PHP xml_parser_set_option函数代码示例

 public function parse($xml_or_arr)
 {
     if (is_array($xml_or_arr)) {
         $output = array();
         foreach ($xml_or_arr as $val) {
             $to = $this->parse($val);
             $output = array_merge($output, $to);
         }
         return $output;
     }
     //		echo '<h1>xml in parser:</h1><pre>'; print_r($xml); echo '</pre>';
     // if we don't start with a processing instruction,
     // we add an outer node just to ensure it's a valid xml document
     // (i.e. only a single root node)
     if (substr($xml_or_arr, 0, 2) != '<?') {
         $xml_or_arr = "<{$this->insertedNode}>" . $xml_or_arr . "</{$this->insertedNode}>";
     }
     $this->parser = xml_parser_create();
     xml_set_object($this->parser, $this);
     xml_set_element_handler($this->parser, "tagOpen", "tagClosed");
     xml_parser_set_option($this->parser, XML_OPTION_CASE_FOLDING, 0);
     xml_parser_set_option($this->parser, XML_OPTION_SKIP_WHITE, 1);
     xml_set_character_data_handler($this->parser, "tagData");
     $this->xmldata = xml_parse($this->parser, $xml_or_arr);
     if (!$this->xmldata) {
         die(sprintf("XML error: %s at line %d", xml_error_string(xml_get_error_code($this->parser)), xml_get_current_line_number($this->parser)));
     }
     xml_parser_free($this->parser);
     $this->parseCompleted();
     return $this->output;
 }

 /**
  * constructor
  * 
  * @param  string  path of input stream 
  * @access public
  */
 function _parse_proppatch($path)
 {
     $this->success = true;
     $this->depth = 0;
     $this->props = array();
     $had_input = false;
     $f_in = fopen($path, "r");
     if (!$f_in) {
         $this->success = false;
         return;
     }
     $xml_parser = xml_parser_create_ns("UTF-8", " ");
     xml_set_element_handler($xml_parser, array(&$this, "_startElement"), array(&$this, "_endElement"));
     xml_set_character_data_handler($xml_parser, array(&$this, "_data"));
     xml_parser_set_option($xml_parser, XML_OPTION_CASE_FOLDING, false);
     while ($this->success && !feof($f_in)) {
         $line = fgets($f_in);
         if (is_string($line)) {
             $had_input = true;
             $this->success &= xml_parse($xml_parser, $line, false);
         }
     }
     if ($had_input) {
         $this->success &= xml_parse($xml_parser, "", true);
     }
     xml_parser_free($xml_parser);
     fclose($f_in);
 }

 /**
  * Run the loader, to load up field-restrictions from the XML file.
  *
  * @param  string			The default breadcrumbs
  * @param  string			The breadcrumb XML data
  */
 function go($current_breadcrumbs, $data)
 {
     $this->tag_stack = array();
     $this->attribute_stack = array();
     $this->substitution_current_match_key = NULL;
     $this->substitution_current_label = NULL;
     $this->links = array();
     $this->substitutions = array();
     $breadcrumb_tpl = do_template('BREADCRUMB_ESCAPED');
     $this->breadcrumb_tpl = $breadcrumb_tpl->evaluate();
     $this->current_breadcrumbs = $current_breadcrumbs;
     // Create and setup our parser
     $xml_parser = @xml_parser_create();
     if ($xml_parser === false) {
         return;
         // PHP5 default build on windows comes with this function disabled, so we need to be able to escape on error
     }
     xml_set_object($xml_parser, $this);
     @xml_parser_set_option($xml_parser, XML_OPTION_TARGET_ENCODING, get_charset());
     @xml_parser_set_option($xml_parser, XML_OPTION_CASE_FOLDING, 0);
     xml_set_element_handler($xml_parser, 'startElement', 'endElement');
     xml_set_character_data_handler($xml_parser, 'startText');
     // Run the parser
     if (@xml_parse($xml_parser, $data, true) == 0) {
         attach_message('breadcrumbs.xml: ' . xml_error_string(xml_get_error_code($xml_parser)), 'warn');
         return;
     }
     @xml_parser_free($xml_parser);
 }

