/**
* Parses a full HTML document.
* @param $text HTML text to parse
* @param $builder Custom builder implementation
* @return Parsed HTML as DOMDocument
*/
public static function parse($text, $builder = null)
{
// Cleanup invalid HTML
$doc = new DOMDocument();
if (mb_detect_encoding($text, "UTF-8", true) == "UTF-8") {
@$doc->loadHTML('<?xml encoding="UTF-8" ?>' . $text);
} else {
@$doc->loadHTML($text);
}
$text = $doc->saveHTML();
$tokenizer = new HTML5_Tokenizer($text, $builder);
$tokenizer->parse();
return $tokenizer->save();
}
public function parse()
{
$this->content_model = $this->_contentModelFlag;
if ($this->_lastStartFlag) {
$this->token = array('type' => self::STARTTAG, 'name' => $this->_lastStartFlag);
}
return parent::parse();
}
public function invoke($test)
{
// this is totally the wrong interface to use, but
// for now we need testing
$tokenizer = new HTML5_Tokenizer($test['data']);
$GLOBALS['TIME'] -= get_microtime();
if (isset($test['document-fragment'])) {
$tokenizer->parseFragment($test['document-fragment']);
} else {
$tokenizer->parse();
}
$GLOBALS['TIME'] += get_microtime();
$this->assertIdentical($test['document'], HTML5_TestData::strDom($tokenizer->save()), $test);
}
/**
* Loads an HTML string
* Parse errors are stored in the global array _dompdf_warnings.
* @todo use the $encoding variable
*
* @param string $str HTML text to load
* @param string $encoding Not used yet
*/
function load_html($str, $encoding = null)
{
$this->save_locale();
// FIXME: Determine character encoding, switch to UTF8, update meta tag. Need better http/file stream encoding detection, currently relies on text or meta tag.
mb_detect_order('auto');
if (mb_detect_encoding($str) !== 'UTF-8') {
$metatags = array('@<meta\\s+http-equiv="Content-Type"\\s+content="(?:[\\w/]+)(?:;\\s*?charset=([^\\s"]+))?@i', '@<meta\\s+content="(?:[\\w/]+)(?:;\\s*?charset=([^\\s"]+))"?\\s+http-equiv="Content-Type"@i', '@<meta [^>]*charset\\s*=\\s*["\']?\\s*([^"\' ]+)@i');
foreach ($metatags as $metatag) {
if (preg_match($metatag, $str, $matches)) {
break;
}
}
if (mb_detect_encoding($str) == '') {
if (isset($matches[1])) {
$encoding = strtoupper($matches[1]);
} else {
$encoding = 'UTF-8';
}
} else {
if (isset($matches[1])) {
$encoding = strtoupper($matches[1]);
} else {
$encoding = 'auto';
}
}
if ($encoding !== 'UTF-8') {
$str = mb_convert_encoding($str, 'UTF-8', $encoding);
}
if (isset($matches[1])) {
$str = preg_replace('/charset=([^\\s"]+)/i', 'charset=UTF-8', $str);
} else {
$str = str_replace('<head>', '<head><meta http-equiv="Content-Type" content="text/html;charset=UTF-8">', $str);
}
} else {
$encoding = 'UTF-8';
}
// remove BOM mark from UTF-8, it's treated as document text by DOMDocument
// FIXME: roll this into the encoding detection using UTF-8/16/32 BOM (http://us2.php.net/manual/en/function.mb-detect-encoding.php#91051)?
if (substr($str, 0, 3) == chr(0xef) . chr(0xbb) . chr(0xbf)) {
$str = substr($str, 3);
}
// Parse embedded php, first-pass
if ($this->get_option("enable_php")) {
ob_start();
eval("?" . ">{$str}");
$str = ob_get_clean();
}
// if the document contains non utf-8 with a utf-8 meta tag chars and was
// detected as utf-8 by mbstring, problems could happen.
// http://devzone.zend.com/article/8855
if ($encoding !== 'UTF-8') {
$re = '/<meta ([^>]*)((?:charset=[^"\' ]+)([^>]*)|(?:charset=["\'][^"\' ]+["\']))([^>]*)>/i';
$str = preg_replace($re, '<meta $1$3>', $str);
}
// Store parsing warnings as messages
set_error_handler("record_warnings");
// @todo Take the quirksmode into account
// http://hsivonen.iki.fi/doctype/
// https://developer.mozilla.org/en/mozilla's_quirks_mode
$quirksmode = false;
if ($this->get_option("enable_html5_parser")) {
$tokenizer = new HTML5_Tokenizer($str);
$tokenizer->parse();
$doc = $tokenizer->save();
// Remove #text children nodes in nodes that shouldn't have
$tag_names = array("html", "table", "tbody", "thead", "tfoot", "tr");
foreach ($tag_names as $tag_name) {
$nodes = $doc->getElementsByTagName($tag_name);
foreach ($nodes as $node) {
self::remove_text_nodes($node);
}
}
$quirksmode = $tokenizer->getTree()->getQuirksMode() > HTML5_TreeBuilder::NO_QUIRKS;
} else {
// loadHTML assumes ISO-8859-1 unless otherwise specified, but there are
// bugs in how DOMDocument determines the actual encoding. Converting to
// HTML-ENTITIES prior to import appears to resolve the issue.
