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 $this->output;
}
/**
* PHP5 constructor.
*/
function __construct($source)
{
# Check if PHP xml isn't compiled
#
if (!function_exists('xml_parser_create')) {
return trigger_error("PHP's XML extension is not available. Please contact your hosting provider to enable PHP's XML extension.");
}
$parser = xml_parser_create();
$this->parser = $parser;
# pass in parser, and a reference to this object
# set up handlers
#
xml_set_object($this->parser, $this);
xml_set_element_handler($this->parser, 'feed_start_element', 'feed_end_element');
xml_set_character_data_handler($this->parser, 'feed_cdata');
$status = xml_parse($this->parser, $source);
if (!$status) {
$errorcode = xml_get_error_code($this->parser);
if ($errorcode != XML_ERROR_NONE) {
$xml_error = xml_error_string($errorcode);
$error_line = xml_get_current_line_number($this->parser);
$error_col = xml_get_current_column_number($this->parser);
$errormsg = "{$xml_error} at line {$error_line}, column {$error_col}";
$this->error($errormsg);
}
}
xml_parser_free($this->parser);
$this->normalize();
}
/**
* Load translation data (TBX file reader)
*
* @param string $filename TBX file to add, full path must be given for access
* @param string $locale Locale has no effect for TBX because TBX defines all languages within
* the source file
* @param array $option OPTIONAL Options to use
* @throws Zend_Translation_Exception
* @return array
*/
protected function _loadTranslationData($filename, $locale, array $options = array())
{
$this->_data = array();
if (!is_readable($filename)) {
require_once 'Zend/Translate/Exception.php';
throw new Zend_Translate_Exception('Translation file \'' . $filename . '\' is not readable.');
}
$encoding = $this->_findEncoding($filename);
$this->_file = xml_parser_create($encoding);
xml_set_object($this->_file, $this);
xml_parser_set_option($this->_file, XML_OPTION_CASE_FOLDING, 0);
xml_set_element_handler($this->_file, "_startElement", "_endElement");
xml_set_character_data_handler($this->_file, "_contentElement");
try {
Zend_Xml_Security::scanFile($filename);
} catch (Zend_Xml_Exception $e) {
require_once 'Zend/Translate/Exception.php';
throw new Zend_Translate_Exception($e->getMessage());
}
if (!xml_parse($this->_file, file_get_contents($filename))) {
$ex = sprintf('XML error: %s at line %d of file %s', xml_error_string(xml_get_error_code($this->_file)), xml_get_current_line_number($this->_file), $filename);
xml_parser_free($this->_file);
require_once 'Zend/Translate/Exception.php';
throw new Zend_Translate_Exception($ex);
}
return $this->_data;
}
function MagpieRSS($source)
{
# if PHP xml isn't compiled in, die
#
if (!function_exists('xml_parser_create')) {
trigger_error("Failed to load PHP's XML Extension. http://www.php.net/manual/en/ref.xml.php");
}
$parser = @xml_parser_create();
if (!is_resource($parser)) {
trigger_error("Failed to create an instance of PHP's XML parser. http://www.php.net/manual/en/ref.xml.php");
}
$this->parser = $parser;
# pass in parser, and a reference to this object
# set up handlers
#
xml_set_object($this->parser, $this);
xml_set_element_handler($this->parser, 'feed_start_element', 'feed_end_element');
xml_set_character_data_handler($this->parser, 'feed_cdata');
$status = xml_parse($this->parser, $source);
if (!$status) {
$errorcode = xml_get_error_code($this->parser);
if ($errorcode != XML_ERROR_NONE) {
$xml_error = xml_error_string($errorcode);
$error_line = xml_get_current_line_number($this->parser);
$error_col = xml_get_current_column_number($this->parser);
$errormsg = "{$xml_error} at line {$error_line}, column {$error_col}";
$this->error($errormsg);
}
}
xml_parser_free($this->parser);
$this->normalize();
}
public function __construct($dataDirectory)
