3 namespace PhpXmlRpc\Helper;
5 use PhpXmlRpc\PhpXmlRpc;
9 * Deals with parsing the XML.
10 * @see http://xmlrpc.com/spec.md
14 const RETURN_XMLRPCVALS = 'xmlrpcvals';
15 const RETURN_EPIVALS = 'epivals';
16 const RETURN_PHP = 'phpvals';
18 const ACCEPT_REQUEST = 1;
19 const ACCEPT_RESPONSE = 2;
20 const ACCEPT_VALUE = 4;
21 const ACCEPT_FAULT = 8;
23 // Used to store state during parsing.
24 // Quick explanation of components:
26 // ac - used to accumulate values
27 // stack - array with genealogy of xml elements names used to validate nesting of xmlrpc elements
28 // valuestack - array used for parsing arrays and structs
29 // lv - used to indicate "looking for a value": implements the logic to allow values with no types to be strings
31 // isf - used to indicate an xml parsing fault (3), invalid xmlrpc fault (2) or xmlrpc response fault (1)
32 // isf_reason - used for storing xmlrpc response fault string
33 // method - used to store method name
34 // params - used to store parameters in method calls
35 // pt - used to store the type of each received parameter. Useful if parameters are automatically decoded to php values
36 // rt - 'methodcall', 'methodresponse', 'value' or 'fault' (the last one used only in EPI emulation mode)
40 'valuestack' => array(),
49 public $xmlrpc_valid_parents = array(
50 'VALUE' => array('MEMBER', 'DATA', 'PARAM', 'FAULT'),
51 'BOOLEAN' => array('VALUE'),
52 'I4' => array('VALUE'),
53 'I8' => array('VALUE'),
54 'EX:I8' => array('VALUE'),
55 'INT' => array('VALUE'),
56 'STRING' => array('VALUE'),
57 'DOUBLE' => array('VALUE'),
58 'DATETIME.ISO8601' => array('VALUE'),
59 'BASE64' => array('VALUE'),
60 'MEMBER' => array('STRUCT'),
61 'NAME' => array('MEMBER'),
62 'DATA' => array('ARRAY'),
63 'ARRAY' => array('VALUE'),
64 'STRUCT' => array('VALUE'),
65 'PARAM' => array('PARAMS'),
66 'METHODNAME' => array('METHODCALL'),
67 'PARAMS' => array('METHODCALL', 'METHODRESPONSE'),
68 'FAULT' => array('METHODRESPONSE'),
69 'NIL' => array('VALUE'), // only used when extension activated
70 'EX:NIL' => array('VALUE'), // only used when extension activated
73 /** @var array $parsing_options */
74 protected $parsing_options = array();
75 /** @var int $accept self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE by default */
76 protected $accept = 3;
77 /** @var int $maxChunkLength 4 MB by default. Any value below 10MB should be good */
78 protected $maxChunkLength = 4194304;
81 * @param array $options passed to the xml parser
83 public function __construct(array $options = array())
85 $this->parsing_options = $options;
90 * @param string $returnType
91 * @param int $accept a bit-combination of self::ACCEPT_REQUEST, self::ACCEPT_RESPONSE, self::ACCEPT_VALUE
94 public function parse($data, $returnType = self::RETURN_XMLRPCVALS, $accept = 3)
99 'valuestack' => array(),
102 'method' => false, // so we can check later if we got a methodname or not
108 $len = strlen($data);
110 // we test for empty documents here to save on resource allocation and simply the chunked-parsing loop below
112 $this->_xh['isf'] = 3;
113 $this->_xh['isf_reason'] = 'XML error 5: empty document';
117 $parser = xml_parser_create();
119 foreach ($this->parsing_options as $key => $val) {
120 xml_parser_set_option($parser, $key, $val);
122 // always set this, in case someone tries to disable it via options...
123 xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 1);
125 xml_set_object($parser, $this);
127 switch($returnType) {
128 case self::RETURN_PHP:
129 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_fast');
131 case self::RETURN_EPIVALS:
132 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_epi');
135 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee');
138 xml_set_character_data_handler($parser, 'xmlrpc_cd');
139 xml_set_default_handler($parser, 'xmlrpc_dh');
141 $this->accept = $accept;
143 // @see ticket #70 - we have to parse big xml docks in chunks to avoid errors
144 for ($offset = 0; $offset < $len; $offset += $this->maxChunkLength) {
145 $chunk = substr($data, $offset, $this->maxChunkLength);
146 // error handling: xml not well formed
147 if (!xml_parse($parser, $chunk, $offset + $this->maxChunkLength >= $len)) {
148 $errCode = xml_get_error_code($parser);
149 $errStr = sprintf('XML error %s: %s at line %d, column %d', $errCode, xml_error_string($errCode),
150 xml_get_current_line_number($parser), xml_get_current_column_number($parser));
152 $this->_xh['isf'] = 3;
153 $this->_xh['isf_reason'] = $errStr;
158 xml_parser_free($parser);
162 * xml parser handler function for opening element tags.