 function load($xml_file)
 {
     $this->_path = dirname($xml_file);
     if (!file_exists($xml_file)) {
         return;
     }
     $fp = @fopen($xml_file, "r");
     if ($fp) {
         $data = '';
         while (!feof($fp)) {
             $data .= fread($fp, 8192);
         }
         fclose($fp);
         if (ini_get("magic_quotes_gpc")) {
             $data = stripslashes($data);
         }
     }
     $this->_parser = xml_parser_create('UTF-8');
     xml_set_object($this->_parser, $this);
     xml_set_element_handler($this->_parser, "_saxStartElement", "_saxEndElement");
     xml_parser_set_option($this->_parser, XML_OPTION_TARGET_ENCODING, "UTF-8");
     if (!xml_parse($this->_parser, $data, true)) {
         trigger_error(sprintf("Language pack loading failed, XML error: %s at line %d.", xml_error_string(xml_get_error_code($this->_parser)), xml_get_current_line_number($this->_parser)), E_USER_ERROR);
     }
     xml_parser_free($this->_parser);
 }

 public function __construct($input, $maxDepth = 20)
 {
     if (!is_string($input)) {
         throw new XmlToArrayException('No valid input.');
     }
     $this->_maxDepth = $maxDepth;
     $XMLParser = xml_parser_create();
     xml_parser_set_option($XMLParser, XML_OPTION_SKIP_WHITE, false);
     xml_parser_set_option($XMLParser, XML_OPTION_CASE_FOLDING, false);
     xml_parser_set_option($XMLParser, XML_OPTION_TARGET_ENCODING, 'UTF-8');
     xml_set_character_data_handler($XMLParser, array($this, '_contents'));
     xml_set_default_handler($XMLParser, array($this, '_default'));
     xml_set_element_handler($XMLParser, array($this, '_start'), array($this, '_end'));
     xml_set_external_entity_ref_handler($XMLParser, array($this, '_externalEntity'));
     xml_set_notation_decl_handler($XMLParser, array($this, '_notationDecl'));
     xml_set_processing_instruction_handler($XMLParser, array($this, '_processingInstruction'));
     xml_set_unparsed_entity_decl_handler($XMLParser, array($this, '_unparsedEntityDecl'));
     if (!xml_parse($XMLParser, $input, true)) {
         $errorCode = xml_get_error_code($XMLParser);
         $message = sprintf('%s. line: %d, char: %d' . ($this->_tagStack ? ', tag: %s' : ''), xml_error_string($errorCode), xml_get_current_line_number($XMLParser), xml_get_current_column_number($XMLParser) + 1, implode('->', $this->_tagStack));
         xml_parser_free($XMLParser);
         throw new XmlToArrayException($message, $errorCode);
     }
     xml_parser_free($XMLParser);
 }

 function parse($xml)
 {
     $this->result = array();
     $this->context = $xml;
     $this->currentTag = '';
     $this->currentState = array();
     $this->xmlParser = xml_parser_create();
     xml_parser_set_option($this->xmlParser, XML_OPTION_TARGET_ENCODING, "UTF-8") . xml_set_object($this->xmlParser, $this);
     xml_set_element_handler($this->xmlParser, 'startElement', 'endElement');
     xml_set_character_data_handler($this->xmlParser, 'content');
     $this->trigger('before:import');
     try {
         if (!xml_parse($this->xmlParser, $xml)) {
             $this->xmlParserError = 'Line ' . xml_get_current_line_number($this->xmlParser) . ': ' . (xml_get_error_code($this->xmlParser) ? xml_error_string(xml_get_error_code($this->xmlParser)) : 'Unknown error');
         }
     } catch (Exception $e) {
         $this->xmlParserError = $e->getMessage();
     }
     xml_parser_free($this->xmlParser);
     if ($this->xmlParserError) {
         $this->raiseError($this->xmlParserError);
     } else {
         if ($this->getCurrentState()) {
             $this->raiseError('Wrong ending state: ' . $this->getCurrentState());
         } else {
         }
     }
     $this->trigger('after:import');
     return $this->result;
 }