{
//
// Default to English for lack of anything better
//
$this->clientLang = 'en-us';
//
// See if we can find anything better
//
$matches = array();
if (array_key_exists('HTTP_ACCEPT_LANGUAGE', $_SERVER)) {
setlocale(LC_ALL, 'en-US');
if (preg_match('/(en-US|fr-FR|de-DE|it-IT|es-ES|sv-SV|sv-SE|nl-NL)/i', $_SERVER['HTTP_ACCEPT_LANGUAGE'], $matches)) {
$this->clientLang = strtolower($matches[1]);
}
}
//
// Pull in all of the strings we need
//
$lang2file = array("es-es" => "es-ES.xml", "en-us" => "en-US.xml", "de-de" => "de-DE.xml", "fr-fr" => "fr-FR.xml", "it-it" => "it-IT.xml", "nl-nl" => "nl-NL.xml", "sv-sv" => "sv-SE.xml", "sv-se" => "sv-SE.xml");
$xml_parser = xml_parser_create();
$l10nValues = array();
$l10nIndexes = array();
$l10nContents = file_get_contents("{$dataDirectory}/" . $lang2file[$this->getClientLang()]);
if (!xml_parse_into_struct($xml_parser, $l10nContents, $l10nValues, $l10nIndexes)) {
die(sprintf("l10n: XML error: %s at line %d", xml_error_string(xml_get_error_code($xml_parser)), xml_get_current_line_number($xml_parser)));
} else {
foreach ($l10nIndexes["STRING"] as $l10nStringIndex) {
$this->l10n[$l10nValues[$l10nStringIndex]["attributes"]["ID"]] = $l10nValues[$l10nStringIndex]["value"];
}
}
xml_parser_free($xml_parser);
}
/**
* Run the loader, to load up field-restrictions from the XML file.
*
* @param string The default breadcrumbs
* @param string The breadcrumb XML data
*/
function go($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);
}
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 $this->node;
}
public function Parse($filename)
{
$root = new CNode(CNode::K_ROOT, $filename, CNode::T_ROOT);
$filename = $root->Get(CNode::FLD_VAL);
$xmlstring = file_get_contents($filename);
if ($xmlstring === FALSE) {
$root->SetErr("failed to read file {$filename}", CNode::E_FATAL);
return $root;
}
$parser = xml_parser_create();
xml_set_object($parser, $this);
xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, false);
xml_set_element_handler($parser, 'startElement', 'endElement');
xml_set_character_data_handler($parser, 'characterData');
// Build a Root node and initialize the node_stack...
$this->node_stack = array();
$this->cur_node = $root;
// parse the data and free the parser...
$result = xml_parse($parser, $xmlstring);
if (!$result) {
$err = 'XML error: ' . xml_error_string(xml_get_error_code($parser)) . ' at line ' . xml_get_current_line_number($parser);
$root->SetErr("failed to parse file {$filename}, {$err}", CNode::E_FATAL);
}
xml_parser_free($parser);
return $root;
}
function xmlize($data, $WHITE=1)
{
$data = trim($data);
$vals = $index = $array = array();
$parser = xml_parser_create();
xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 0);
xml_parser_set_option($parser, XML_OPTION_SKIP_WHITE, $WHITE);
if ( !xml_parse_into_struct($parser, $data, $vals, $index) ) {
die(sprintf("XML error: %s at line %d",
xml_error_string(xml_get_error_code($parser)),
xml_get_current_line_number($parser)));
}
xml_parser_free($parser);
$i = 0;
$tagname = $vals[$i]['tag'];
if ( isset ($vals[$i]['attributes'] ) ) {
$array[$tagname]['@'] = $vals[$i]['attributes'];
} else {
$array[$tagname]['@'] = array();
}
$array[$tagname]["#"] = xml_depth($vals, $i);
return $array;
}
public function parse($html)
{
// replace entities
$html = preg_replace('/&([a-z0-9#]{2,5});/i', '+$1;', $html);
//before sending to xml parser make sure we have valid xml by tidying it up
$html = Kwf_Util_Tidy::repairHtml($html);
$this->_stack = array();
$this->_ret = '';
$this->_parser = xml_parser_create();
xml_set_object($this->_parser, $this);
xml_set_element_handler($this->_parser, 'startElement', 'endElement');
xml_set_character_data_handler($this->_parser, 'characterData');
