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.
162 * @param resource $parser
163 * @param string $name
165 * @param bool $acceptSingleVals DEPRECATED use the $accept parameter instead
167 public function xmlrpc_se($parser, $name, $attrs, $acceptSingleVals = false)
169 // if invalid xmlrpc already detected, skip all processing
170 if ($this->_xh['isf'] < 2) {
172 // check for correct element nesting
173 if (count($this->_xh['stack']) == 0) {
174 // top level element can only be of 2 types
175 /// @todo optimization creep: save this check into a bool variable, instead of using count() every time:
176 /// there is only a single top level element in xml anyway
178 if ($acceptSingleVals === false) {
179 $accept = $this->accept;
181 $accept = self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE | self::ACCEPT_VALUE;
183 if (($name == 'METHODCALL' && ($accept & self::ACCEPT_REQUEST)) ||
184 ($name == 'METHODRESPONSE' && ($accept & self::ACCEPT_RESPONSE)) ||
185 ($name == 'VALUE' && ($accept & self::ACCEPT_VALUE)) ||
186 ($name == 'FAULT' && ($accept & self::ACCEPT_FAULT))) {
187 $this->_xh['rt'] = strtolower($name);
189 $this->_xh['isf'] = 2;
190 $this->_xh['isf_reason'] = 'missing top level xmlrpc element. Found: ' . $name;
195 // not top level element: see if parent is OK
196 $parent = end($this->_xh['stack']);
197 if (!array_key_exists($name, $this->xmlrpc_valid_parents) || !in_array($parent, $this->xmlrpc_valid_parents[$name])) {
198 $this->_xh['isf'] = 2;
199 $this->_xh['isf_reason'] = "xmlrpc element $name cannot be child of $parent";
206 // optimize for speed switch cases: most common cases first
208 /// @todo we could check for 2 VALUE elements inside a MEMBER or PARAM element
209 $this->_xh['vt'] = 'value'; // indicator: no value found yet
210 $this->_xh['ac'] = '';
211 $this->_xh['lv'] = 1;
212 $this->_xh['php_class'] = null;
216 if (PHP_INT_SIZE === 4) {
217 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
218 $this->_xh['isf'] = 2;
219 $this->_xh['isf_reason'] = "Received i8 element but php is compiled in 32 bit mode";
223 // fall through voluntarily
229 case 'DATETIME.ISO8601':
231 if ($this->_xh['vt'] != 'value') {
232 // two data elements inside a value: an error occurred!
233 $this->_xh['isf'] = 2;
234 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
238 $this->_xh['ac'] = ''; // reset the accumulator
242 if ($this->_xh['vt'] != 'value') {
243 // two data elements inside a value: an error occurred!
244 $this->_xh['isf'] = 2;
245 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
249 // create an empty array to hold child values, and push it onto appropriate stack
251 $curVal['values'] = array();
252 $curVal['type'] = $name;
253 // check for out-of-band information to rebuild php objs
254 // and in case it is found, save it
255 if (@isset($attrs['PHP_CLASS'])) {
256 $curVal['php_class'] = $attrs['PHP_CLASS'];
258 $this->_xh['valuestack'][] = $curVal;
259 $this->_xh['vt'] = 'data'; // be prepared for a data element next
262 if ($this->_xh['vt'] != 'data') {
263 // two data elements inside a value: an error occurred!
264 $this->_xh['isf'] = 2;
265 $this->_xh['isf_reason'] = "found two data elements inside an array element";
270 case 'METHODRESPONSE':
272 // valid elements that add little to processing
276 /// @todo we could check for 2 NAME elements inside a MEMBER element
277 $this->_xh['ac'] = '';
280 $this->_xh['isf'] = 1;
283 // set member name to null, in case we do not find in the xml later on
284 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = '';
285 //$this->_xh['ac']='';
286 // Drop trough intentionally
288 // clear value type, so we can check later if no value has been passed for this param/member
289 $this->_xh['vt'] = null;
293 if (PhpXmlRpc::$xmlrpc_null_extension) {
294 if ($this->_xh['vt'] != 'value') {
295 // two data elements inside a value: an error occurred!