// http://devzone.zend.com/1538/php-dom-xml-extension-encoding-processing/ (see #4)
// http://stackoverflow.com/a/11310258/264628
$doc = new DOMDocument();
$doc->preserveWhiteSpace = true;
$doc->loadHTML(mb_convert_encoding($str, 'HTML-ENTITIES', 'UTF-8'));
// If some text is before the doctype, we are in quirksmode
if (preg_match("/^(.+)<!doctype/i", ltrim($str), $matches)) {
$quirksmode = true;
} elseif (!preg_match("/^<!doctype/i", ltrim($str), $matches)) {
$quirksmode = true;
} else {
// HTML5 <!DOCTYPE html>
if (!$doc->doctype->publicId && !$doc->doctype->systemId) {
$quirksmode = false;
}
// not XHTML
if (!preg_match("/xhtml/i", $doc->doctype->publicId)) {
$quirksmode = true;
}
}
}
$this->_xml = $doc;
$this->_quirksmode = $quirksmode;
$this->_tree = new Frame_Tree($this->_xml);
restore_error_handler();
$this->restore_locale();
}
/**
* Parses a full HTML document.
* @param $text HTML text to parse
* @param $builder Custom builder implementation
* @return Parsed HTML as DOMDocument
*/
public static function parse($text, $builder = null)
{
$tokenizer = new HTML5_Tokenizer($text, $builder);
$tokenizer->parse();
return $tokenizer->save();
}
function load_html($str, $encoding = null)
{
$this->save_locale();
mb_detect_order('auto');
if (mb_detect_encoding($str) !== 'UTF-8') {
$metatags = array('@<meta\\s+http-equiv="Content-Type"\\s+content="(?:[\\w/]+)(?:;\\s*?charset=([^\\s"]+))?@i', '@<meta\\s+content="(?:[\\w/]+)(?:;\\s*?charset=([^\\s"]+))"?\\s+http-equiv="Content-Type"@i', '@<meta [^>]*charset\\s*=\\s*["\']?\\s*([^"\' ]+)@i');
foreach ($metatags as $metatag) {
if (preg_match($metatag, $str, $matches)) {
break;
}
}
if (mb_detect_encoding($str) == '') {
if (isset($matches[1])) {
$encoding = strtoupper($matches[1]);
} else {
$encoding = 'UTF-8';
}
} else {
if (isset($matches[1])) {
$encoding = strtoupper($matches[1]);
} else {
$encoding = 'auto';
}
}
if ($encoding !== 'UTF-8') {
$str = mb_convert_encoding($str, 'UTF-8', $encoding);
}
if (isset($matches[1])) {
$str = preg_replace('/charset=([^\\s"]+)/i', 'charset=UTF-8', $str);
} else {
$str = str_replace('<head>', '<head><meta http-equiv="Content-Type" content="text/html;charset=UTF-8">', $str);
}
} else {
$encoding = 'UTF-8';
}
if (substr($str, 0, 3) == chr(0xef) . chr(0xbb) . chr(0xbf)) {
$str = substr($str, 3);
}
if (DOMPDF_ENABLE_PHP) {
ob_start();
eval("?" . ">{$str}");
$str = ob_get_clean();
}
if ($encoding !== 'UTF-8') {
$re = '/<meta ([^>]*)((?:charset=[^"\' ]+)([^>]*)|(?:charset=["\'][^"\' ]+["\']))([^>]*)>/i';
$str = preg_replace($re, '<meta $1$3>', $str);
}
set_error_handler("record_warnings");
$quirksmode = false;
if (DOMPDF_ENABLE_HTML5PARSER) {
$tokenizer = new HTML5_Tokenizer($str);
$tokenizer->parse();
$doc = $tokenizer->save();
$tag_names = array("html", "table", "tbody", "thead", "tfoot", "tr");
foreach ($tag_names as $tag_name) {
$nodes = $doc->getElementsByTagName($tag_name);
foreach ($nodes as $node) {
self::remove_text_nodes($node);
}
}
$quirksmode = $tokenizer->getTree()->getQuirksMode() > HTML5_TreeBuilder::NO_QUIRKS;
} else {
$doc = new DOMDocument();
$doc->preserveWhiteSpace = true;
$doc->loadHTML($str);
if (preg_match("/^(.+)<(!doctype|html)/i", ltrim($str), $matches)) {
$quirksmode = true;
} else {
if (!$doc->doctype->publicId && !$doc->doctype->systemId) {
$quirksmode = false;
}
if (!preg_match("/xhtml/i", $doc->doctype->publicId)) {
$quirksmode = true;
}
}
}
$this->_xml = $doc;
$this->_quirksmode = $quirksmode;
$this->_tree = new Frame_Tree($this->_xml);
restore_error_handler();
$this->restore_locale();
}