xml_set_default_handler($this->_parser, 'characterData');
$result = xml_parse($this->_parser, '<body>' . $html . '</body>', true);
if (!$result) {
// wenn man ein nicht geschlossenes <br> rein gibt, schreit er hier,
// macht aber normal weiter. wenns zu oft vorkommt, evtl. exception
// entfernen und ignorieren, oder was andres überlegen :-)
$errorCode = xml_get_error_code($this->_parser);
$ex = new Kwf_Exception("HtmlExport UrlParser XML Error {$errorCode}: " . xml_error_string($errorCode) . "in line " . xml_get_current_line_number($this->_parser) . " parsed html: " . $html);
$ex->logOrThrow();
}
// re-replace entities
$this->_ret = preg_replace('/\\+([a-z0-9#]{2,5});/i', '&$1;', $this->_ret);
return $this->_ret;
}
function _parseXML()
{
$this->items = array();
$this->item = array();
$this->is_item = FALSE;
$this->tag = '';
$this->level_base = 0;
$buf = $this->src;
// Detect encoding
$matches = array();
if (preg_match('/<\\?xml [^>]*\\bencoding="([a-z0-9-_]+)"/i', $buf, $matches)) {
$this->encoding = $matches[1];
} else {
$this->encoding = mb_detect_encoding($buf, $this->detect_order);
}
// Normalize to UTF-8 / ASCII
if (!in_array(strtolower($this->encoding), array('us-ascii', 'iso-8859-1', 'utf-8'))) {
$buf = mb_convert_encoding($buf, 'utf-8', $this->encoding);
$this->encoding = 'utf-8';
}
// Parsing
$xml_parser = xml_parser_create($this->encoding);
xml_set_element_handler($xml_parser, array(&$this, 'start_element'), array(&$this, 'end_element'));
xml_set_character_data_handler($xml_parser, array(&$this, 'character_data'));
if (!xml_parse($xml_parser, $buf, 1)) {
return sprintf('XML error: %s at line %d in %s', xml_error_string(xml_get_error_code($xml_parser)), xml_get_current_line_number($xml_parser), $buf);
}
xml_parser_free($xml_parser);
//return $this->items;
}
function GetXMLTreeProfile($xmlloc)
{
if (file_exists($xmlloc)) {
$data = implode('', file($xmlloc));
} else {
$fp = fopen($xmlloc, 'r');
$data = fread($fp, 100000000);
fclose($fp);
}
$data = preg_replace("/<knows>.*<\\/knows>/is", "", $data);
$parser = xml_parser_create('UTF-8');
xml_parser_set_option($parser, XML_OPTION_SKIP_WHITE, 0);
xml_parse_into_struct($parser, $data, $vals, $index);
xml_parser_free($parser);
$code = xml_get_error_code($parser);
if ($code != XML_ERROR_NONE) {
global $messages;
$messages[] = gettext("XML error: ") . xml_error_string($code);
}
$tree = array();
$i = 0;
if (isset($vals[$i]['attributes'])) {
$tree[$vals[$i]['tag']][]['ATTRIBUTES'] = $vals[$i]['attributes'];
$index = count($tree[$vals[$i]['tag']]) - 1;
$tree[$vals[$i]['tag']][$index] = array_merge($tree[$vals[$i]['tag']][$index], GetChildren($vals, $i));
} else {
$tree[$vals[$i]['tag']][] = GetChildren($vals, $i);
}
return $tree;
}
function parse()
{
set_error_handler(array(&$this, 'error_handler'));
array_unshift($this->ns_contexts, array());
$parser = xml_parser_create_ns();
xml_set_object($parser, $this);
xml_set_element_handler($parser, "start_element", "end_element");
xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, 0);
xml_parser_set_option($parser, XML_OPTION_SKIP_WHITE, 0);
xml_set_character_data_handler($parser, "cdata");
xml_set_default_handler($parser, "_default");
xml_set_start_namespace_decl_handler($parser, "start_ns");
xml_set_end_namespace_decl_handler($parser, "end_ns");
$this->content = '';
$ret = true;
$fp = fopen($this->FILE, "r");
while ($data = fread($fp, 4096)) {
if ($this->debug) {
$this->content .= $data;
}
if (!xml_parse($parser, $data, feof($fp))) {
trigger_error(sprintf(__('XML error: %s at line %d') . "\n", xml_error_string(xml_get_error_code($xml_parser)), xml_get_current_line_number($xml_parser)));
$ret = false;
break;
}
}
fclose($fp);
xml_parser_free($parser);
restore_error_handler();
return $ret;
}
/**
* Method parses a mirror.xml file.
