3 namespace PhpXmlRpc\Helper;
5 use PhpXmlRpc\PhpXmlRpc;
9 * Deals with parsing the XML.
10 * @see http://xmlrpc.com/spec.md
12 * @todo implement an interface to allow for alternative implementations
13 * - make access to $_xh protected, return more high-level data structures
14 * - add parseRequest, parseResponse, parseValue methods
18 const RETURN_XMLRPCVALS = 'xmlrpcvals';
19 const RETURN_EPIVALS = 'epivals';
20 const RETURN_PHP = 'phpvals';
22 const ACCEPT_REQUEST = 1;
23 const ACCEPT_RESPONSE = 2;
24 const ACCEPT_VALUE = 4;
25 const ACCEPT_FAULT = 8;
27 // Used to store state during parsing and to pass parsing results to callers.
28 // Quick explanation of components:
30 // ac - used to accumulate values
31 // stack - array with genealogy of xml elements names used to validate nesting of xmlrpc elements
32 // valuestack - array used for parsing arrays and structs
33 // lv - used to indicate "looking for a value": implements the logic to allow values with no types to be strings
35 // isf - used to indicate an xml parsing fault (3), invalid xmlrpc fault (2) or xmlrpc response fault (1)
36 // isf_reason - used for storing xmlrpc response fault string
37 // value - used to store the value in responses
38 // method - used to store method name in requests
39 // params - used to store parameters in requests
40 // pt - used to store the type of each received parameter. Useful if parameters are automatically decoded to php values
41 // rt - 'methodcall', 'methodresponse', 'value' or 'fault' (the last one used only in EPI emulation mode)
45 'valuestack' => array(),
55 public $xmlrpc_valid_parents = array(
56 'VALUE' => array('MEMBER', 'DATA', 'PARAM', 'FAULT'),
57 'BOOLEAN' => array('VALUE'),
58 'I4' => array('VALUE'),
59 'I8' => array('VALUE'),
60 'EX:I8' => array('VALUE'),
61 'INT' => array('VALUE'),
62 'STRING' => array('VALUE'),
63 'DOUBLE' => array('VALUE'),
64 'DATETIME.ISO8601' => array('VALUE'),
65 'BASE64' => array('VALUE'),
66 'MEMBER' => array('STRUCT'),
67 'NAME' => array('MEMBER'),
68 'DATA' => array('ARRAY'),
69 'ARRAY' => array('VALUE'),
70 'STRUCT' => array('VALUE'),
71 'PARAM' => array('PARAMS'),
72 'METHODNAME' => array('METHODCALL'),
73 'PARAMS' => array('METHODCALL', 'METHODRESPONSE'),
74 'FAULT' => array('METHODRESPONSE'),
75 'NIL' => array('VALUE'), // only used when extension activated
76 'EX:NIL' => array('VALUE'), // only used when extension activated
79 /** @var array $parsing_options */
80 protected $parsing_options = array();
81 /** @var int $accept self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE by default */
82 protected $accept = 3;
83 /** @var int $maxChunkLength 4 MB by default. Any value below 10MB should be good */
84 protected $maxChunkLength = 4194304;
87 * @param array $options passed to the xml parser
89 public function __construct(array $options = array())
91 $this->parsing_options = $options;
96 * @param string $returnType
97 * @param int $accept a bit-combination of self::ACCEPT_REQUEST, self::ACCEPT_RESPONSE, self::ACCEPT_VALUE
99 public function parse($data, $returnType = self::RETURN_XMLRPCVALS, $accept = 3)
104 'valuestack' => array(),
108 'method' => false, // so we can check later if we got a methodname or not
114 $len = strlen($data);
116 // we test for empty documents here to save on resource allocation and simply the chunked-parsing loop below
118 $this->_xh['isf'] = 3;
119 $this->_xh['isf_reason'] = 'XML error 5: empty document';
123 $parser = xml_parser_create();
125 foreach ($this->parsing_options as $key => $val) {
126 xml_parser_set_option($parser, $key, $val);
128 // always set this, in case someone tries to disable it via options...