 public function fetch()
 {
     $this->query = $this->url . '?pid=' . $this->pid . '&noredirect=true&format=unixref&id=doi%3A' . $this->doi;
     $request_options = array('method' => 'POST');
     $result = drupal_http_request($this->query, $request_options);
     if ($result->code != 200) {
         drupal_set_message(t('HTTP error: !error when trying to contact crossref.org for XML input', array('!error' => $result->code)), 'error');
         return;
     }
     if (empty($result->data)) {
         drupal_set_message(t('Did not get any data from crossref.org'), 'error');
         return;
     }
     $sxml = @simplexml_load_string($result->data);
     if ($error = (string) $sxml->doi_record->crossref->error) {
         drupal_set_message($error, 'error');
         return;
     }
     $this->nodes = array();
     $this->parser = drupal_xml_parser_create($result->data);
     // use case-folding so we are sure to find the tag in
     xml_parser_set_option($this->parser, XML_OPTION_CASE_FOLDING, FALSE);
     xml_parser_set_option($this->parser, XML_OPTION_SKIP_WHITE, TRUE);
     xml_set_object($this->parser, $this);
     xml_set_element_handler($this->parser, 'unixref_startElement', 'unixref_endElement');
     xml_set_character_data_handler($this->parser, 'unixref_characterData');
     if (!xml_parse($this->parser, $result->data)) {
         drupal_set_message(sprintf("XML error: %s at line %d", xml_error_string(xml_get_error_code($this->parser)), xml_get_current_line_number($this->parser)), 'error');
     }
     xml_parser_free($this->parser);
     return $this->node;
 }

 public function xml() {
     $this->parser = xml_parser_create();    
     xml_parser_set_option($this->parser, XML_OPTION_CASE_FOLDING, false);    
     xml_set_object($this->parser, $this);    
     xml_set_element_handler($this->parser, 'open','close');    
     xml_set_character_data_handler($this->parser, 'data');    
 }

 /**
  * 获取xml树
  * @param string $xmldata xml数据
  * @param array $result xml内容
  */
 static function get_xml_tree($xmldata, &$result)
 {
     ini_set('track_errors', '1');
     $xmlreaderror = false;
     $parser = xml_parser_create();
     xml_parser_set_option($parser, XML_OPTION_SKIP_WHITE, 1);
     xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 0);
     if (!xml_parse_into_struct($parser, $xmldata, $vals, $index)) {
         $xmlreaderror = true;
         return 0;
     }
     xml_parser_free($parser);
     if (!$xmlreaderror) {
         $result = array();
         $i = 0;
         if (isset($vals[$i]['attributes'])) {
             foreach (array_keys($vals[$i]['attributes']) as $attkey) {
                 $attributes[$attkey] = $vals[$i]['attributes'][$attkey];
             }
         }
         $result[$vals[$i]['tag']] = array_merge((array) $attributes, self::get_children($vals, $i, 'open'));
     }
     ini_set('track_errors', '0');
     return 1;
 }

 function AbstractSAXParser()
 {
     // create a parser
     $this->parserResource = xml_parser_create();
     if (isset($this->parserResource)) {
         // allowing object instance to use the xml parser
         xml_set_object($this->parserResource, $this);
         // set tag event handler
         xml_set_element_handler($this->parserResource, "startTagElement", "endTagElement");
         // set CDATA event handler
         xml_set_character_data_handler($this->parserResource, "cdataElement");
         // set processing instruction handler
         xml_set_processing_instruction_handler($this->parserResource, "instructionElement");
         // set undeclare entity
         xml_set_unparsed_entity_decl_handler($this->parserResource, "undeclaredEntityElement");
         // set notation delcaration handler
         xml_set_notation_decl_handler($this->parserResource, "notationDeclarationElement");
         // set external entity handler
         xml_set_external_entity_ref_handler($this->parserResource, "externalEntityElement");
         // seat default parser option
         xml_parser_set_option($this->parserResource, XML_OPTION_SKIP_WHITE, 1);
         xml_parser_set_option($this->parserResource, XML_OPTION_CASE_FOLDING, 0);
         xml_parser_set_option($this->parserResource, XML_OPTION_TARGET_ENCODING, 'UTF-8');
     }
 }