*
* @param string $file GZIP stream resource
* @return void
* @throws \TYPO3\CMS\Extensionmanager\Exception\ExtensionManagerException in case of XML parser errors
*/
public function parseXml($file)
{
$this->createParser();
if (!is_resource($this->objXml)) {
throw new \TYPO3\CMS\Extensionmanager\Exception\ExtensionManagerException('Unable to create XML parser.', 1342641009);
}
// Disables the functionality to allow external entities to be loaded when parsing the XML, must be kept
$previousValueOfEntityLoader = libxml_disable_entity_loader(true);
// keep original character case of XML document
xml_parser_set_option($this->objXml, XML_OPTION_CASE_FOLDING, false);
xml_parser_set_option($this->objXml, XML_OPTION_SKIP_WHITE, false);
xml_parser_set_option($this->objXml, XML_OPTION_TARGET_ENCODING, 'utf-8');
xml_set_element_handler($this->objXml, 'startElement', 'endElement');
xml_set_character_data_handler($this->objXml, 'characterData');
if (!($fp = fopen($file, 'r'))) {
throw new \TYPO3\CMS\Extensionmanager\Exception\ExtensionManagerException(sprintf('Unable to open file resource %s.', $file), 1342641010);
}
while ($data = fread($fp, 4096)) {
if (!xml_parse($this->objXml, $data, feof($fp))) {
throw new \TYPO3\CMS\Extensionmanager\Exception\ExtensionManagerException(sprintf('XML error %s in line %u of file resource %s.', xml_error_string(xml_get_error_code($this->objXml)), xml_get_current_line_number($this->objXml), $file), 1342641011);
}
}
libxml_disable_entity_loader($previousValueOfEntityLoader);
xml_parser_free($this->objXml);
}
public function __construct($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 parseFile($persist = false)
{
if (empty($this->parser)) {
error_log("No parser, so allocating new xml parser");
$this->parser = xml_parser_create();
xml_set_object($this->parser, $this);
xml_set_element_handler($this->parser, "startHandler", "endHandler");
xml_set_character_data_handler($this->parser, "cdataHandler");
}
if (!empty($this->file)) {
// process & parse
$fp = fopen($this->file, 'r');
if ($fp) {
// could read the file successfully
while ($data = fread($fp, 4096)) {
xml_parse($this->parser, $data, feof($fp)) or error_log(sprintf("Error in parsing XML: %s at line %d", xml_error_string(xml_get_error_code($this->parser)), xml_get_current_line_number($this->parser)));
}
fclose($fp);
} else {
error_log("Could not read file (name: {$this->file}).");
}
} else {
error_log("Could not parse. No file specified.");
}
if (!$persist) {
xml_parser_free($this->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);
}
/**
* Format resource metadata.