296 $this->_xh['isf'] = 2;
297 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
301 $this->_xh['ac'] = ''; // reset the accumulator
304 // we do not support the <NIL/> extension, so
305 // drop through intentionally
307 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
308 $this->_xh['isf'] = 2;
309 $this->_xh['isf_reason'] = "found not-xmlrpc xml element $name";
313 // Save current element name to stack, to validate nesting
314 $this->_xh['stack'][] = $name;
316 /// @todo optimization creep: move this inside the big switch() above
317 if ($name != 'VALUE') {
318 $this->_xh['lv'] = 0;
324 * xml parser handler function for opening element tags.
325 * Used in decoding xml chunks that might represent single xmlrpc values as well as requests, responses.
327 * @param resource $parser
331 public function xmlrpc_se_any($parser, $name, $attrs)
333 $this->xmlrpc_se($parser, $name, $attrs, true);
337 * xml parser handler function for close element tags.
338 * @param resource $parser
339 * @param string $name
340 * @param int $rebuildXmlrpcvals >1 for rebuilding xmlrpcvals, 0 for rebuilding php values, -1 for xmlrpc-extension compatibility
342 public function xmlrpc_ee($parser, $name, $rebuildXmlrpcvals = 1)
344 if ($this->_xh['isf'] < 2) {
345 // push this element name from stack
346 // NB: if XML validates, correct opening/closing is guaranteed and
347 // we do not have to check for $name == $currElem.
348 // we also checked for proper nesting at start of elements...
349 $currElem = array_pop($this->_xh['stack']);
353 // This if() detects if no scalar was inside <VALUE></VALUE>
354 if ($this->_xh['vt'] == 'value') {
355 $this->_xh['value'] = $this->_xh['ac'];
356 $this->_xh['vt'] = Value::$xmlrpcString;
359 if ($rebuildXmlrpcvals > 0) {
360 // build the xmlrpc val out of the data received, and substitute it
361 $temp = new Value($this->_xh['value'], $this->_xh['vt']);
362 // in case we got info about underlying php class, save it
363 // in the object we're rebuilding
364 if (isset($this->_xh['php_class'])) {
365 $temp->_php_class = $this->_xh['php_class'];
367 $this->_xh['value'] = $temp;
368 } elseif ($rebuildXmlrpcvals < 0) {
369 if ($this->_xh['vt'] == Value::$xmlrpcDateTime) {
370 $this->_xh['value'] = (object)array(
371 'xmlrpc_type' => 'datetime',
372 'scalar' => $this->_xh['value'],
373 'timestamp' => \PhpXmlRpc\Helper\Date::iso8601Decode($this->_xh['value'])
375 } elseif ($this->_xh['vt'] == Value::$xmlrpcBase64) {
376 $this->_xh['value'] = (object)array(
377 'xmlrpc_type' => 'base64',
378 'scalar' => $this->_xh['value']
382 /// @todo this should handle php-serialized objects,
383 /// since std deserializing is done by php_xmlrpc_decode,
384 /// which we will not be calling...
385 //if (isset($this->_xh['php_class'])) {
389 // check if we are inside an array or struct:
390 // if value just built is inside an array, let's move it into array on the stack
391 $vscount = count($this->_xh['valuestack']);
392 if ($vscount && $this->_xh['valuestack'][$vscount - 1]['type'] == 'ARRAY') {
393 $this->_xh['valuestack'][$vscount - 1]['values'][] = $this->_xh['value'];
403 case 'DATETIME.ISO8601':
405 $this->_xh['vt'] = strtolower($name);
406 /// @todo: optimization creep - remove the if/elseif cycle below
407 /// since the case() in which we are already did that
408 if ($name == 'STRING') {
409 $this->_xh['value'] = $this->_xh['ac'];
410 } elseif ($name == 'DATETIME.ISO8601') {
411 if (!preg_match('/^[0-9]{8}T[0-9]{2}:[0-9]{2}:[0-9]{2}$/', $this->_xh['ac'])) {
412 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in DATETIME: ' . $this->_xh['ac']);
414 $this->_xh['vt'] = Value::$xmlrpcDateTime;
415 $this->_xh['value'] = $this->_xh['ac'];
416 } elseif ($name == 'BASE64') {
417 /// @todo check for failure of base64 decoding / catch warnings
418 $this->_xh['value'] = base64_decode($this->_xh['ac']);
419 } elseif ($name == 'BOOLEAN') {
420 // special case here: we translate boolean 1 or 0 into PHP
421 // constants true or false.