function processXML($filename)
{
    // read the xml document
    $data = implode("", file($filename));
    $parser = xml_parser_create();
    xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 0);
    xml_parser_set_option($parser, XML_OPTION_SKIP_WHITE, 1);
    xml_parse_into_struct($parser, $data, $values, $tags);
    xml_parser_free($parser);
    $talks = array();
    // loop through the structures
    foreach ($tags as $key => $val) {
        if ($key == "talk") {
            $ranges = $val;
            $nranges = count($ranges);
            // each contiguous pair of array entries are the
            // lower and upper range for each talk elements
            for ($i = 0; $i < $nranges; $i += 2) {
                $offset = $ranges[$i] + 1;
                $len = $ranges[$i + 1] - $offset;
                $talks[] =& parseTalk(array_slice($values, $offset, $len));
            }
        } else {
            continue;
        }
    }
    return $talks;
}

 public function replaceTranslation()
 {
     //write xml header
     fwrite($this->outputFP, '<?xml version="1.0" encoding="UTF-8"?>');
     //create parser
     $xml_parser = xml_parser_create('UTF-8');
     //configure parser
     //pass this object to parser to make its variables and functions visible inside callbacks
     xml_set_object($xml_parser, $this);
     //avoid uppercasing all tags name
     xml_parser_set_option($xml_parser, XML_OPTION_CASE_FOLDING, false);
     //define callbacks for tags
     xml_set_element_handler($xml_parser, "tagOpen", "tagClose");
     //define callback for data
     xml_set_character_data_handler($xml_parser, "characterData");
     //read a chunk of text
     while ($this->currentBuffer = fread($this->originalFP, 4096)) {
         /*
           preprocess file
         */
         // obfuscate entities because sax automatically does html_entity_decode
         $temporary_check_buffer = preg_replace("/&(.*?);/", '#%$1#%', $this->currentBuffer);
         $lastByte = $temporary_check_buffer[strlen($temporary_check_buffer) - 1];
         //avoid cutting entities in half:
         //the last fread could have truncated an entity (say, '&lt;' in '&l'), thus invalidating the escaping
         //***** and if there is an & that it is not an entity, this is an infinite loop !!!!!
         $escape_AMP = false;
         // 9 is the max length of an entity. So, suppose that the & is at the end of buffer,
         // add 9 Bytes and substitute the entities, if the & is present and it is not at the end
         //it can't be a entity, exit the loop
         while (true) {
             $_ampPos = strpos($temporary_check_buffer, '&');
             //check for real entity or escape it to safely exit from the loop!!!
             if ($_ampPos === false || strlen(substr($temporary_check_buffer, $_ampPos)) > 9) {
                 $escape_AMP = true;
                 break;
             }
             //if an entity is still present, fetch some more and repeat the escaping
             $this->currentBuffer .= fread($this->originalFP, 9);
             $temporary_check_buffer = preg_replace("/&(.*?);/", '#%$1#%', $this->currentBuffer);
         }
         //free stuff outside the loop
         unset($temporary_check_buffer);
         $this->currentBuffer = preg_replace("/&(.*?);/", '#%$1#%', $this->currentBuffer);
         if ($escape_AMP) {
             $this->currentBuffer = str_replace("&", '#%amp#%', $this->currentBuffer);
         }
         //get lenght of chunk
         $this->len = strlen($this->currentBuffer);
         //parse chunk of text
         if (!xml_parse($xml_parser, $this->currentBuffer, feof($this->originalFP))) {
             //if unable, die
             die(sprintf("XML error: %s at line %d", xml_error_string(xml_get_error_code($xml_parser)), xml_get_current_line_number($xml_parser)));
         }
         //get accumulated this->offset in document: as long as SAX pointer advances, we keep track of total bytes it has seen so far; this way, we can translate its global pointer in an address local to the current buffer of text to retrieve last char of tag
         $this->offset += $this->len;
     }
     //close parser
     xml_parser_free($xml_parser);
 }

 /**
  * rss push
  */
 public function getPushList($id)
 {
     $rssInfo = $this->where('id=' . $id)->find();
     $ch = curl_init();
     curl_setopt($ch, CURLOPT_URL, $rssInfo['url']);
     curl_setopt($ch, CURLOPT_RETURNTRANSFER, 1);
     $result = curl_exec($ch);
     curl_close($ch);
     $parser = xml_parser_create();
     xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 0);
     xml_parser_set_option($parser, XML_OPTION_SKIP_WHITE, 1);
     xml_parse_into_struct($parser, $result, $values, $tags);
     xml_parser_free($parser);
     for ($i = 0; $i < $rssInfo['count']; $i++) {
         foreach ($tags as $k => $v) {
             if ($k == 'title') {
                 $newsList[$i]['title'] = $values[$v[$i + 2]]['value'];
             } elseif ($k == 'description') {
                 $description = $values[$v[$i + 1]]['value'];
                 //如果简介中有图片，就将其作为图文封面
                 $pattern = '/<img(.*)src="(.*)"/Us';
                 preg_match($pattern, $description, $content);
                 $newsList[$i]['cover'] = $content['2'];
                 $newsList[$i]['description'] = $values[$v[$i + 1]]['value'];
             } elseif ($k == 'link') {
                 $newsList[$i]['url'] = $values[$v[$i + 2]]['value'];
             }
         }
     }
     return $newsList;
 }

 /**
  * @param $fname
  */
 private function run($fname)
 {
     $parser = xml_parser_create_ns('UTF-8');
     // case folding violates XML standard, turn it off
     xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, false);
     xml_set_element_handler($parser, array($this, 'rootElementOpen'), false);
     if (file_exists($fname)) {
         $file = fopen($fname, "rb");
         if ($file) {
             do {
                 $chunk = fread($file, 32768);
                 $ret = xml_parse($parser, $chunk, feof($file));
                 if ($ret == 0) {
                     // XML isn't well-formed!
                     fclose($file);
                     xml_parser_free($parser);
                     return;
                 }
             } while (!feof($file));
             fclose($file);
         }
     }
     $this->wellFormed = true;
     xml_parser_free($parser);
 }

 function XMLToArray($data, $ignore = array(), $replace = array(), $showattribs = false, $toupper = false)
 {
     $this->_showAttribs = $showattribs;
     $this->_parser = xml_parser_create();
     xml_set_object($this->_parser, $this);
     if ($toupper) {
         foreach ((array) $ignore as $key => $value) {
             $this->_ignore[strtoupper($key)] = strtoupper($value);
         }
         foreach ((array) $replace as $key => $value) {
             $this->_replace[strtoupper($key)] = strtoupper($value);
         }
         xml_parser_set_option($this->_parser, XML_OPTION_CASE_FOLDING, true);
     } else {
         $this->_ignore =& $ignore;
         $this->_replace =& $replace;
         xml_parser_set_option($this->_parser, XML_OPTION_CASE_FOLDING, false);
     }
     xml_set_element_handler($this->_parser, '_startElement', '_endElement');
     xml_set_character_data_handler($this->_parser, '_cdata');
     $this->_data = array();
     $this->_level = 0;
     if (!xml_parse($this->_parser, $data, true)) {
         //new Error("XML Parse Error: ".xml_error_string(xml_get_error_code($this->_parser))."n on line: ".xml_get_current_line_number($this->_parser),true);
         return false;
     }
     xml_parser_free($this->_parser);
 }