*
* @param resource $value
*
* @return string
*/
protected function formatMetadata($value)
{
$props = array();
switch (get_resource_type($value)) {
case 'stream':
$props = stream_get_meta_data($value);
break;
case 'curl':
$props = curl_getinfo($value);
break;
case 'xml':
$props = array('current_byte_index' => xml_get_current_byte_index($value), 'current_column_number' => xml_get_current_column_number($value), 'current_line_number' => xml_get_current_line_number($value), 'error_code' => xml_get_error_code($value));
break;
}
if (empty($props)) {
return '{}';
}
$formatted = array();
foreach ($props as $name => $value) {
$formatted[] = sprintf('%s: %s', $name, $this->indentValue($this->presentSubValue($value)));
}
$template = sprintf('{%s%s%%s%s}', PHP_EOL, self::INDENT, PHP_EOL);
$glue = sprintf(',%s%s', PHP_EOL, self::INDENT);
return sprintf($template, implode($glue, $formatted));
}
/**
* Load translation data (TMX file reader)
*
* @param string $filename TMX file to add, full path must be given for access
* @param string $locale Locale has no effect for TMX because TMX defines all languages within
* the source file
* @param array $option OPTIONAL Options to use
* @throws Zend_Translation_Exception
*/
protected function _loadTranslationData($filename, $locale, array $options = array())
{
$options = array_merge($this->_options, $options);
if ($options['clear']) {
$this->_translate = array();
}
if (in_array('defined_language', $options) and !empty($options['defined_language'])) {
$this->_defined = true;
}
if (!is_readable($filename)) {
require_once 'Zend/Translate/Exception.php';
throw new Zend_Translate_Exception('Translation file \'' . $filename . '\' is not readable.');
}
$this->_file = xml_parser_create();
xml_set_object($this->_file, $this);
xml_parser_set_option($this->_file, XML_OPTION_CASE_FOLDING, 0);
xml_set_element_handler($this->_file, "_startElement", "_endElement");
xml_set_character_data_handler($this->_file, "_contentElement");
if (!xml_parse($this->_file, file_get_contents($filename))) {
$ex = sprintf('XML error: %s at line %d', xml_error_string(xml_get_error_code($this->_file)), xml_get_current_line_number($this->_file));
xml_parser_free($this->_file);
require_once 'Zend/Translate/Exception.php';
throw new Zend_Translate_Exception($ex);
}
}
function open($xml, $encoding = null)
{
if (!empty($encoding) && strtolower($encoding) != 'utf-8' && !Utils_Unicode::validate($xml)) {
if (preg_match('/^<\\?xml[^<]*\\s+encoding=["\']?([\\w-]+)["\']?/', $xml, $matches)) {
$encoding = $matches[1];
$xml = preg_replace('/^(<\\?xml[^<]*\\s+encoding=)["\']?[\\w-]+["\']?/', '$1"utf-8"', $xml, 1);
}
if (strcasecmp($encoding, 'utf-8')) {
$xml = Utils_Unicode::bring($xml, $encoding);
if (is_null($xml)) {
$this->error = XML_ERROR_UNKNOWN_ENCODING;
return false;
}
}
} else {
if (substr($xml, 0, 3) == "") {
$xml = substr($xml, 3);
}
}
$p = xml_parser_create();
xml_set_object($p, $this);
xml_parser_set_option($p, XML_OPTION_CASE_FOLDING, 0);
xml_set_element_handler($p, 'o', 'c');
xml_set_character_data_handler($p, 'd');
xml_set_default_handler($p, 'x');
$this->tree = array('children' => array());
$this->_cursor =& $this->tree;
$this->_cdata = false;
xml_parse($p, $xml);
unset($this->_cursor);
unset($this->_cdata);
$this->error = xml_get_error_code($p);
xml_parser_free($p);
return $this->error == XML_ERROR_NONE;
}
function parse($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 $this->result;
}
function parse($path, $data = '')
{
$this->state = 0;
/* reader */
if (!$this->v('reader')) {
ARC2::inc('Reader');
$this->reader =& new ARC2_Reader($this->a, $this);
}
$this->reader->setAcceptHeader('Accept: sparql-results+xml; q=0.9, */*; q=0.1');
$this->reader->activate($path, $data);
$this->x_base = isset($this->a['base']) && $this->a['base'] ? $this->a['base'] : $this->reader->base;
/* xml parser */
$this->initXMLParser();
/* parse */
$first = true;
while ($d = $this->reader->readStream()) {
if (!xml_parse($this->xml_parser, $d, false)) {
$error_str = xml_error_string(xml_get_error_code($this->xml_parser));
$line = xml_get_current_line_number($this->xml_parser);
$this->tmp_error = 'XML error: "' . $error_str . '" at line ' . $line . ' (parsing as ' . $this->getEncoding() . ')';
return $this->addError($this->tmp_error);
}
}
$this->target_encoding = xml_parser_get_option($this->xml_parser, XML_OPTION_TARGET_ENCODING);
xml_parser_free($this->xml_parser);
$this->reader->closeStream();
return $this->done();
}
function loadXML($file)
{
$this->_parser = xml_parser_create($this->getEncoding($file));
// Auto detect for PHP4/PHP5
xml_set_object($this->_parser, $this);
xml_set_element_handler($this->_parser, "_saxStartElement", "_saxEndElement");
xml_set_character_data_handler($this->_parser, "_saxCharacterData");
xml_parser_set_option($this->_parser, XML_OPTION_TARGET_ENCODING, "UTF-8");
if ($fp = fopen($file, "r")) {
$data = '';
while (!feof($fp)) {
$data .= fread($fp, 8192);
}
fclose($fp);
// Strip slashes
if (ini_get("magic_quotes_gpc")) {
$data = stripslashes($data);
}
// XML parse
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)), FATAL);
}
} else {
trigger_error("Could not open XML language pack: " . $file, FATAL);
}
xml_parser_free($this->_parser);
}
function fetch($raw_xml, $final = true)
{
$this->log .= 'fetch() called.' . PHP_EOL;
$this->log .= 'Raw XML:' . PHP_EOL . $raw_xml . PHP_EOL . PHP_EOL;
$this->index = 0;
$this->data = new stdClass();
$this->stack = array();
$this->stack[] =& $this->data;
$this->parser = xml_parser_create('UTF-8');
xml_set_object($this->parser, $this);
xml_set_element_handler($this->parser, 'tag_open', 'tag_close');
xml_set_character_data_handler($this->parser, 'cdata');
xml_parser_set_option($this->parser, XML_OPTION_TARGET_ENCODING, 'UTF-8');
xml_parser_set_option($this->parser, XML_OPTION_SKIP_WHITE, 1);
xml_parser_set_option($this->parser, XML_OPTION_CASE_FOLDING, false);
$start = 0;
$length = strlen($raw_xml);
$chunk_size = 32 * 1024 * 1024;
$ret = true;
while (true) {
if (!($parsed_xml = xml_parse($this->parser, substr($raw_xml, $start, $chunk_size), $final = ($start += $chunk_size) >= $length ? true : false))) {
$this->log .= $ret = sprintf('XML error: %s at line %d.', xml_error_string(xml_get_error_code($this->parser)), xml_get_current_line_number($this->parser));
break;
} else {
if ($final) {
break;
}
}
}
xml_parser_free($this->parser);
return $ret;
}
/**
* Load translation data (XMLTM file reader)
*
* @param string $locale Locale/Language to add data for, identical with locale identifier,
* see Zend_Locale for more information
* @param string $filename XMLTM file to add, full path must be given for access
* @param array $option OPTIONAL Options to use
* @throws \Zend\Translator\Adapter\Exception\InvalidArgumentException
* @throws \Zend\Translator\Adapter\Exception\InvalidFileTypeException
* @return array
*/
protected function _loadTranslationData($filename, $locale, array $options = array())
{
$this->_data = array();
$this->_lang = $locale;
if (!is_readable($filename)) {
throw new InvalidArgumentException('Translation file \'' . $filename . '\' is not readable.');
}
$encoding = $this->_findEncoding($filename);
$this->_file = xml_parser_create($encoding);
xml_set_object($this->_file, $this);
xml_parser_set_option($this->_file, XML_OPTION_CASE_FOLDING, 0);
xml_set_element_handler($this->_file, "_startElement", "_endElement");
xml_set_character_data_handler($this->_file, "_contentElement");
if (!xml_parse($this->_file, file_get_contents($filename))) {
$ex = sprintf('XML error: %s at line %d',
xml_error_string(xml_get_error_code($this->_file)),
xml_get_current_line_number($this->_file));
xml_parser_free($this->_file);
throw new InvalidFileTypeException($ex);
}
return $this->_data;
}
/**
* Pour tester la validité d'un document XML, on peut utiliser un analyseur syntaxique XML : http://fr3.php.net/manual/fr/book.xml.php
* Voir en particulier l'exemple http://fr3.php.net/manual/fr/example.xml-structure.php
*
* Mais ceci ne permet pas de vérifier la conformité d'un XML avec une DTD.