422 // Strings 'true' and 'false' are accepted, even though the
423 // spec never mentions them (see eg. Blogger api docs)
424 // NB: this simple checks helps a lot sanitizing input, ie no
425 // security problems around here
426 if ($this->_xh['ac'] == '1' || strcasecmp($this->_xh['ac'], 'true') == 0) {
427 $this->_xh['value'] = true;
429 // log if receiving something strange, even though we set the value to false anyway
430 if ($this->_xh['ac'] != '0' && strcasecmp($this->_xh['ac'], 'false') != 0) {
431 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in BOOLEAN: ' . $this->_xh['ac']);
433 $this->_xh['value'] = false;
435 } elseif ($name == 'DOUBLE') {
437 // we must check that only 0123456789-.<space> are characters here
438 // NOTE: regexp could be much stricter than this...
439 if (!preg_match('/^[+-eE0123456789 \t.]+$/', $this->_xh['ac'])) {
440 /// @todo: find a better way of throwing an error than this!
441 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in DOUBLE: ' . $this->_xh['ac']);
442 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
444 // it's ok, add it on
445 $this->_xh['value'] = (double)$this->_xh['ac'];
448 // we have an I4/I8/INT
449 // we must check that only 0123456789-<space> are characters here
450 if (!preg_match('/^[+-]?[0123456789 \t]+$/', $this->_xh['ac'])) {
451 /// @todo find a better way of throwing an error than this!
452 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in INT: ' . $this->_xh['ac']);
453 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
455 // it's ok, add it on
456 $this->_xh['value'] = (int)$this->_xh['ac'];
459 $this->_xh['lv'] = 3; // indicate we've found a value
462 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = $this->_xh['ac'];
465 // add to array in the stack the last element built,
466 // unless no VALUE was found
467 if ($this->_xh['vt']) {
468 $vscount = count($this->_xh['valuestack']);
469 $this->_xh['valuestack'][$vscount - 1]['values'][$this->_xh['valuestack'][$vscount - 1]['name']] = $this->_xh['value'];
471 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside STRUCT in received xml');
475 $this->_xh['vt'] = null; // reset this to check for 2 data elements in a row - even if they're empty
479 // fetch out of stack array of values, and promote it to current value
480 $currVal = array_pop($this->_xh['valuestack']);
481 $this->_xh['value'] = $currVal['values'];
482 $this->_xh['vt'] = strtolower($name);
483 if (isset($currVal['php_class'])) {
484 $this->_xh['php_class'] = $currVal['php_class'];
488 // add to array of params the current value,
489 // unless no VALUE was found
490 if ($this->_xh['vt']) {
491 $this->_xh['params'][] = $this->_xh['value'];
492 $this->_xh['pt'][] = $this->_xh['vt'];
494 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside PARAM in received xml');
498 $this->_xh['method'] = preg_replace('/^[\n\r\t ]+/', '', $this->_xh['ac']);
502 if (PhpXmlRpc::$xmlrpc_null_extension) {
503 $this->_xh['vt'] = 'null';
504 $this->_xh['value'] = null;
505 $this->_xh['lv'] = 3;
508 // drop through intentionally if nil extension not enabled
512 case 'METHORESPONSE':
515 // End of INVALID ELEMENT!
516 // shall we add an assert here for unreachable code???
523 * Used in decoding xmlrpc requests/responses without rebuilding xmlrpc Values.
524 * @param resource $parser
525 * @param string $name
527 public function xmlrpc_ee_fast($parser, $name)
529 $this->xmlrpc_ee($parser, $name, 0);
533 * Used in decoding xmlrpc requests/responses while building xmlrpc-extension Values (plain php for all but base64 and datetime).
534 * @param resource $parser
535 * @param string $name
537 public function xmlrpc_ee_epi($parser, $name)
539 $this->xmlrpc_ee($parser, $name, -1);
543 * xml parser handler function for character data.
544 * @param resource $parser
545 * @param string $data
547 public function xmlrpc_cd($parser, $data)
549 // skip processing if xml fault already detected
550 if ($this->_xh['isf'] < 2) {
551 // "lookforvalue==3" means that we've found an entire value
552 // and should discard any further character data
553 if ($this->_xh['lv'] != 3) {
554 $this->_xh['ac'] .= $data;
560 * xml parser handler function for 'other stuff', ie. not char data or
561 * element start/end tag. In fact it only gets called on unknown entities...
565 public function xmlrpc_dh($parser, $data)
567 // skip processing if xml fault already detected
568 if ($this->_xh['isf'] < 2) {
569 if (substr($data, 0, 1) == '&' && substr($data, -1, 1) == ';') {
570 $this->_xh['ac'] .= $data;
578 * xml charset encoding guessing helper function.
