->getStringTableEntry($this->getMBUInt());
372	return $element;
373
374	case WBXML_LITERAL_A:
375	$element[EN_TYPE] = EN_TYPE_STARTTAG;
376	$element[EN_TAG] = $this->getStringTableEntry($this->getMBUInt());
377	$element[EN_ATTRIBUTES] = $this->getAttributes();
378	$element[EN_FLAGS] = EN_FLAGS_ATTRIBUTES;
379	return $element;
380	case WBXML_EXT_0:
381	case WBXML_EXT_1:
382	case WBXML_EXT_2:
383	continue;
384
385	case WBXML_OPAQUE:
386	$length = $this->getMBUInt();
387	$element[EN_TYPE] = EN_TYPE_CONTENT;
388	$element[EN_CONTENT] = $this->getOpaque($length);
389	return $element;
390
391	case WBXML_LITERAL_AC:
392	$element[EN_TYPE] = EN_TYPE_STARTTAG;
393	$element[EN_TAG] = $this->getStringTableEntry($this->getMBUInt());
394	$element[EN_ATTRIBUTES] = $this->getAttributes();
395	$element[EN_FLAGS] = EN_FLAGS_ATTRIBUTES \| EN_FLAGS_CONTENT;
396	return $element;
397
398	default:
399	$element[EN_TYPE] = EN_TYPE_STARTTAG;
400	$element[EN_TAG] = $this->getMapping($this->tagcp, $byte & 0x3f);
401	$element[EN_FLAGS] = ($byte & 0x80 ? EN_FLAGS_ATTRIBUTES : 0) \| ($byte & 0x40 ? EN_FLAGS_CONTENT : 0);
402	if($byte & 0x80)
403	$element[EN_ATTRIBUTES] = $this->getAttributes();
404	return $element;
405	}
406	}
407	}
408
409	/**
410	* Gets attributes
411	*
412	* @access private
413	* @return
414	*/
415	private function getAttributes() {
416	$attributes = array();
417	$attr = "";
418
419	while(1) {
420	$byte = $this->getByte();
421
422	if(count($byte) == 0)
423	break;
424
425	switch($byte) {
426	case WBXML_SWITCH_PAGE:
427	$this->attrcp = $this->getByte();
428	break;
429
430	case WBXML_END:
431	if($attr != "")
432	$attributes += $this->splitAttribute($attr);
433
434	return $attributes;
435
436	case WBXML_ENTITY:
437	$entity = $this->getMBUInt();
438	$attr .= $this->entityToCharset($entity);
439	return $attr; /* fmbiete's contribution r1534, ZP-324 */
440
441	case WBXML_STR_I:
442	$attr .= $this->getTermStr();
443	return $attr; /* fmbiete's contribution r1534, ZP-324 */
444
445	case WBXML_LITERAL:
446	if($attr != "")
447	$attributes += $this->splitAttribute($attr);
448
449	$attr = $this->getStringTableEntry($this->getMBUInt());
450	return $attr; /* fmbiete's contribution r1534, ZP-324 */
451
452	case WBXML_EXT_I_0:
453	case WBXML_EXT_I_1:
454	case WBXML_EXT_I_2:
455	$this->getTermStr();
456	continue;
457
458	case WBXML_PI:
459	case WBXML_LITERAL_C:
460	// Invalid
461	return false;
462
463	case WBXML_EXT_T_0:
464	case WBXML_EXT_T_1:
465	case WBXML_EXT_T_2:
466	$this->getMBUInt();
467	continue;
468
469	case WBXML_STR_T:
470	$attr .= $this->getStringTableEntry($this->getMBUInt());
471	return $attr; /* fmbiete's contribution r1534, ZP-324 */
472
473	case WBXML_LITERAL_A:
474	return false;
475
476	case WBXML_EXT_0:
477	case WBXML_EXT_1:
478	case WBXML_EXT_2:
479	continue;
480
481	case WBXML_OPAQUE:
482	$length = $this->getMBUInt();
483	$attr .= $this->getOpaque($length);
484	return $attr; /* fmbiete's contribution r1534, ZP-324 */
485
486	case WBXML_LITERAL_AC:
487	return false;
488
489	default:
490	if($byte < 128) {
491	if($attr != "") {
492	$attributes += $this->splitAttribute($attr);
493	$attr = "";
494	}
495	}
496	$attr .= $this->getMapping($this->attrcp, $byte);
497	break;
498	}
499	}
500	}
501
502	/**
503	* Splits an attribute
504	*
505	* @param string $attr attribute to be splitted
506	*
507	* @access private
508	* @return array
509	*/
510	private function splitAttribute($attr) {
511	$attributes = array();
512
513	$pos = strpos($attr,chr(61)); // equals sign
514
515	if($pos)
516	$attributes[substr($attr, 0, $pos)] = substr($attr, $pos+1);
517	else
518	$attributes[$attr] = null;