129 xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 1);
131 xml_set_object($parser, $this);
133 switch($returnType) {
134 case self::RETURN_PHP:
135 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_fast');
137 case self::RETURN_EPIVALS:
138 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_epi');
141 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee');
144 xml_set_character_data_handler($parser, 'xmlrpc_cd');
145 xml_set_default_handler($parser, 'xmlrpc_dh');
147 $this->accept = $accept;
149 // @see ticket #70 - we have to parse big xml docks in chunks to avoid errors
150 for ($offset = 0; $offset < $len; $offset += $this->maxChunkLength) {
151 $chunk = substr($data, $offset, $this->maxChunkLength);
152 // error handling: xml not well formed
153 if (!xml_parse($parser, $chunk, $offset + $this->maxChunkLength >= $len)) {
154 $errCode = xml_get_error_code($parser);
155 $errStr = sprintf('XML error %s: %s at line %d, column %d', $errCode, xml_error_string($errCode),
156 xml_get_current_line_number($parser), xml_get_current_column_number($parser));
158 $this->_xh['isf'] = 3;
159 $this->_xh['isf_reason'] = $errStr;
164 xml_parser_free($parser);
168 * xml parser handler function for opening element tags.
170 * @param resource $parser
171 * @param string $name
173 * @param bool $acceptSingleVals DEPRECATED use the $accept parameter instead
175 public function xmlrpc_se($parser, $name, $attrs, $acceptSingleVals = false)
177 // if invalid xmlrpc already detected, skip all processing
178 if ($this->_xh['isf'] < 2) {
180 // check for correct element nesting
181 if (count($this->_xh['stack']) == 0) {
182 // top level element can only be of 2 types
183 /// @todo optimization creep: save this check into a bool variable, instead of using count() every time:
184 /// there is only a single top level element in xml anyway
186 if ($acceptSingleVals === false) {
187 $accept = $this->accept;
189 $accept = self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE | self::ACCEPT_VALUE;
191 if (($name == 'METHODCALL' && ($accept & self::ACCEPT_REQUEST)) ||
192 ($name == 'METHODRESPONSE' && ($accept & self::ACCEPT_RESPONSE)) ||
193 ($name == 'VALUE' && ($accept & self::ACCEPT_VALUE)) ||
194 ($name == 'FAULT' && ($accept & self::ACCEPT_FAULT))) {
195 $this->_xh['rt'] = strtolower($name);
197 $this->_xh['isf'] = 2;
198 $this->_xh['isf_reason'] = 'missing top level xmlrpc element. Found: ' . $name;
203 // not top level element: see if parent is OK
204 $parent = end($this->_xh['stack']);
205 if (!array_key_exists($name, $this->xmlrpc_valid_parents) || !in_array($parent, $this->xmlrpc_valid_parents[$name])) {
206 $this->_xh['isf'] = 2;
207 $this->_xh['isf_reason'] = "xmlrpc element $name cannot be child of $parent";
214 // optimize for speed switch cases: most common cases first
216 /// @todo we could check for 2 VALUE elements inside a MEMBER or PARAM element
217 $this->_xh['vt'] = 'value'; // indicator: no value found yet
218 $this->_xh['ac'] = '';
219 $this->_xh['lv'] = 1;
220 $this->_xh['php_class'] = null;
224 if (PHP_INT_SIZE === 4) {
225 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
226 $this->_xh['isf'] = 2;
227 $this->_xh['isf_reason'] = "Received i8 element but php is compiled in 32 bit mode";
231 // fall through voluntarily
237 case 'DATETIME.ISO8601':
239 if ($this->_xh['vt'] != 'value') {
240 // two data elements inside a value: an error occurred!
241 $this->_xh['isf'] = 2;
242 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
246 $this->_xh['ac'] = ''; // reset the accumulator
250 if ($this->_xh['vt'] != 'value') {
251 // two data elements inside a value: an error occurred!
252 $this->_xh['isf'] = 2;
253 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
257 // create an empty array to hold child values, and push it onto appropriate stack
259 $curVal['values'] = array();
260 $curVal['type'] = $name;
261 // check for out-of-band information to rebuild php objs
262 // and in case it is found, save it
263 if (@isset($attrs['PHP_CLASS'])) {
264 $curVal['php_class'] = $attrs['PHP_CLASS'];
266 $this->_xh['valuestack'][] = $curVal;
267 $this->_xh['vt'] = 'data'; // be prepared for a data element next
270 if ($this->_xh['vt'] != 'data') {
271 // two data elements inside a value: an error occurred!
272 $this->_xh['isf'] = 2;
273 $this->_xh['isf_reason'] = "found two data elements inside an array element";
278 case 'METHODRESPONSE':
280 // valid elements that add little to processing
284 /// @todo we could check for 2 NAME elements inside a MEMBER element
285 $this->_xh['ac'] = '';
288 $this->_xh['isf'] = 1;
291 // set member name to null, in case we do not find in the xml later on
292 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = '';
293 //$this->_xh['ac']='';
294 // Drop trough intentionally
296 // clear value type, so we can check later if no value has been passed for this param/member
297 $this->_xh['vt'] = null;
301 if (PhpXmlRpc::$xmlrpc_null_extension) {
302 if ($this->_xh['vt'] != 'value') {
303 // two data elements inside a value: an error occurred!
304 $this->_xh['isf'] = 2;
305 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
309 $this->_xh['ac'] = ''; // reset the accumulator
312 // we do not support the <NIL/> extension, so
313 // drop through intentionally
315 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
316 $this->_xh['isf'] = 2;
317 $this->_xh['isf_reason'] = "found not-xmlrpc xml element $name";
321 // Save current element name to stack, to validate nesting
322 $this->_xh['stack'][] = $name;
324 /// @todo optimization creep: move this inside the big switch() above
325 if ($name != 'VALUE') {
326 $this->_xh['lv'] = 0;
332 * xml parser handler function for opening element tags.
333 * Used in decoding xml chunks that might represent single xmlrpc values as well as requests, responses.
335 * @param resource $parser
339 public function xmlrpc_se_any($parser, $name, $attrs)
341 $this->xmlrpc_se($parser, $name, $attrs, true);
345 * xml parser handler function for close element tags.
347 * @param resource $parser
348 * @param string $name
349 * @param int $rebuildXmlrpcvals >1 for rebuilding xmlrpcvals, 0 for rebuilding php values, -1 for xmlrpc-extension compatibility
351 public function xmlrpc_ee($parser, $name, $rebuildXmlrpcvals = 1)
353 if ($this->_xh['isf'] < 2) {
354 // push this element name from stack
355 // NB: if XML validates, correct opening/closing is guaranteed and
356 // we do not have to check for $name == $currElem.
357 // we also checked for proper nesting at start of elements...
358 $currElem = array_pop($this->_xh['stack']);
362 // This if() detects if no scalar was inside <VALUE></VALUE>
363 if ($this->_xh['vt'] == 'value') {
364 $this->_xh['value'] = $this->_xh['ac'];
365 $this->_xh['vt'] = Value::$xmlrpcString;
368 if ($rebuildXmlrpcvals > 0) {
369 // build the xmlrpc val out of the data received, and substitute it
370 $temp = new Value($this->_xh['value'], $this->_xh['vt']);
371 // in case we got info about underlying php class, save it
372 // in the object we're rebuilding
373 if (isset($this->_xh['php_class'])) {
374 $temp->_php_class = $this->_xh['php_class'];
376 $this->_xh['value'] = $temp;
377 } elseif ($rebuildXmlrpcvals < 0) {
378 if ($this->_xh['vt'] == Value::$xmlrpcDateTime) {
379 $this->_xh['value'] = (object)array(
380 'xmlrpc_type' => 'datetime',
381 'scalar' => $this->_xh['value'],
382 'timestamp' => \PhpXmlRpc\Helper\Date::iso8601Decode($this->_xh['value'])
384 } elseif ($this->_xh['vt'] == Value::$xmlrpcBase64) {
385 $this->_xh['value'] = (object)array(
386 'xmlrpc_type' => 'base64',
387 'scalar' => $this->_xh['value']
391 /// @todo this should handle php-serialized objects,
392 /// since std deserializing is done by php_xmlrpc_decode,
393 /// which we will not be calling...
394 //if (isset($this->_xh['php_class'])) {
398 // check if we are inside an array or struct:
399 // if value just built is inside an array, let's move it into array on the stack
400 $vscount = count($this->_xh['valuestack']);
401 if ($vscount && $this->_xh['valuestack'][$vscount - 1]['type'] == 'ARRAY') {
402 $this->_xh['valuestack'][$vscount - 1]['values'][] = $this->_xh['value'];
412 case 'DATETIME.ISO8601':
414 $this->_xh['vt'] = strtolower($name);
415 /// @todo: optimization creep - remove the if/elseif cycle below
416 /// since the case() in which we are already did that
417 if ($name == 'STRING') {
418 $this->_xh['value'] = $this->_xh['ac'];
419 } elseif ($name == 'DATETIME.ISO8601') {
420 if (!preg_match('/^[0-9]{8}T[0-9]{2}:[0-9]{2}:[0-9]{2}$/', $this->_xh['ac'])) {
421 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in DATETIME: ' . $this->_xh['ac']);
423 $this->_xh['vt'] = Value::$xmlrpcDateTime;
424 $this->_xh['value'] = $this->_xh['ac'];
425 } elseif ($name == 'BASE64') {
426 /// @todo check for failure of base64 decoding / catch warnings
427 $this->_xh['value'] = base64_decode($this->_xh['ac']);
428 } elseif ($name == 'BOOLEAN') {
429 // special case here: we translate boolean 1 or 0 into PHP
430 // constants true or false.
431 // Strings 'true' and 'false' are accepted, even though the
432 // spec never mentions them (see eg. Blogger api docs)
433 // NB: this simple checks helps a lot sanitizing input, ie no
434 // security problems around here
435 if ($this->_xh['ac'] == '1' || strcasecmp($this->_xh['ac'], 'true') == 0) {
436 $this->_xh['value'] = true;
438 // log if receiving something strange, even though we set the value to false anyway
439 if ($this->_xh['ac'] != '0' && strcasecmp($this->_xh['ac'], 'false') != 0) {
440 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in BOOLEAN: ' . $this->_xh['ac']);
442 $this->_xh['value'] = false;
444 } elseif ($name == 'DOUBLE') {
446 // we must check that only 0123456789-.<space> are characters here
447 // NOTE: regexp could be much stricter than this...
448 if (!preg_match('/^[+-eE0123456789 \t.]+$/', $this->_xh['ac'])) {
449 /// @todo: find a better way of throwing an error than this!
450 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in DOUBLE: ' . $this->_xh['ac']);
451 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
453 // it's ok, add it on
454 $this->_xh['value'] = (double)$this->_xh['ac'];
457 // we have an I4/I8/INT
458 // we must check that only 0123456789-<space> are characters here
459 if (!preg_match('/^[+-]?[0123456789 \t]+$/', $this->_xh['ac'])) {
460 /// @todo find a better way of throwing an error than this!
461 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in INT: ' . $this->_xh['ac']);
462 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
464 // it's ok, add it on
465 $this->_xh['value'] = (int)$this->_xh['ac'];
468 $this->_xh['lv'] = 3; // indicate we've found a value
471 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = $this->_xh['ac'];
474 // add to array in the stack the last element built,
475 // unless no VALUE was found
476 if ($this->_xh['vt']) {
477 $vscount = count($this->_xh['valuestack']);
478 $this->_xh['valuestack'][$vscount - 1]['values'][$this->_xh['valuestack'][$vscount - 1]['name']] = $this->_xh['value'];
480 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside STRUCT in received xml');
484 $this->_xh['vt'] = null; // reset this to check for 2 data elements in a row - even if they're empty
488 // fetch out of stack array of values, and promote it to current value
489 $currVal = array_pop($this->_xh['valuestack']);
490 $this->_xh['value'] = $currVal['values'];
491 $this->_xh['vt'] = strtolower($name);
492 if (isset($currVal['php_class'])) {
493 $this->_xh['php_class'] = $currVal['php_class'];
497 // add to array of params the current value,
498 // unless no VALUE was found
499 if ($this->_xh['vt']) {
500 $this->_xh['params'][] = $this->_xh['value'];
501 $this->_xh['pt'][] = $this->_xh['vt'];
503 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside PARAM in received xml');
507 $this->_xh['method'] = preg_replace('/^[\n\r\t ]+/', '', $this->_xh['ac']);
511 if (PhpXmlRpc::$xmlrpc_null_extension) {
512 $this->_xh['vt'] = 'null';
513 $this->_xh['value'] = null;
514 $this->_xh['lv'] = 3;
517 // drop through intentionally if nil extension not enabled
521 case 'METHORESPONSE':
524 // End of INVALID ELEMENT!
525 // shall we add an assert here for unreachable code???
532 * Used in decoding xmlrpc requests/responses without rebuilding xmlrpc Values.
534 * @param resource $parser
535 * @param string $name
537 public function xmlrpc_ee_fast($parser, $name)
539 $this->xmlrpc_ee($parser, $name, 0);
543 * Used in decoding xmlrpc requests/responses while building xmlrpc-extension Values (plain php for all but base64 and datetime).
545 * @param resource $parser
546 * @param string $name
548 public function xmlrpc_ee_epi($parser, $name)
550 $this->xmlrpc_ee($parser, $name, -1);
554 * xml parser handler function for character data.
556 * @param resource $parser
557 * @param string $data
559 public function xmlrpc_cd($parser, $data)
561 // skip processing if xml fault already detected
562 if ($this->_xh['isf'] < 2) {
563 // "lookforvalue==3" means that we've found an entire value
564 // and should discard any further character data
565 if ($this->_xh['lv'] != 3) {
566 $this->_xh['ac'] .= $data;
572 * xml parser handler function for 'other stuff', ie. not char data or
573 * element start/end tag. In fact it only gets called on unknown entities...