164 * @param resource $parser
165 * @param string $name
167 * @param bool $acceptSingleVals DEPRECATED use the $accept parameter instead
169 public function xmlrpc_se($parser, $name, $attrs, $acceptSingleVals = false)
171 // if invalid xmlrpc already detected, skip all processing
172 if ($this->_xh['isf'] < 2) {
174 // check for correct element nesting
175 if (count($this->_xh['stack']) == 0) {
176 // top level element can only be of 2 types
177 /// @todo optimization creep: save this check into a bool variable, instead of using count() every time:
178 /// there is only a single top level element in xml anyway
180 if ($acceptSingleVals === false) {
181 $accept = $this->accept;
183 $accept = self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE | self::ACCEPT_VALUE;
185 if (($name == 'METHODCALL' && ($accept & self::ACCEPT_REQUEST)) ||
186 ($name == 'METHODRESPONSE' && ($accept & self::ACCEPT_RESPONSE)) ||
187 ($name == 'VALUE' && ($accept & self::ACCEPT_VALUE)) ||
188 ($name == 'FAULT' && ($accept & self::ACCEPT_FAULT))) {
189 $this->_xh['rt'] = strtolower($name);
191 $this->_xh['isf'] = 2;
192 $this->_xh['isf_reason'] = 'missing top level xmlrpc element. Found: ' . $name;
197 // not top level element: see if parent is OK
198 $parent = end($this->_xh['stack']);
199 if (!array_key_exists($name, $this->xmlrpc_valid_parents) || !in_array($parent, $this->xmlrpc_valid_parents[$name])) {
200 $this->_xh['isf'] = 2;
201 $this->_xh['isf_reason'] = "xmlrpc element $name cannot be child of $parent";
208 // optimize for speed switch cases: most common cases first
210 /// @todo we could check for 2 VALUE elements inside a MEMBER or PARAM element
211 $this->_xh['vt'] = 'value'; // indicator: no value found yet
212 $this->_xh['ac'] = '';
213 $this->_xh['lv'] = 1;
214 $this->_xh['php_class'] = null;
218 if (PHP_INT_SIZE === 4) {
219 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
220 $this->_xh['isf'] = 2;
221 $this->_xh['isf_reason'] = "Received i8 element but php is compiled in 32 bit mode";
225 // fall through voluntarily
231 case 'DATETIME.ISO8601':
233 if ($this->_xh['vt'] != 'value') {
234 // two data elements inside a value: an error occurred!
235 $this->_xh['isf'] = 2;
236 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
240 $this->_xh['ac'] = ''; // reset the accumulator
244 if ($this->_xh['vt'] != 'value') {
245 // two data elements inside a value: an error occurred!
246 $this->_xh['isf'] = 2;
247 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
251 // create an empty array to hold child values, and push it onto appropriate stack
253 $curVal['values'] = array();
254 $curVal['type'] = $name;
255 // check for out-of-band information to rebuild php objs
256 // and in case it is found, save it
257 if (@isset($attrs['PHP_CLASS'])) {
258 $curVal['php_class'] = $attrs['PHP_CLASS'];
260 $this->_xh['valuestack'][] = $curVal;
261 $this->_xh['vt'] = 'data'; // be prepared for a data element next
264 if ($this->_xh['vt'] != 'data') {
265 // two data elements inside a value: an error occurred!
266 $this->_xh['isf'] = 2;
267 $this->_xh['isf_reason'] = "found two data elements inside an array element";
272 case 'METHODRESPONSE':
274 // valid elements that add little to processing
278 /// @todo we could check for 2 NAME elements inside a MEMBER element
279 $this->_xh['ac'] = '';
282 $this->_xh['isf'] = 1;
285 // set member name to null, in case we do not find in the xml later on
286 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = '';
287 //$this->_xh['ac']='';
288 // Drop trough intentionally
290 // clear value type, so we can check later if no value has been passed for this param/member
291 $this->_xh['vt'] = null;
295 if (PhpXmlRpc::$xmlrpc_null_extension) {
296 if ($this->_xh['vt'] != 'value') {
297 // two data elements inside a value: an error occurred!