* DOMDocument le permet : http://fr2.php.net/manual/fr/domdocument.validate.php
* Mais d'une part ça emmet des warnings et d'autre part ça ne retourne qu'un booléen sans détails sur les erreurs trouvées
*
* Pour y remédier on peut utiliser cette extention de classe "MyDOMDocument" : http://fr2.php.net/manual/fr/domdocument.validate.php#85792
* Mais attention : il faut lui fournir un objet DOMDocument et load ou loadXML provoquent des warnings préliminaires si le XML est mal formé.
*
* Ma solution est d'utiliser :
* 1. dans un premier temps l'analyseur syntaxique XML xml_parse pour vérifier que le XML est bien formé
* 2. dans un second temps l'extention de classe MyDOMDocument pour vérifier la conformité avec la DTD
*
* J'en ai fait la fonction ci-dessous "analyser_XML($fichier)".
* La classe "MyDOMDocument" est dans autochargée (elle se trouve ici : _inc/class.domdocument.php).
*
* @param string $fichier_adresse
* @return string 'ok' ou un message d'erreur
*/
private static function analyser_XML($fichier_adresse)
{
// Récupération du contenu du fichier
$fichier_contenu = file_get_contents($fichier_adresse);
$fichier_contenu = To::deleteBOM(To::utf8($fichier_contenu)); // Mettre en UTF-8 si besoin et retirer le BOM éventuel
FileSystem::ecrire_fichier($fichier_adresse,$fichier_contenu); // Mettre à jour le fichier au cas où.
// Analyse XML (s'arrête à la 1ère erreur trouvée)
$xml_parser = xml_parser_create();
$valid_XML = xml_parse($xml_parser , $fichier_contenu , TRUE);
if(!$valid_XML)
{
return sprintf("Erreur XML ligne %d (%s)" , xml_get_current_line_number($xml_parser) , xml_error_string(xml_get_error_code($xml_parser)));
}
xml_parser_free($xml_parser);
// Analyse DTD (renvoie un tableau d'erreurs, affiche la dernière)
$xml = new DOMDocument;
$xml -> load($fichier_adresse);
$xml = new MyDOMDocument($xml);
$valid_DTD = $xml->validate();
if(!$valid_DTD)
{
return 'Erreur DTD : '.end($xml->errors);
}
// Tout est ok
return 'ok';
}
function analyser($file = "")
{
global $charset;
if (!($fp = @fopen($this->file, "r"))) {
die("impossible d'ouvrir le fichier {$this->file}");
}
$data = fread($fp, filesize($this->file));
$rx = "/<?xml.*encoding=[\\'\"](.*?)[\\'\"].*?>/m";
if (preg_match($rx, $data, $m)) {
$encoding = strtoupper($m[1]);
} else {
$encoding = "ISO-8859-1";
}
$this->analyseur = xml_parser_create($encoding);
xml_parser_set_option($this->analyseur, XML_OPTION_TARGET_ENCODING, $charset);
xml_parser_set_option($this->analyseur, XML_OPTION_CASE_FOLDING, true);
xml_set_object($this->analyseur, $this);
xml_set_element_handler($this->analyseur, "debutBalise", "finBalise");
xml_set_character_data_handler($this->analyseur, "texte");
fclose($fp);
if (!xml_parse($this->analyseur, $data, TRUE)) {
die(sprintf("erreur XML %s à la ligne: %d ( {$this->file} )\n\n", xml_error_string(xml_get_error_code($this->analyseur)), xml_get_current_line_number($this->analyseur)));
}
xml_parser_free($this->analyseur);
}
function parseFile($fileLocation)
{
// reset it.