578 public function xmlrpc_dh($parser, $data)
580 // skip processing if xml fault already detected
581 if ($this->_xh['isf'] < 2) {
582 if (substr($data, 0, 1) == '&' && substr($data, -1, 1) == ';') {
583 $this->_xh['ac'] .= $data;
591 * xml charset encoding guessing helper function.
592 * Tries to determine the charset encoding of an XML chunk received over HTTP.
593 * NB: according to the spec (RFC 3023), if text/xml content-type is received over HTTP without a content-type,
594 * we SHOULD assume it is strictly US-ASCII. But we try to be more tolerant of non conforming (legacy?) clients/servers,
595 * which will be most probably using UTF-8 anyway...
596 * In order of importance checks:
600 * 4. guesses using mb_detect_encoding()
602 * @param string $httpHeader the http Content-type header
603 * @param string $xmlChunk xml content buffer
604 * @param string $encodingPrefs comma separated list of character encodings to be used as default (when mb extension is enabled).
605 * This can also be set globally using PhpXmlRpc::$xmlrpc_detectencodings
606 * @return string the encoding determined. Null if it can't be determined and mbstring is enabled,
607 * PhpXmlRpc::$xmlrpc_defencoding if it can't be determined and mbstring is not enabled
609 * @todo explore usage of mb_http_input(): does it detect http headers + post data? if so, use it instead of hand-detection!!!
611 public static function guessEncoding($httpHeader = '', $xmlChunk = '', $encodingPrefs = null)
613 // discussion: see http://www.yale.edu/pclt/encoding/
614 // 1 - test if encoding is specified in HTTP HEADERS
617 // LWS: (\13\10)?( |\t)+
618 // token: (any char but excluded stuff)+
619 // quoted string: " (any char but double quotes and control chars)* "
620 // header: Content-type = ...; charset=value(; ...)*
621 // where value is of type token, no LWS allowed between 'charset' and value
622 // Note: we do not check for invalid chars in VALUE:
623 // this had better be done using pure ereg as below
624 // Note 2: we might be removing whitespace/tabs that ought to be left in if
625 // the received charset is a quoted string. But nobody uses such charset names...
627 /// @todo this test will pass if ANY header has charset specification, not only Content-Type. Fix it?
629 if (preg_match('/;\s*charset\s*=([^;]+)/i', $httpHeader, $matches)) {
630 return strtoupper(trim($matches[1], " \t\""));
633 // 2 - scan the first bytes of the data for a UTF-16 (or other) BOM pattern
634 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
635 // NOTE: actually, according to the spec, even if we find the BOM and determine
636 // an encoding, we should check if there is an encoding specified
637 // in the xml declaration, and verify if they match.
638 /// @todo implement check as described above?
639 /// @todo implement check for first bytes of string even without a BOM? (It sure looks harder than for cases WITH a BOM)
640 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
642 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
644 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
648 // 3 - test if encoding is specified in the xml declaration
650 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
651 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
652 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
653 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
654 $xmlChunk, $matches)) {
655 return strtoupper(substr($matches[2], 1, -1));
658 // 4 - if mbstring is available, let it do the guesswork
659 if (extension_loaded('mbstring')) {
660 if ($encodingPrefs == null && PhpXmlRpc::$xmlrpc_detectencodings != null) {
661 $encodingPrefs = PhpXmlRpc::$xmlrpc_detectencodings;
663 if ($encodingPrefs) {
664 $enc = mb_detect_encoding($xmlChunk, $encodingPrefs);
666 $enc = mb_detect_encoding($xmlChunk);
668 // NB: mb_detect likes to call it ascii, xml parser likes to call it US_ASCII...
669 // IANA also likes better US-ASCII, so go with it
670 if ($enc == 'ASCII') {
676 // no encoding specified: as per HTTP1.1 assume it is iso-8859-1?
677 // Both RFC 2616 (HTTP 1.1) and 1945 (HTTP 1.0) clearly state that for text/xxx content types
678 // this should be the standard. And we should be getting text/xml as request and response.
679 // BUT we have to be backward compatible with the lib, which always used UTF-8 as default...
680 return PhpXmlRpc::$xmlrpc_defencoding;
685 * Helper function: checks if an xml chunk as a charset declaration (BOM or in the xml declaration)
687 * @param string $xmlChunk
690 public static function hasEncoding($xmlChunk)
692 // scan the first bytes of the data for a UTF-16 (or other) BOM pattern
693 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
694 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
696 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
698 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
702 // test if encoding is specified in the xml declaration
704 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
705 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
706 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
707 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
708 $xmlChunk, $matches)) {