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
15 * @todo if iconv() or mb_string() are available, we could allow to convert the received xml to a custom charset encoding
16 * while parsing, which is faster than doing it later by going over the rebuilt data structure
20 const RETURN_XMLRPCVALS = 'xmlrpcvals';
21 const RETURN_EPIVALS = 'epivals';
22 const RETURN_PHP = 'phpvals';
24 const ACCEPT_REQUEST = 1;
25 const ACCEPT_RESPONSE = 2;
26 const ACCEPT_VALUE = 4;
27 const ACCEPT_FAULT = 8;
29 // Used to store state during parsing and to pass parsing results to callers.
30 // Quick explanation of components:
32 // ac - used to accumulate values
33 // stack - array with genealogy of xml elements names used to validate nesting of xmlrpc elements
34 // valuestack - array used for parsing arrays and structs
35 // lv - used to indicate "looking for a value": implements the logic to allow values with no types to be strings
37 // isf - used to indicate an xml parsing fault (3), invalid xmlrpc fault (2) or xmlrpc response fault (1)
38 // isf_reason - used for storing xmlrpc response fault string
39 // value - used to store the value in responses
40 // method - used to store method name in requests
41 // params - used to store parameters in requests
42 // pt - used to store the type of each received parameter. Useful if parameters are automatically decoded to php values
43 // rt - 'methodcall', 'methodresponse', 'value' or 'fault' (the last one used only in EPI emulation mode)
47 'valuestack' => array(),
57 public $xmlrpc_valid_parents = array(
58 'VALUE' => array('MEMBER', 'DATA', 'PARAM', 'FAULT'),
59 'BOOLEAN' => array('VALUE'),
60 'I4' => array('VALUE'),
61 'I8' => array('VALUE'),
62 'EX:I8' => array('VALUE'),
63 'INT' => array('VALUE'),
64 'STRING' => array('VALUE'),
65 'DOUBLE' => array('VALUE'),
66 'DATETIME.ISO8601' => array('VALUE'),
67 'BASE64' => array('VALUE'),
68 'MEMBER' => array('STRUCT'),
69 'NAME' => array('MEMBER'),
70 'DATA' => array('ARRAY'),
71 'ARRAY' => array('VALUE'),
72 'STRUCT' => array('VALUE'),
73 'PARAM' => array('PARAMS'),
74 'METHODNAME' => array('METHODCALL'),
75 'PARAMS' => array('METHODCALL', 'METHODRESPONSE'),
76 'FAULT' => array('METHODRESPONSE'),
77 'NIL' => array('VALUE'), // only used when extension activated
78 'EX:NIL' => array('VALUE'), // only used when extension activated
81 /** @var array $parsing_options */
82 protected $parsing_options = array();
83 /** @var int $accept self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE by default */
84 protected $accept = 3;
85 /** @var int $maxChunkLength 4 MB by default. Any value below 10MB should be good */
86 protected $maxChunkLength = 4194304;
89 * @param array $options passed to the xml parser
91 public function __construct(array $options = array())
93 $this->parsing_options = $options;
98 * @param string $returnType
99 * @param int $accept a bit-combination of self::ACCEPT_REQUEST, self::ACCEPT_RESPONSE, self::ACCEPT_VALUE
101 public function parse($data, $returnType = self::RETURN_XMLRPCVALS, $accept = 3)
106 'valuestack' => array(),
110 'method' => false, // so we can check later if we got a methodname or not
116 $len = strlen($data);
118 // we test for empty documents here to save on resource allocation and simply the chunked-parsing loop below
120 $this->_xh['isf'] = 3;
121 $this->_xh['isf_reason'] = 'XML error 5: empty document';
125 $parser = xml_parser_create();
127 foreach ($this->parsing_options as $key => $val) {
128 xml_parser_set_option($parser, $key, $val);
130 // always set this, in case someone tries to disable it via options...