519
520	return $attributes;
521	}
522
523	/**
524	* Reads from the stream until getting a string terminator
525	*
526	* @access private
527	* @return string
528	*/
529	private function getTermStr() {
530	$str = "";
531	while(1) {
532	$in = $this->getByte();
533
534	if($in == 0)
535	break;
536	else
537	$str .= chr($in);
538	}
539
540	return $str;
541	}
542
543	/**
544	* Reads $len from the input stream
545	*
546	* @param int $len
547	*
548	* @access private
549	* @return string
550	*/
551	private function getOpaque($len) {
552	// TODO check if it's possible to do it other way
553	// fread stops reading because the following condition is true (from php.net):
554	// if the stream is read buffered and it does not represent a plain file,
555	// at most one read of up to a number of bytes equal to the chunk size
556	// (usually 8192) is made; depending on the previously buffered data,
557	// the size of the returned data may be larger than the chunk size.
558
559	// using only return fread it will return only a part of stream if chunk is smaller
560	// than $len. Read from stream in a loop until the $len is reached.
561	$d = "";
562	$l = 0;
563	while (1) {
564	$l = (($len - strlen($d)) > 8192) ? 8192 : ($len - strlen($d));
565	if ($l > 0) {
566	$data = fread($this->in, $l);
567
568	// Stream ends prematurely on instable connections and big mails
569	if ($data === false \|\| feof($this->in))
570	throw new HTTPReturnCodeException(sprintf("WBXMLDecoder->getOpaque() connection unavailable while trying to read %d bytes from stream. Aborting after %d bytes read.", $len, strlen($d)), HTTP_CODE_500, null, LOGLEVEL_WARN);
571	else
572	$d .= $data;
573	}
574	if (strlen($d) >= $len) break;
575	}
576	return $d;
577	}
578
579	/**
580	* Reads one byte from the input stream
581	*
582	* @access private
583	* @return int
584	*/
585	private function getByte() {
586	$ch = fread($this->in, 1);
587	if(strlen($ch) > 0)
588	return ord($ch);
589	else
590	return;
591	}
592
593	/**
594	* Reads string length from the input stream
595	*
596	* @access private
597	* @return
598	*/
599	private function getMBUInt() {
600	$uint = 0;
601
602	while(1) {
603	$byte = $this->getByte();
604
605	$uint \|= $byte & 0x7f;
606
607	if($byte & 0x80)
608	$uint = $uint << 7;
609	else
610	break;
611	}
612
613	return $uint;
614	}
615
616	/**
617	* Reads string table from the input stream
618	*
619	* @access private
620	* @return int
621	*/
622	private function getStringTable() {
623	$stringtable = "";
624
625	$length = $this->getMBUInt();
626	if($length > 0)
627	$stringtable = fread($this->in, $length);
628
629	return $stringtable;
630	}
631
632	/**
633	* Returns the mapping for a specified codepage and id
634	*
635	* @param $cp codepage
636	* @param $id
637	*
638	* @access public
639	* @return string
640	*/
641	private function getMapping($cp, $id) {
642	if(!isset($this->dtd["codes"][$cp]) \|\| !isset($this->dtd["codes"][$cp][$id]))
643	return false;
644	else {
645	if(isset($this->dtd["namespaces"][$cp])) {
646	return $this->dtd["namespaces"][$cp] . ":" . $this->dtd["codes"][$cp][$id];
647	} else
648	return $this->dtd["codes"][$cp][$id];
649	}
650	}
651
652	/**
653	* Reads one byte from the input stream
654	*
655	* @access private
656	* @return void
657	*/
658	private function readVersion() {
659	$ch = $this->getByte();
660
661	if($ch != NULL) {
662	$this->inputBuffer .= chr($ch);
663	$this->version = $ch;
664	}
665	}
666	}
667
668	?>

Note: See TracBrowser for help on using the repository browser.

Context Navigation

source: trunk/zpush/lib/wbxml/wbxmldecoder.php @ 7589

Download in other formats: