3 namespace PhpXmlRpc\Helper;
5 use PhpXmlRpc\PhpXmlRpc;
9 * Deals with parsing the XML.
13 const RETURN_XMLRPCVALS = 'xmlrpcvals';
14 const RETURN_EPIVALS = 'epivals';
15 const RETURN_PHP = 'phpvals';
17 const ACCEPT_REQUEST = 1;
18 const ACCEPT_RESPONSE = 2;
19 const ACCEPT_VALUE = 4;
20 const ACCEPT_FAULT = 8;
22 // Used to store state during parsing.
23 // Quick explanation of components:
25 // ac - used to accumulate values
26 // stack - array with genealogy of xml elements names used to validate nesting of xmlrpc elements
27 // valuestack - array used for parsing arrays and structs
28 // lv - used to indicate "looking for a value": implements the logic to allow values with no types to be strings
30 // isf - used to indicate an xml parsing fault (3), invalid xmlrpc fault (2) or xmlrpc response fault (1)
31 // isf_reason - used for storing xmlrpc response fault string
32 // method - used to store method name
33 // params - used to store parameters in method calls
34 // pt - used to store the type of each received parameter. Useful if parameters are automatically decoded to php values
35 // rt - 'methodcall', 'methodresponse', 'value' or 'fault' (the last one used only in EPI emulation mode)
39 'valuestack' => array(),
48 public $xmlrpc_valid_parents = array(
49 'VALUE' => array('MEMBER', 'DATA', 'PARAM', 'FAULT'),
50 'BOOLEAN' => array('VALUE'),
51 'I4' => array('VALUE'),
52 'I8' => array('VALUE'),
53 'EX:I8' => array('VALUE'),
54 'INT' => array('VALUE'),
55 'STRING' => array('VALUE'),
56 'DOUBLE' => array('VALUE'),
57 'DATETIME.ISO8601' => array('VALUE'),
58 'BASE64' => array('VALUE'),
59 'MEMBER' => array('STRUCT'),
60 'NAME' => array('MEMBER'),
61 'DATA' => array('ARRAY'),
62 'ARRAY' => array('VALUE'),
63 'STRUCT' => array('VALUE'),
64 'PARAM' => array('PARAMS'),
65 'METHODNAME' => array('METHODCALL'),
66 'PARAMS' => array('METHODCALL', 'METHODRESPONSE'),
67 'FAULT' => array('METHODRESPONSE'),
68 'NIL' => array('VALUE'), // only used when extension activated
69 'EX:NIL' => array('VALUE'), // only used when extension activated
72 /** @var array $parsing_options */
73 protected $parsing_options = array();
74 /** @var int $accept self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE by default */
75 protected $accept = 3;
76 /** @var int $maxChunkLength 4 MB by default. Any value below 10MB should be good */
77 protected $maxChunkLength = 4194304;
80 * @param array $options passed to the xml parser
82 public function __construct(array $options = array())
84 $this->parsing_options = $options;
89 * @param string $returnType
90 * @param int $accept a bit-combination of self::ACCEPT_REQUEST, self::ACCEPT_RESPONSE, self::ACCEPT_VALUE
93 public function parse($data, $returnType = self::RETURN_XMLRPCVALS, $accept = 3)
98 'valuestack' => array(),
101 'method' => false, // so we can check later if we got a methodname or not
107 $len = strlen($data);
109 // we test for empty documents here to save on resource allocation and simply the chunked-parsing loop below
111 $this->_xh['isf'] = 3;
112 $this->_xh['isf_reason'] = 'XML error 5: empty document';
116 $parser = xml_parser_create();
118 foreach ($this->parsing_options as $key => $val) {
119 xml_parser_set_option($parser, $key, $val);
121 // always set this, in case someone tries to disable it via options...
122 xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 1);
124 xml_set_object($parser, $this);
126 switch($returnType) {
127 case self::RETURN_PHP:
128 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_fast');
130 case self::RETURN_EPIVALS:
131 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_epi');
134 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee');
137 xml_set_character_data_handler($parser, 'xmlrpc_cd');
138 xml_set_default_handler($parser, 'xmlrpc_dh');
140 $this->accept = $accept;
142 // @see ticket #70 - we have to parse big xml docks in chunks to avoid errors
143 for ($offset = 0; $offset < $len; $offset += $this->maxChunkLength) {
144 $chunk = substr($data, $offset, $this->maxChunkLength);
145 // error handling: xml not well formed
146 if (!xml_parse($parser, $chunk, $offset + $this->maxChunkLength >= $len)) {
147 $errCode = xml_get_error_code($parser);
148 $errStr = sprintf('XML error %s: %s at line %d, column %d', $errCode, xml_error_string($errCode),
149 xml_get_current_line_number($parser), xml_get_current_column_number($parser));
151 $this->_xh['isf'] = 3;
152 $this->_xh['isf_reason'] = $errStr;
157 xml_parser_free($parser);
161 * xml parser handler function for opening element tags.
163 * @param resource $parser
164 * @param string $name
166 * @param bool $acceptSingleVals DEPRECATED use the $accept parameter instead
168 public function xmlrpc_se($parser, $name, $attrs, $acceptSingleVals = false)
170 // if invalid xmlrpc already detected, skip all processing
171 if ($this->_xh['isf'] < 2) {
173 // check for correct element nesting
174 if (count($this->_xh['stack']) == 0) {
175 // top level element can only be of 2 types
176 /// @todo optimization creep: save this check into a bool variable, instead of using count() every time:
177 /// there is only a single top level element in xml anyway
179 if ($acceptSingleVals === false) {
180 $accept = $this->accept;
182 $accept = self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE | self::ACCEPT_VALUE;
184 if (($name == 'METHODCALL' && ($accept & self::ACCEPT_REQUEST)) ||
185 ($name == 'METHODRESPONSE' && ($accept & self::ACCEPT_RESPONSE)) ||
186 ($name == 'VALUE' && ($accept & self::ACCEPT_VALUE)) ||
187 ($name == 'FAULT' && ($accept & self::ACCEPT_FAULT))) {
188 $this->_xh['rt'] = strtolower($name);
190 $this->_xh['isf'] = 2;
191 $this->_xh['isf_reason'] = 'missing top level xmlrpc element. Found: ' . $name;
196 // not top level element: see if parent is OK
197 $parent = end($this->_xh['stack']);
198 if (!array_key_exists($name, $this->xmlrpc_valid_parents) || !in_array($parent, $this->xmlrpc_valid_parents[$name])) {
199 $this->_xh['isf'] = 2;
200 $this->_xh['isf_reason'] = "xmlrpc element $name cannot be child of $parent";
207 // optimize for speed switch cases: most common cases first
209 /// @todo we could check for 2 VALUE elements inside a MEMBER or PARAM element
210 $this->_xh['vt'] = 'value'; // indicator: no value found yet
211 $this->_xh['ac'] = '';
212 $this->_xh['lv'] = 1;
213 $this->_xh['php_class'] = null;
217 if (PHP_INT_SIZE === 4) {
218 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
219 $this->_xh['isf'] = 2;
220 $this->_xh['isf_reason'] = "Received i8 element but php is compiled in 32 bit mode";
224 // fall through voluntarily
230 case 'DATETIME.ISO8601':
232 if ($this->_xh['vt'] != 'value') {
233 // two data elements inside a value: an error occurred!
234 $this->_xh['isf'] = 2;
235 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
239 $this->_xh['ac'] = ''; // reset the accumulator
243 if ($this->_xh['vt'] != 'value') {
244 // two data elements inside a value: an error occurred!
245 $this->_xh['isf'] = 2;
246 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
250 // create an empty array to hold child values, and push it onto appropriate stack
252 $curVal['values'] = array();
253 $curVal['type'] = $name;
254 // check for out-of-band information to rebuild php objs
255 // and in case it is found, save it
256 if (@isset($attrs['PHP_CLASS'])) {
257 $curVal['php_class'] = $attrs['PHP_CLASS'];
259 $this->_xh['valuestack'][] = $curVal;
260 $this->_xh['vt'] = 'data'; // be prepared for a data element next
263 if ($this->_xh['vt'] != 'data') {
264 // two data elements inside a value: an error occurred!
265 $this->_xh['isf'] = 2;
266 $this->_xh['isf_reason'] = "found two data elements inside an array element";
271 case 'METHODRESPONSE':
273 // valid elements that add little to processing
277 /// @todo we could check for 2 NAME elements inside a MEMBER element
278 $this->_xh['ac'] = '';
281 $this->_xh['isf'] = 1;
284 // set member name to null, in case we do not find in the xml later on
285 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = '';
286 //$this->_xh['ac']='';
287 // Drop trough intentionally
289 // clear value type, so we can check later if no value has been passed for this param/member
290 $this->_xh['vt'] = null;
294 if (PhpXmlRpc::$xmlrpc_null_extension) {
295 if ($this->_xh['vt'] != 'value') {
296 // two data elements inside a value: an error occurred!