131 xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 1);
133 xml_set_object($parser, $this);
135 switch($returnType) {
136 case self::RETURN_PHP:
137 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_fast');
139 case self::RETURN_EPIVALS:
140 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee_epi');
143 xml_set_element_handler($parser, 'xmlrpc_se', 'xmlrpc_ee');
146 xml_set_character_data_handler($parser, 'xmlrpc_cd');
147 xml_set_default_handler($parser, 'xmlrpc_dh');
149 $this->accept = $accept;
151 // @see ticket #70 - we have to parse big xml docks in chunks to avoid errors
152 for ($offset = 0; $offset < $len; $offset += $this->maxChunkLength) {
153 $chunk = substr($data, $offset, $this->maxChunkLength);
154 // error handling: xml not well formed
155 if (!xml_parse($parser, $chunk, $offset + $this->maxChunkLength >= $len)) {
156 $errCode = xml_get_error_code($parser);
157 $errStr = sprintf('XML error %s: %s at line %d, column %d', $errCode, xml_error_string($errCode),
158 xml_get_current_line_number($parser), xml_get_current_column_number($parser));
160 $this->_xh['isf'] = 3;
161 $this->_xh['isf_reason'] = $errStr;
166 xml_parser_free($parser);
170 * xml parser handler function for opening element tags.
172 * @param resource $parser
173 * @param string $name
175 * @param bool $acceptSingleVals DEPRECATED use the $accept parameter instead
177 public function xmlrpc_se($parser, $name, $attrs, $acceptSingleVals = false)
179 // if invalid xmlrpc already detected, skip all processing
180 if ($this->_xh['isf'] < 2) {
182 // check for correct element nesting
183 if (count($this->_xh['stack']) == 0) {
184 // top level element can only be of 2 types
185 /// @todo optimization creep: save this check into a bool variable, instead of using count() every time:
186 /// there is only a single top level element in xml anyway
188 if ($acceptSingleVals === false) {
189 $accept = $this->accept;
191 $accept = self::ACCEPT_REQUEST | self::ACCEPT_RESPONSE | self::ACCEPT_VALUE;
193 if (($name == 'METHODCALL' && ($accept & self::ACCEPT_REQUEST)) ||
194 ($name == 'METHODRESPONSE' && ($accept & self::ACCEPT_RESPONSE)) ||
195 ($name == 'VALUE' && ($accept & self::ACCEPT_VALUE)) ||
196 ($name == 'FAULT' && ($accept & self::ACCEPT_FAULT))) {
197 $this->_xh['rt'] = strtolower($name);
199 $this->_xh['isf'] = 2;
200 $this->_xh['isf_reason'] = 'missing top level xmlrpc element. Found: ' . $name;
205 // not top level element: see if parent is OK
206 $parent = end($this->_xh['stack']);
207 if (!array_key_exists($name, $this->xmlrpc_valid_parents) || !in_array($parent, $this->xmlrpc_valid_parents[$name])) {
208 $this->_xh['isf'] = 2;
209 $this->_xh['isf_reason'] = "xmlrpc element $name cannot be child of $parent";
216 // optimize for speed switch cases: most common cases first
218 /// @todo we could check for 2 VALUE elements inside a MEMBER or PARAM element
219 $this->_xh['vt'] = 'value'; // indicator: no value found yet
220 $this->_xh['ac'] = '';
221 $this->_xh['lv'] = 1;
222 $this->_xh['php_class'] = null;
226 if (PHP_INT_SIZE === 4) {
227 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
228 $this->_xh['isf'] = 2;
229 $this->_xh['isf_reason'] = "Received i8 element but php is compiled in 32 bit mode";
233 // fall through voluntarily
239 case 'DATETIME.ISO8601':
241 if ($this->_xh['vt'] != 'value') {
242 // two data elements inside a value: an error occurred!
243 $this->_xh['isf'] = 2;
244 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
248 $this->_xh['ac'] = ''; // reset the accumulator
252 if ($this->_xh['vt'] != 'value') {
253 // two data elements inside a value: an error occurred!
254 $this->_xh['isf'] = 2;
255 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
259 // create an empty array to hold child values, and push it onto appropriate stack
261 $curVal['values'] = array();
262 $curVal['type'] = $name;
263 // check for out-of-band information to rebuild php objs
264 // and in case it is found, save it
265 if (@isset($attrs['PHP_CLASS'])) {
266 $curVal['php_class'] = $attrs['PHP_CLASS'];
268 $this->_xh['valuestack'][] = $curVal;
269 $this->_xh['vt'] = 'data'; // be prepared for a data element next
272 if ($this->_xh['vt'] != 'data') {
273 // two data elements inside a value: an error occurred!
274 $this->_xh['isf'] = 2;
275 $this->_xh['isf_reason'] = "found two data elements inside an array element";
280 case 'METHODRESPONSE':
282 // valid elements that add little to processing
286 /// @todo we could check for 2 NAME elements inside a MEMBER element
287 $this->_xh['ac'] = '';
290 $this->_xh['isf'] = 1;
293 // set member name to null, in case we do not find in the xml later on
294 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = '';
295 //$this->_xh['ac']='';
296 // Drop trough intentionally
298 // clear value type, so we can check later if no value has been passed for this param/member
299 $this->_xh['vt'] = null;
303 if (PhpXmlRpc::$xmlrpc_null_extension) {
304 if ($this->_xh['vt'] != 'value') {
305 // two data elements inside a value: an error occurred!
306 $this->_xh['isf'] = 2;
307 $this->_xh['isf_reason'] = "$name element following a {$this->_xh['vt']} element inside a single value";
311 $this->_xh['ac'] = ''; // reset the accumulator
314 // we do not support the <NIL/> extension, so
315 // drop through intentionally
317 // INVALID ELEMENT: RAISE ISF so that it is later recognized!!!
318 $this->_xh['isf'] = 2;
319 $this->_xh['isf_reason'] = "found not-xmlrpc xml element $name";
323 // Save current element name to stack, to validate nesting
324 $this->_xh['stack'][] = $name;
326 /// @todo optimization creep: move this inside the big switch() above
327 if ($name != 'VALUE') {
328 $this->_xh['lv'] = 0;
334 * xml parser handler function for opening element tags.
335 * Used in decoding xml chunks that might represent single xmlrpc values as well as requests, responses.
337 * @param resource $parser
341 public function xmlrpc_se_any($parser, $name, $attrs)
343 $this->xmlrpc_se($parser, $name, $attrs, true);
347 * xml parser handler function for close element tags.
349 * @param resource $parser
350 * @param string $name
351 * @param int $rebuildXmlrpcvals >1 for rebuilding xmlrpcvals, 0 for rebuilding php values, -1 for xmlrpc-extension compatibility
353 public function xmlrpc_ee($parser, $name, $rebuildXmlrpcvals = 1)
355 if ($this->_xh['isf'] < 2) {
356 // push this element name from stack
357 // NB: if XML validates, correct opening/closing is guaranteed and
358 // we do not have to check for $name == $currElem.
359 // we also checked for proper nesting at start of elements...
360 $currElem = array_pop($this->_xh['stack']);
364 // This if() detects if no scalar was inside <VALUE></VALUE>
365 if ($this->_xh['vt'] == 'value') {
366 $this->_xh['value'] = $this->_xh['ac'];
367 $this->_xh['vt'] = Value::$xmlrpcString;
370 if ($rebuildXmlrpcvals > 0) {
371 // build the xmlrpc val out of the data received, and substitute it
372 $temp = new Value($this->_xh['value'], $this->_xh['vt']);
373 // in case we got info about underlying php class, save it
374 // in the object we're rebuilding
375 if (isset($this->_xh['php_class'])) {
376 $temp->_php_class = $this->_xh['php_class'];
378 $this->_xh['value'] = $temp;
379 } elseif ($rebuildXmlrpcvals < 0) {
380 if ($this->_xh['vt'] == Value::$xmlrpcDateTime) {
381 $this->_xh['value'] = (object)array(
382 'xmlrpc_type' => 'datetime',
383 'scalar' => $this->_xh['value'],
384 'timestamp' => \PhpXmlRpc\Helper\Date::iso8601Decode($this->_xh['value'])
386 } elseif ($this->_xh['vt'] == Value::$xmlrpcBase64) {
387 $this->_xh['value'] = (object)array(
388 'xmlrpc_type' => 'base64',
389 'scalar' => $this->_xh['value']
393 /// @todo this should handle php-serialized objects,
394 /// since std deserializing is done by php_xmlrpc_decode,
395 /// which we will not be calling...
396 //if (isset($this->_xh['php_class'])) {
400 // check if we are inside an array or struct:
401 // if value just built is inside an array, let's move it into array on the stack
402 $vscount = count($this->_xh['valuestack']);
403 if ($vscount && $this->_xh['valuestack'][$vscount - 1]['type'] == 'ARRAY') {
404 $this->_xh['valuestack'][$vscount - 1]['values'][] = $this->_xh['value'];
414 case 'DATETIME.ISO8601':
416 $this->_xh['vt'] = strtolower($name);
417 /// @todo: optimization creep - remove the if/elseif cycle below
418 /// since the case() in which we are already did that
419 if ($name == 'STRING') {
420 $this->_xh['value'] = $this->_xh['ac'];
421 } elseif ($name == 'DATETIME.ISO8601') {
422 if (!preg_match('/^[0-9]{8}T[0-9]{2}:[0-9]{2}:[0-9]{2}$/', $this->_xh['ac'])) {
423 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in DATETIME: ' . $this->_xh['ac']);
425 $this->_xh['vt'] = Value::$xmlrpcDateTime;
426 $this->_xh['value'] = $this->_xh['ac'];
427 } elseif ($name == 'BASE64') {
428 /// @todo check for failure of base64 decoding / catch warnings
429 $this->_xh['value'] = base64_decode($this->_xh['ac']);
430 } elseif ($name == 'BOOLEAN') {
431 // special case here: we translate boolean 1 or 0 into PHP
432 // constants true or false.
433 // Strings 'true' and 'false' are accepted, even though the
434 // spec never mentions them (see eg. Blogger api docs)
435 // NB: this simple checks helps a lot sanitizing input, ie no
436 // security problems around here
437 if ($this->_xh['ac'] == '1' || strcasecmp($this->_xh['ac'], 'true') == 0) {
438 $this->_xh['value'] = true;
440 // log if receiving something strange, even though we set the value to false anyway
441 if ($this->_xh['ac'] != '0' && strcasecmp($this->_xh['ac'], 'false') != 0) {
442 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': invalid value received in BOOLEAN: ' . $this->_xh['ac']);
444 $this->_xh['value'] = false;
446 } elseif ($name == 'DOUBLE') {
448 // we must check that only 0123456789-.<space> are characters here
449 // NOTE: regexp could be much stricter than this...
450 if (!preg_match('/^[+-eE0123456789 \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 DOUBLE: ' . $this->_xh['ac']);
453 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
455 // it's ok, add it on
456 $this->_xh['value'] = (double)$this->_xh['ac'];
459 // we have an I4/I8/INT
460 // we must check that only 0123456789-<space> are characters here
461 if (!preg_match('/^[+-]?[0123456789 \t]+$/', $this->_xh['ac'])) {
462 /// @todo find a better way of throwing an error than this!