298 $this->_xh['isf'] = 2;
299 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
303 $this->_xh['ac'] = ''; // reset the accumulator
306 // we do not support the <NIL/> extension, so
307 // drop through intentionally
309 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
310 $this->_xh['isf'] = 2;
311 $this->_xh['isf_reason'] = "found not-xmlrpc xml element $name";
315 // Save current element name to stack, to validate nesting
316 $this->_xh['stack'][] = $name;
318 /// @todo optimization creep: move this inside the big switch() above
319 if ($name != 'VALUE') {
320 $this->_xh['lv'] = 0;
326 * xml parser handler function for opening element tags.
327 * Used in decoding xml chunks that might represent single xmlrpc values as well as requests, responses.
329 * @param resource $parser
333 public function xmlrpc_se_any($parser, $name, $attrs)
335 $this->xmlrpc_se($parser, $name, $attrs, true);
339 * xml parser handler function for close element tags.
341 * @param resource $parser
342 * @param string $name
343 * @param int $rebuildXmlrpcvals >1 for rebuilding xmlrpcvals, 0 for rebuilding php values, -1 for xmlrpc-extension compatibility
345 public function xmlrpc_ee($parser, $name, $rebuildXmlrpcvals = 1)
347 if ($this->_xh['isf'] < 2) {
348 // push this element name from stack
349 // NB: if XML validates, correct opening/closing is guaranteed and
350 // we do not have to check for $name == $currElem.
351 // we also checked for proper nesting at start of elements...
352 $currElem = array_pop($this->_xh['stack']);
356 // This if() detects if no scalar was inside <VALUE></VALUE>
357 if ($this->_xh['vt'] == 'value') {
358 $this->_xh['value'] = $this->_xh['ac'];
359 $this->_xh['vt'] = Value::$xmlrpcString;
362 if ($rebuildXmlrpcvals > 0) {
363 // build the xmlrpc val out of the data received, and substitute it
364 $temp = new Value($this->_xh['value'], $this->_xh['vt']);
365 // in case we got info about underlying php class, save it
366 // in the object we're rebuilding
367 if (isset($this->_xh['php_class'])) {
368 $temp->_php_class = $this->_xh['php_class'];
370 $this->_xh['value'] = $temp;
371 } elseif ($rebuildXmlrpcvals < 0) {
372 if ($this->_xh['vt'] == Value::$xmlrpcDateTime) {
373 $this->_xh['value'] = (object)array(
374 'xmlrpc_type' => 'datetime',
375 'scalar' => $this->_xh['value'],
376 'timestamp' => \PhpXmlRpc\Helper\Date::iso8601Decode($this->_xh['value'])
378 } elseif ($this->_xh['vt'] == Value::$xmlrpcBase64) {
379 $this->_xh['value'] = (object)array(
380 'xmlrpc_type' => 'base64',
381 'scalar' => $this->_xh['value']
385 /// @todo this should handle php-serialized objects,
386 /// since std deserializing is done by php_xmlrpc_decode,
387 /// which we will not be calling...
388 //if (isset($this->_xh['php_class'])) {
392 // check if we are inside an array or struct:
393 // if value just built is inside an array, let's move it into array on the stack
394 $vscount = count($this->_xh['valuestack']);
395 if ($vscount && $this->_xh['valuestack'][$vscount - 1]['type'] == 'ARRAY') {
396 $this->_xh['valuestack'][$vscount - 1]['values'][] = $this->_xh['value'];
406 case 'DATETIME.ISO8601':
408 $this->_xh['vt'] = strtolower($name);
409 /// @todo: optimization creep - remove the if/elseif cycle below
410 /// since the case() in which we are already did that
411 if ($name == 'STRING') {
412 $this->_xh['value'] = $this->_xh['ac'];
413 } elseif ($name == 'DATETIME.ISO8601') {
414 if (!preg_match('/^[0-9]{8}T[0-9]{2}:[0-9]{2}:[0-9]{2}$/', $this->_xh['ac'])) {
415 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in DATETIME: ' . $this->_xh['ac']);
417 $this->_xh['vt'] = Value::$xmlrpcDateTime;
418 $this->_xh['value'] = $this->_xh['ac'];
419 } elseif ($name == 'BASE64') {
420 /// @todo check for failure of base64 decoding / catch warnings
421 $this->_xh['value'] = base64_decode($this->_xh['ac']);
422 } elseif ($name == 'BOOLEAN') {
423 // special case here: we translate boolean 1 or 0 into PHP
424 // constants true or false.