297 $this->_xh['isf'] = 2;
298 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
302 $this->_xh['ac'] = ''; // reset the accumulator
305 // we do not support the <NIL/> extension, so
306 // drop through intentionally
308 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
309 $this->_xh['isf'] = 2;
310 $this->_xh['isf_reason'] = "found not-xmlrpc xml element $name";
314 // Save current element name to stack, to validate nesting
315 $this->_xh['stack'][] = $name;
317 /// @todo optimization creep: move this inside the big switch() above
318 if ($name != 'VALUE') {
319 $this->_xh['lv'] = 0;
325 * xml parser handler function for opening element tags.
326 * Used in decoding xml chunks that might represent single xmlrpc values as well as requests, responses.
328 * @param resource $parser
332 public function xmlrpc_se_any($parser, $name, $attrs)
334 $this->xmlrpc_se($parser, $name, $attrs, true);
338 * xml parser handler function for close element tags.
340 * @param resource $parser
341 * @param string $name
342 * @param int $rebuildXmlrpcvals >1 for rebuilding xmlrpcvals, 0 for rebuilding php values, -1 for xmlrpc-extension compatibility
344 public function xmlrpc_ee($parser, $name, $rebuildXmlrpcvals = 1)
346 if ($this->_xh['isf'] < 2) {
347 // push this element name from stack
348 // NB: if XML validates, correct opening/closing is guaranteed and
349 // we do not have to check for $name == $currElem.
350 // we also checked for proper nesting at start of elements...
351 $currElem = array_pop($this->_xh['stack']);
355 // This if() detects if no scalar was inside <VALUE></VALUE>
356 if ($this->_xh['vt'] == 'value') {
357 $this->_xh['value'] = $this->_xh['ac'];
358 $this->_xh['vt'] = Value::$xmlrpcString;
361 if ($rebuildXmlrpcvals > 0) {
362 // build the xmlrpc val out of the data received, and substitute it
363 $temp = new Value($this->_xh['value'], $this->_xh['vt']);
364 // in case we got info about underlying php class, save it
365 // in the object we're rebuilding
366 if (isset($this->_xh['php_class'])) {
367 $temp->_php_class = $this->_xh['php_class'];
369 $this->_xh['value'] = $temp;
370 } elseif ($rebuildXmlrpcvals < 0) {
371 if ($this->_xh['vt'] == Value::$xmlrpcDateTime) {
372 $this->_xh['value'] = (object)array(
373 'xmlrpc_type' => 'datetime',
374 'scalar' => $this->_xh['value'],
375 'timestamp' => \PhpXmlRpc\Helper\Date::iso8601Decode($this->_xh['value'])
377 } elseif ($this->_xh['vt'] == Value::$xmlrpcBase64) {
378 $this->_xh['value'] = (object)array(
379 'xmlrpc_type' => 'base64',
380 'scalar' => $this->_xh['value']
384 /// @todo this should handle php-serialized objects,
385 /// since std deserializing is done by php_xmlrpc_decode,
386 /// which we will not be calling...
387 //if (isset($this->_xh['php_class'])) {
391 // check if we are inside an array or struct:
392 // if value just built is inside an array, let's move it into array on the stack
393 $vscount = count($this->_xh['valuestack']);
394 if ($vscount && $this->_xh['valuestack'][$vscount - 1]['type'] == 'ARRAY') {
395 $this->_xh['valuestack'][$vscount - 1]['values'][] = $this->_xh['value'];
405 case 'DATETIME.ISO8601':
407 $this->_xh['vt'] = strtolower($name);
408 /// @todo: optimization creep - remove the if/elseif cycle below
409 /// since the case() in which we are already did that
410 if ($name == 'STRING') {
411 $this->_xh['value'] = $this->_xh['ac'];
412 } elseif ($name == 'DATETIME.ISO8601') {
413 if (!preg_match('/^[0-9]{8}T[0-9]{2}:[0-9]{2}:[0-9]{2}$/', $this->_xh['ac'])) {
414 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in DATETIME: ' . $this->_xh['ac']);
416 $this->_xh['vt'] = Value::$xmlrpcDateTime;
417 $this->_xh['value'] = $this->_xh['ac'];
418 } elseif ($name == 'BASE64') {
419 /// @todo check for failure of base64 decoding / catch warnings
420 $this->_xh['value'] = base64_decode($this->_xh['ac']);
421 } elseif ($name == 'BOOLEAN') {
422 // special case here: we translate boolean 1 or 0 into PHP
423 // constants true or false.
424 // Strings 'true' and 'false' are accepted, even though the
425 // spec never mentions them (see eg. Blogger api docs)
426 // NB: this simple checks helps a lot sanitizing input, ie no
427 // security problems around here
428 if ($this->_xh['ac'] == '1' || strcasecmp($this->_xh['ac'], 'true') == 0) {
429 $this->_xh['value'] = true;
431 // log if receiving something strange, even though we set the value to false anyway
432 if ($this->_xh['ac'] != '0' && strcasecmp($this->_xh['ac'], 'false') != 0) {
433 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in BOOLEAN: ' . $this->_xh['ac']);
435 $this->_xh['value'] = false;
437 } elseif ($name == 'DOUBLE') {
439 // we must check that only 0123456789-.<space> are characters here
440 // NOTE: regexp could be much stricter than this...
441 if (!preg_match('/^[+-eE0123456789 \t.]+$/', $this->_xh['ac'])) {
442 /// @todo: find a better way of throwing an error than this!
443 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in DOUBLE: ' . $this->_xh['ac']);
444 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
446 // it's ok, add it on
447 $this->_xh['value'] = (double)$this->_xh['ac'];
450 // we have an I4/I8/INT
451 // we must check that only 0123456789-<space> are characters here
452 if (!preg_match('/^[+-]?[0123456789 \t]+$/', $this->_xh['ac'])) {
453 /// @todo find a better way of throwing an error than this!
454 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in INT: ' . $this->_xh['ac']);
455 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
457 // it's ok, add it on
458 $this->_xh['value'] = (int)$this->_xh['ac'];
461 $this->_xh['lv'] = 3; // indicate we've found a value
464 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = $this->_xh['ac'];
467 // add to array in the stack the last element built,
468 // unless no VALUE was found
469 if ($this->_xh['vt']) {
470 $vscount = count($this->_xh['valuestack']);
471 $this->_xh['valuestack'][$vscount - 1]['values'][$this->_xh['valuestack'][$vscount - 1]['name']] = $this->_xh['value'];
473 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside STRUCT in received xml');
477 $this->_xh['vt'] = null; // reset this to check for 2 data elements in a row - even if they're empty
481 // fetch out of stack array of values, and promote it to current value
482 $currVal = array_pop($this->_xh['valuestack']);
483 $this->_xh['value'] = $currVal['values'];
484 $this->_xh['vt'] = strtolower($name);
485 if (isset($currVal['php_class'])) {
486 $this->_xh['php_class'] = $currVal['php_class'];
490 // add to array of params the current value,
491 // unless no VALUE was found
492 if ($this->_xh['vt']) {
493 $this->_xh['params'][] = $this->_xh['value'];
494 $this->_xh['pt'][] = $this->_xh['vt'];
496 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside PARAM in received xml');
500 $this->_xh['method'] = preg_replace('/^[\n\r\t ]+/', '', $this->_xh['ac']);
504 if (PhpXmlRpc::$xmlrpc_null_extension) {
505 $this->_xh['vt'] = 'null';
506 $this->_xh['value'] = null;
507 $this->_xh['lv'] = 3;
510 // drop through intentionally if nil extension not enabled
514 case 'METHORESPONSE':
517 // End of INVALID ELEMENT!
518 // shall we add an assert here for unreachable code???
525 * Used in decoding xmlrpc requests/responses without rebuilding xmlrpc Values.
527 * @param resource $parser
528 * @param string $name
530 public function xmlrpc_ee_fast($parser, $name)
532 $this->xmlrpc_ee($parser, $name, 0);
536 * Used in decoding xmlrpc requests/responses while building xmlrpc-extension Values (plain php for all but base64 and datetime).
538 * @param resource $parser
539 * @param string $name
541 public function xmlrpc_ee_epi($parser, $name)
543 $this->xmlrpc_ee($parser, $name, -1);
547 * xml parser handler function for character data.
549 * @param resource $parser
550 * @param string $data
552 public function xmlrpc_cd($parser, $data)
554 // skip processing if xml fault already detected
555 if ($this->_xh['isf'] < 2) {
556 // "lookforvalue==3" means that we've found an entire value
557 // and should discard any further character data
558 if ($this->_xh['lv'] != 3) {
559 $this->_xh['ac'] .= $data;
565 * xml parser handler function for 'other stuff', ie. not char data or
566 * element start/end tag. In fact it only gets called on unknown entities...