579 * Tries to determine the charset encoding of an XML chunk received over HTTP.
580 * NB: according to the spec (RFC 3023), if text/xml content-type is received over HTTP without a content-type,
581 * we SHOULD assume it is strictly US-ASCII. But we try to be more tolerant of non conforming (legacy?) clients/servers,
582 * which will be most probably using UTF-8 anyway...
583 * In order of importance checks:
587 * 4. guesses using mb_detect_encoding()
589 * @param string $httpHeader the http Content-type header
590 * @param string $xmlChunk xml content buffer
591 * @param string $encodingPrefs comma separated list of character encodings to be used as default (when mb extension is enabled).
592 * This can also be set globally using PhpXmlRpc::$xmlrpc_detectencodings
593 * @return string the encoding determined. Null if it can't be determined and mbstring is enabled,
594 * PhpXmlRpc::$xmlrpc_defencoding if it can't be determined and mbstring is not enabled
596 * @todo explore usage of mb_http_input(): does it detect http headers + post data? if so, use it instead of hand-detection!!!
598 public static function guessEncoding($httpHeader = '', $xmlChunk = '', $encodingPrefs = null)
600 // discussion: see http://www.yale.edu/pclt/encoding/
601 // 1 - test if encoding is specified in HTTP HEADERS
604 // LWS: (\13\10)?( |\t)+
605 // token: (any char but excluded stuff)+
606 // quoted string: " (any char but double quotes and control chars)* "
607 // header: Content-type = ...; charset=value(; ...)*
608 // where value is of type token, no LWS allowed between 'charset' and value
609 // Note: we do not check for invalid chars in VALUE:
610 // this had better be done using pure ereg as below
611 // Note 2: we might be removing whitespace/tabs that ought to be left in if
612 // the received charset is a quoted string. But nobody uses such charset names...
614 /// @todo this test will pass if ANY header has charset specification, not only Content-Type. Fix it?
616 if (preg_match('/;\s*charset\s*=([^;]+)/i', $httpHeader, $matches)) {
617 return strtoupper(trim($matches[1], " \t\""));
620 // 2 - scan the first bytes of the data for a UTF-16 (or other) BOM pattern
621 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
622 // NOTE: actually, according to the spec, even if we find the BOM and determine
623 // an encoding, we should check if there is an encoding specified
624 // in the xml declaration, and verify if they match.
625 /// @todo implement check as described above?
626 /// @todo implement check for first bytes of string even without a BOM? (It sure looks harder than for cases WITH a BOM)
627 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
629 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
631 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
635 // 3 - test if encoding is specified in the xml declaration
637 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
638 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
639 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
640 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
641 $xmlChunk, $matches)) {
642 return strtoupper(substr($matches[2], 1, -1));
645 // 4 - if mbstring is available, let it do the guesswork
646 if (extension_loaded('mbstring')) {
647 if ($encodingPrefs == null && PhpXmlRpc::$xmlrpc_detectencodings != null) {
648 $encodingPrefs = PhpXmlRpc::$xmlrpc_detectencodings;
650 if ($encodingPrefs) {
651 $enc = mb_detect_encoding($xmlChunk, $encodingPrefs);
653 $enc = mb_detect_encoding($xmlChunk);
655 // NB: mb_detect likes to call it ascii, xml parser likes to call it US_ASCII...
656 // IANA also likes better US-ASCII, so go with it
657 if ($enc == 'ASCII') {
663 // no encoding specified: as per HTTP1.1 assume it is iso-8859-1?
664 // Both RFC 2616 (HTTP 1.1) and 1945 (HTTP 1.0) clearly state that for text/xxx content types
665 // this should be the standard. And we should be getting text/xml as request and response.
666 // BUT we have to be backward compatible with the lib, which always used UTF-8 as default...
667 return PhpXmlRpc::$xmlrpc_defencoding;
672 * Helper function: checks if an xml chunk as a charset declaration (BOM or in the xml declaration)
674 * @param string $xmlChunk
677 public static function hasEncoding($xmlChunk)
679 // scan the first bytes of the data for a UTF-16 (or other) BOM pattern
680 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
681 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
683 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
685 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
689 // test if encoding is specified in the xml declaration
691 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
692 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
693 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
694 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
695 $xmlChunk, $matches)) {