425 // Strings 'true' and 'false' are accepted, even though the
426 // spec never mentions them (see eg. Blogger api docs)
427 // NB: this simple checks helps a lot sanitizing input, ie no
428 // security problems around here
429 if ($this->_xh['ac'] == '1' || strcasecmp($this->_xh['ac'], 'true') == 0) {
430 $this->_xh['value'] = true;
432 // log if receiving something strange, even though we set the value to false anyway
433 if ($this->_xh['ac'] != '0' && strcasecmp($this->_xh['ac'], 'false') != 0) {
434 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in BOOLEAN: ' . $this->_xh['ac']);
436 $this->_xh['value'] = false;
438 } elseif ($name == 'DOUBLE') {
440 // we must check that only 0123456789-.<space> are characters here
441 // NOTE: regexp could be much stricter than this...
442 if (!preg_match('/^[+-eE0123456789 \t.]+$/', $this->_xh['ac'])) {
443 /// @todo: find a better way of throwing an error than this!
444 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in DOUBLE: ' . $this->_xh['ac']);
445 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
447 // it's ok, add it on
448 $this->_xh['value'] = (double)$this->_xh['ac'];
451 // we have an I4/I8/INT
452 // we must check that only 0123456789-<space> are characters here
453 if (!preg_match('/^[+-]?[0123456789 \t]+$/', $this->_xh['ac'])) {
454 /// @todo find a better way of throwing an error than this!
455 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in INT: ' . $this->_xh['ac']);
456 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
458 // it's ok, add it on
459 $this->_xh['value'] = (int)$this->_xh['ac'];
462 $this->_xh['lv'] = 3; // indicate we've found a value
465 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = $this->_xh['ac'];
468 // add to array in the stack the last element built,
469 // unless no VALUE was found
470 if ($this->_xh['vt']) {
471 $vscount = count($this->_xh['valuestack']);
472 $this->_xh['valuestack'][$vscount - 1]['values'][$this->_xh['valuestack'][$vscount - 1]['name']] = $this->_xh['value'];
474 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside STRUCT in received xml');
478 $this->_xh['vt'] = null; // reset this to check for 2 data elements in a row - even if they're empty
482 // fetch out of stack array of values, and promote it to current value
483 $currVal = array_pop($this->_xh['valuestack']);
484 $this->_xh['value'] = $currVal['values'];
485 $this->_xh['vt'] = strtolower($name);
486 if (isset($currVal['php_class'])) {
487 $this->_xh['php_class'] = $currVal['php_class'];
491 // add to array of params the current value,
492 // unless no VALUE was found
493 if ($this->_xh['vt']) {
494 $this->_xh['params'][] = $this->_xh['value'];
495 $this->_xh['pt'][] = $this->_xh['vt'];
497 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside PARAM in received xml');
501 $this->_xh['method'] = preg_replace('/^[\n\r\t ]+/', '', $this->_xh['ac']);
505 if (PhpXmlRpc::$xmlrpc_null_extension) {
506 $this->_xh['vt'] = 'null';
507 $this->_xh['value'] = null;
508 $this->_xh['lv'] = 3;
511 // drop through intentionally if nil extension not enabled
515 case 'METHORESPONSE':
518 // End of INVALID ELEMENT!
519 // shall we add an assert here for unreachable code???
526 * Used in decoding xmlrpc requests/responses without rebuilding xmlrpc Values.
528 * @param resource $parser
529 * @param string $name
531 public function xmlrpc_ee_fast($parser, $name)
533 $this->xmlrpc_ee($parser, $name, 0);
537 * Used in decoding xmlrpc requests/responses while building xmlrpc-extension Values (plain php for all but base64 and datetime).
539 * @param resource $parser
540 * @param string $name
542 public function xmlrpc_ee_epi($parser, $name)
544 $this->xmlrpc_ee($parser, $name, -1);
548 * xml parser handler function for character data.
550 * @param resource $parser
551 * @param string $data
553 public function xmlrpc_cd($parser, $data)
555 // skip processing if xml fault already detected
556 if ($this->_xh['isf'] < 2) {
557 // "lookforvalue==3" means that we've found an entire value
558 // and should discard any further character data
559 if ($this->_xh['lv'] != 3) {
560 $this->_xh['ac'] .= $data;
566 * xml parser handler function for 'other stuff', ie. not char data or
567 * element start/end tag. In fact it only gets called on unknown entities...