463 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': non numeric value received in INT: ' . $this->_xh['ac']);
464 $this->_xh['value'] = 'ERROR_NON_NUMERIC_FOUND';
466 // it's ok, add it on
467 $this->_xh['value'] = (int)$this->_xh['ac'];
470 $this->_xh['lv'] = 3; // indicate we've found a value
473 $this->_xh['valuestack'][count($this->_xh['valuestack']) - 1]['name'] = $this->_xh['ac'];
476 // add to array in the stack the last element built,
477 // unless no VALUE was found
478 if ($this->_xh['vt']) {
479 $vscount = count($this->_xh['valuestack']);
480 $this->_xh['valuestack'][$vscount - 1]['values'][$this->_xh['valuestack'][$vscount - 1]['name']] = $this->_xh['value'];
482 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside STRUCT in received xml');
486 $this->_xh['vt'] = null; // reset this to check for 2 data elements in a row - even if they're empty
490 // fetch out of stack array of values, and promote it to current value
491 $currVal = array_pop($this->_xh['valuestack']);
492 $this->_xh['value'] = $currVal['values'];
493 $this->_xh['vt'] = strtolower($name);
494 if (isset($currVal['php_class'])) {
495 $this->_xh['php_class'] = $currVal['php_class'];
499 // add to array of params the current value,
500 // unless no VALUE was found
501 if ($this->_xh['vt']) {
502 $this->_xh['params'][] = $this->_xh['value'];
503 $this->_xh['pt'][] = $this->_xh['vt'];
505 Logger::instance()->errorLog('XML-RPC: ' . __METHOD__ . ': missing VALUE inside PARAM in received xml');
509 $this->_xh['method'] = preg_replace('/^[\n\r\t ]+/', '', $this->_xh['ac']);
513 if (PhpXmlRpc::$xmlrpc_null_extension) {
514 $this->_xh['vt'] = 'null';
515 $this->_xh['value'] = null;
516 $this->_xh['lv'] = 3;
519 // drop through intentionally if nil extension not enabled
523 case 'METHORESPONSE':
526 // End of INVALID ELEMENT!
527 // shall we add an assert here for unreachable code???
534 * Used in decoding xmlrpc requests/responses without rebuilding xmlrpc Values.
536 * @param resource $parser
537 * @param string $name
539 public function xmlrpc_ee_fast($parser, $name)
541 $this->xmlrpc_ee($parser, $name, 0);
545 * Used in decoding xmlrpc requests/responses while building xmlrpc-extension Values (plain php for all but base64 and datetime).
547 * @param resource $parser
548 * @param string $name
550 public function xmlrpc_ee_epi($parser, $name)
552 $this->xmlrpc_ee($parser, $name, -1);
556 * xml parser handler function for character data.
558 * @param resource $parser
559 * @param string $data
561 public function xmlrpc_cd($parser, $data)
563 // skip processing if xml fault already detected
564 if ($this->_xh['isf'] < 2) {
565 // "lookforvalue==3" means that we've found an entire value
566 // and should discard any further character data
567 if ($this->_xh['lv'] != 3) {
568 $this->_xh['ac'] .= $data;
574 * xml parser handler function for 'other stuff', ie. not char data or
575 * element start/end tag. In fact it only gets called on unknown entities...
580 public function xmlrpc_dh($parser, $data)
582 // skip processing if xml fault already detected
583 if ($this->_xh['isf'] < 2) {
584 if (substr($data, 0, 1) == '&' && substr($data, -1, 1) == ';') {
585 $this->_xh['ac'] .= $data;
593 * xml charset encoding guessing helper function.
594 * Tries to determine the charset encoding of an XML chunk received over HTTP.
595 * NB: according to the spec (RFC 3023), if text/xml content-type is received over HTTP without a content-type,
596 * we SHOULD assume it is strictly US-ASCII. But we try to be more tolerant of non conforming (legacy?) clients/servers,
597 * which will be most probably using UTF-8 anyway...
598 * In order of importance checks:
602 * 4. guesses using mb_detect_encoding()
604 * @param string $httpHeader the http Content-type header
605 * @param string $xmlChunk xml content buffer
606 * @param string $encodingPrefs comma separated list of character encodings to be used as default (when mb extension is enabled).
607 * This can also be set globally using PhpXmlRpc::$xmlrpc_detectencodings
608 * @return string the encoding determined. Null if it can't be determined and mbstring is enabled,
609 * PhpXmlRpc::$xmlrpc_defencoding if it can't be determined and mbstring is not enabled
611 * @todo explore usage of mb_http_input(): does it detect http headers + post data? if so, use it instead of hand-detection!!!