$this->_docType = NULL;
$this->_nameSpace = NULL;
$this->_errors = NULL;
$fp = @fopen($fileLocation, 'r');
if ($fp) {
$parser = xml_parser_create('ISO-8859-1');
xml_set_object($parser, $this);
xml_parser_set_option($parser, XML_OPTION_CASE_FOLDING, FALSE);
xml_set_element_handler($parser, "_startElement", "_endElement");
while ($data = fread($fp, 1024)) {
if (strlen($this->_docType) > 0) {
break;
}
if (!xml_parse($parser, $data, feof($fp))) {
$error = xml_error_string(xml_get_error_code($parser));
break;
}
}
xml_parser_free($parser);
@fclose(fp);
return TRUE;
} else {
$this->_errors[] = 'File ' . $fileLocation . ' could not be opened.';
return FALSE;
}
}
/**
* Load translation data (XLIFF file reader)
*
* @param string $locale Locale/Language to add data for, identical with locale identifier,
* see Zend_Locale for more information
* @param string $filename XLIFF file to add, full path must be given for access
* @param array $option OPTIONAL Options to use
* @throws Zend_Translation_Exception
* @return array
*/
protected function _loadTranslationData($filename, $locale, array $options = array())
{
$this->_data = array();
if (!is_readable($filename)) {
require_once 'Zend/Translate/Exception.php';
throw new Zend_Translate_Exception('Translation file \'' . $filename . '\' is not readable.');
}
if (empty($options['useId'])) {
$this->_useId = false;
} else {
$this->_useId = true;
}
$encoding = $this->_findEncoding($filename);
$this->_target = $locale;
$this->_file = xml_parser_create($encoding);
xml_set_object($this->_file, $this);
xml_parser_set_option($this->_file, XML_OPTION_CASE_FOLDING, 0);
xml_set_element_handler($this->_file, "_startElement", "_endElement");
xml_set_character_data_handler($this->_file, "_contentElement");
if (!xml_parse($this->_file, file_get_contents($filename))) {
$ex = sprintf('XML error: %s at line %d', xml_error_string(xml_get_error_code($this->_file)), xml_get_current_line_number($this->_file));
xml_parser_free($this->_file);
require_once 'Zend/Translate/Exception.php';
throw new Zend_Translate_Exception($ex);
}
return $this->_data;
}
function parse($fbfile)
{
// Read the data from file
$data = NULL;
$fp = fopen($fbfile, "r");
if ($fp) {
$data = fread($fp, 1024);
fclose($fp);
}
if (!$data) {
return false;
}
// Initialize the XML parser
$parser = xml_parser_create();
xml_set_object($parser, $this);
xml_set_element_handler($parser, "tagStart", "tagEnd");
xml_set_character_data_handler($parser, "tagContent");
if (!xml_parse($parser, $data, false)) {
$errMsg = xml_error_string(xml_get_error_code($parser));
error_log("FB2PDF ERROR. Unable parse XML: {$errMsg}");
xml_parser_free($parser);
return false;
}
xml_parser_free($parser);
if ($this->rootElement != $this->ROOT) {
return false;
}
return $this->getBookInfo();
}