571 public function xmlrpc_dh($parser, $data)
573 // skip processing if xml fault already detected
574 if ($this->_xh['isf'] < 2) {
575 if (substr($data, 0, 1) == '&' && substr($data, -1, 1) == ';') {
576 $this->_xh['ac'] .= $data;
584 * xml charset encoding guessing helper function.
585 * Tries to determine the charset encoding of an XML chunk received over HTTP.
586 * NB: according to the spec (RFC 3023), if text/xml content-type is received over HTTP without a content-type,
587 * we SHOULD assume it is strictly US-ASCII. But we try to be more tolerant of non conforming (legacy?) clients/servers,
588 * which will be most probably using UTF-8 anyway...
589 * In order of importance checks:
593 * 4. guesses using mb_detect_encoding()
595 * @param string $httpHeader the http Content-type header
596 * @param string $xmlChunk xml content buffer
597 * @param string $encodingPrefs comma separated list of character encodings to be used as default (when mb extension is enabled).
598 * This can also be set globally using PhpXmlRpc::$xmlrpc_detectencodings
599 * @return string the encoding determined. Null if it can't be determined and mbstring is enabled,
600 * PhpXmlRpc::$xmlrpc_defencoding if it can't be determined and mbstring is not enabled
602 * @todo explore usage of mb_http_input(): does it detect http headers + post data? if so, use it instead of hand-detection!!!
604 public static function guessEncoding($httpHeader = '', $xmlChunk = '', $encodingPrefs = null)
606 // discussion: see http://www.yale.edu/pclt/encoding/
607 // 1 - test if encoding is specified in HTTP HEADERS
610 // LWS: (\13\10)?( |\t)+
611 // token: (any char but excluded stuff)+
612 // quoted string: " (any char but double quotes and control chars)* "
613 // header: Content-type = ...; charset=value(; ...)*
614 // where value is of type token, no LWS allowed between 'charset' and value
615 // Note: we do not check for invalid chars in VALUE:
616 // this had better be done using pure ereg as below
617 // Note 2: we might be removing whitespace/tabs that ought to be left in if
618 // the received charset is a quoted string. But nobody uses such charset names...
620 /// @todo this test will pass if ANY header has charset specification, not only Content-Type. Fix it?
622 if (preg_match('/;\s*charset\s*=([^;]+)/i', $httpHeader, $matches)) {
623 return strtoupper(trim($matches[1], " \t\""));
626 // 2 - scan the first bytes of the data for a UTF-16 (or other) BOM pattern
627 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
628 // NOTE: actually, according to the spec, even if we find the BOM and determine
629 // an encoding, we should check if there is an encoding specified
630 // in the xml declaration, and verify if they match.
631 /// @todo implement check as described above?
632 /// @todo implement check for first bytes of string even without a BOM? (It sure looks harder than for cases WITH a BOM)
633 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
635 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
637 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
641 // 3 - test if encoding is specified in the xml declaration
643 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
644 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
645 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
646 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
647 $xmlChunk, $matches)) {
648 return strtoupper(substr($matches[2], 1, -1));
651 // 4 - if mbstring is available, let it do the guesswork
652 if (extension_loaded('mbstring')) {
653 if ($encodingPrefs == null && PhpXmlRpc::$xmlrpc_detectencodings != null) {
654 $encodingPrefs = PhpXmlRpc::$xmlrpc_detectencodings;
656 if ($encodingPrefs) {
657 $enc = mb_detect_encoding($xmlChunk, $encodingPrefs);
659 $enc = mb_detect_encoding($xmlChunk);
661 // NB: mb_detect likes to call it ascii, xml parser likes to call it US_ASCII...
662 // IANA also likes better US-ASCII, so go with it
663 if ($enc == 'ASCII') {
669 // no encoding specified: as per HTTP1.1 assume it is iso-8859-1?
670 // Both RFC 2616 (HTTP 1.1) and 1945 (HTTP 1.0) clearly state that for text/xxx content types
671 // this should be the standard. And we should be getting text/xml as request and response.
672 // BUT we have to be backward compatible with the lib, which always used UTF-8 as default...
673 return PhpXmlRpc::$xmlrpc_defencoding;
678 * Helper function: checks if an xml chunk as a charset declaration (BOM or in the xml declaration)
680 * @param string $xmlChunk
683 public static function hasEncoding($xmlChunk)
685 // scan the first bytes of the data for a UTF-16 (or other) BOM pattern
686 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
687 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
689 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
691 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
695 // test if encoding is specified in the xml declaration
697 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
698 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
699 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
700 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
701 $xmlChunk, $matches)) {