613 public static function guessEncoding($httpHeader = '', $xmlChunk = '', $encodingPrefs = null)
615 // discussion: see http://www.yale.edu/pclt/encoding/
616 // 1 - test if encoding is specified in HTTP HEADERS
619 // LWS: (\13\10)?( |\t)+
620 // token: (any char but excluded stuff)+
621 // quoted string: " (any char but double quotes and control chars)* "
622 // header: Content-type = ...; charset=value(; ...)*
623 // where value is of type token, no LWS allowed between 'charset' and value
624 // Note: we do not check for invalid chars in VALUE:
625 // this had better be done using pure ereg as below
626 // Note 2: we might be removing whitespace/tabs that ought to be left in if
627 // the received charset is a quoted string. But nobody uses such charset names...
629 /// @todo this test will pass if ANY header has charset specification, not only Content-Type. Fix it?
631 if (preg_match('/;\s*charset\s*=([^;]+)/i', $httpHeader, $matches)) {
632 return strtoupper(trim($matches[1], " \t\""));
635 // 2 - scan the first bytes of the data for a UTF-16 (or other) BOM pattern
636 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
637 // NOTE: actually, according to the spec, even if we find the BOM and determine
638 // an encoding, we should check if there is an encoding specified
639 // in the xml declaration, and verify if they match.
640 /// @todo implement check as described above?
641 /// @todo implement check for first bytes of string even without a BOM? (It sure looks harder than for cases WITH a BOM)
642 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
644 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
646 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
650 // 3 - test if encoding is specified in the xml declaration
652 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
653 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
654 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
655 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
656 $xmlChunk, $matches)) {
657 return strtoupper(substr($matches[2], 1, -1));
660 // 4 - if mbstring is available, let it do the guesswork
661 if (extension_loaded('mbstring')) {
662 if ($encodingPrefs == null && PhpXmlRpc::$xmlrpc_detectencodings != null) {
663 $encodingPrefs = PhpXmlRpc::$xmlrpc_detectencodings;
665 if ($encodingPrefs) {
666 $enc = mb_detect_encoding($xmlChunk, $encodingPrefs);
668 $enc = mb_detect_encoding($xmlChunk);
670 // NB: mb_detect likes to call it ascii, xml parser likes to call it US_ASCII...
671 // IANA also likes better US-ASCII, so go with it
672 if ($enc == 'ASCII') {
678 // no encoding specified: as per HTTP1.1 assume it is iso-8859-1?
679 // Both RFC 2616 (HTTP 1.1) and 1945 (HTTP 1.0) clearly state that for text/xxx content types
680 // this should be the standard. And we should be getting text/xml as request and response.
681 // BUT we have to be backward compatible with the lib, which always used UTF-8 as default...
682 return PhpXmlRpc::$xmlrpc_defencoding;
687 * Helper function: checks if an xml chunk as a charset declaration (BOM or in the xml declaration)
689 * @param string $xmlChunk
692 public static function hasEncoding($xmlChunk)
694 // scan the first bytes of the data for a UTF-16 (or other) BOM pattern
695 // (source: http://www.w3.org/TR/2000/REC-xml-20001006)
696 if (preg_match('/^(\x00\x00\xFE\xFF|\xFF\xFE\x00\x00|\x00\x00\xFF\xFE|\xFE\xFF\x00\x00)/', $xmlChunk)) {
698 } elseif (preg_match('/^(\xFE\xFF|\xFF\xFE)/', $xmlChunk)) {
700 } elseif (preg_match('/^(\xEF\xBB\xBF)/', $xmlChunk)) {
704 // test if encoding is specified in the xml declaration
706 // SPACE: (#x20 | #x9 | #xD | #xA)+ === [ \x9\xD\xA]+
707 // EQ: SPACE?=SPACE? === [ \x9\xD\xA]*=[ \x9\xD\xA]*
708 if (preg_match('/^<\?xml\s+version\s*=\s*' . "((?:\"[a-zA-Z0-9_.:-]+\")|(?:'[a-zA-Z0-9_.:-]+'))" .
709 '\s+encoding\s*=\s*' . "((?:\"[A-Za-z][A-Za-z0-9._-]*\")|(?:'[A-Za-z][A-Za-z0-9._-]*'))/",
710 $xmlChunk, $matches)) {