572 public function xmlrpc_dh($parser, $data)
574 // skip processing if xml fault already detected
575 if ($this->_xh['isf'] < 2) {
576 if (substr($data, 0, 1) == '&' && substr($data, -1, 1) == ';') {
577 $this->_xh['ac'] .= $data;
585 * xml charset encoding guessing helper function.
586 * Tries to determine the charset encoding of an XML chunk received over HTTP.
587 * NB: according to the spec (RFC 3023), if text/xml content-type is received over HTTP without a content-type,
588 * we SHOULD assume it is strictly US-ASCII. But we try to be more tolerant of non conforming (legacy?) clients/servers,
589 * which will be most probably using UTF-8 anyway...
590 * In order of importance checks:
594 * 4. guesses using mb_detect_encoding()
596 * @param string $httpHeader the http Content-type header
597 * @param string $xmlChunk xml content buffer
598 * @param string $encodingPrefs comma separated list of character encodings to be used as default (when mb extension is enabled).
599 * This can also be set globally using PhpXmlRpc::$xmlrpc_detectencodings
600 * @return string the encoding determined. Null if it can't be determined and mbstring is enabled,
601 * PhpXmlRpc::$xmlrpc_defencoding if it can't be determined and mbstring is not enabled
603 * @todo explore usage of mb_http_input(): does it detect http headers + post data? if so, use it instead of hand-detection!!!
605 public static function guessEncoding($httpHeader = '', $xmlChunk = '', $encodingPrefs = null)
607 // discussion: see http://www.yale.edu/pclt/encoding/
608 // 1 - test if encoding is specified in HTTP HEADERS
611 // LWS: (\13\10)?( |\t)+
612 // token: (any char but excluded stuff)+
613 // quoted string: " (any char but double quotes and control chars)* "
614 // header: Content-type = ...; charset=value(; ...)*
615 // where value is of type token, no LWS allowed between 'charset' and value
616 // Note: we do not check for invalid chars in VALUE:
617 // this had better be done using pure ereg as below
618 // Note 2: we might be removing whitespace/tabs that ought to be left in if
619 // the received charset is a quoted string. But nobody uses such charset names...
621 /// @todo this test will pass if ANY header has charset specification, not only Content-Type. Fix it?
623 if (preg_match('/;\s*charset\s*=([^;]+)/i', $httpHeader, $matches)) {
624 return strtoupper(trim($matches[1], " \t\""));
627 // 2 - scan the first bytes of the data for a UTF-16 (or other) BOM pattern
628 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
629 // NOTE: actually, according to the spec, even if we find the BOM and determine
630 // an encoding, we should check if there is an encoding specified
631 // in the xml declaration, and verify if they match.
632 /// @todo implement check as described above?
633 /// @todo implement check for first bytes of string even without a BOM? (It sure looks harder than for cases WITH a BOM)
634 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
636 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
638 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
642 // 3 - test if encoding is specified in the xml declaration
644 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
645 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
646 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
647 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
648 $xmlChunk, $matches)) {
649 return strtoupper(substr($matches[2], 1, -1));
652 // 4 - if mbstring is available, let it do the guesswork
653 if (extension_loaded('mbstring')) {
654 if ($encodingPrefs == null && PhpXmlRpc::$xmlrpc_detectencodings != null) {
655 $encodingPrefs = PhpXmlRpc::$xmlrpc_detectencodings;
657 if ($encodingPrefs) {
658 $enc = mb_detect_encoding($xmlChunk, $encodingPrefs);
660 $enc = mb_detect_encoding($xmlChunk);
662 // NB: mb_detect likes to call it ascii, xml parser likes to call it US_ASCII...
663 // IANA also likes better US-ASCII, so go with it
664 if ($enc == 'ASCII') {
670 // no encoding specified: as per HTTP1.1 assume it is iso-8859-1?
671 // Both RFC 2616 (HTTP 1.1) and 1945 (HTTP 1.0) clearly state that for text/xxx content types
672 // this should be the standard. And we should be getting text/xml as request and response.
673 // BUT we have to be backward compatible with the lib, which always used UTF-8 as default...
674 return PhpXmlRpc::$xmlrpc_defencoding;
679 * Helper function: checks if an xml chunk as a charset declaration (BOM or in the xml declaration)
681 * @param string $xmlChunk
684 public static function hasEncoding($xmlChunk)
686 // scan the first bytes of the data for a UTF-16 (or other) BOM pattern
687 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
688 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
690 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
692 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
696 // test if encoding is specified in the xml declaration
698 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
699 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
700 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
701 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
702 $xmlChunk, $matches)) {