diff --git a/src/HTML5DOMDocument.php b/src/HTML5DOMDocument.php
index bf4f310..08ceb65 100644
--- a/src/HTML5DOMDocument.php
+++ b/src/HTML5DOMDocument.php
@@ -57,7 +57,7 @@ public function loadHTML($source, $options = 0)
$source = '' . $source . '';
}
- if (stripos($source, '' . $source;
}
@@ -89,16 +89,8 @@ public function loadHTML($source, $options = 0)
}
// Preserve html entities
- $matches = [];
- preg_match_all('/&[a-zA-Z]*;/', $source, $matches);
- foreach ($matches[0] as $match) {
- $source = str_replace($match, 'html5-dom-document-internal-entity1-' . trim($match, '&;') . '-end', $source);
- }
- $matches = [];
- preg_match_all('/&#[0-9]*;/', $source, $matches);
- foreach ($matches[0] as $match) {
- $source = str_replace($match, 'html5-dom-document-internal-entity2-' . trim($match, '&#;') . '-end', $source);
- }
+ $source = preg_replace('/&([a-zA-Z]*);/', 'html5-dom-document-internal-entity1-$1-end', $source);
+ $source = preg_replace('/&#([0-9]*);/', 'html5-dom-document-internal-entity2-$1-end', $source);
$result = parent::loadHTML('' . $source, $options);
if ($internalErrorsOptionValue === false) {
@@ -295,24 +287,15 @@ public function saveHTML(\DOMNode $node = NULL)
$html = str_replace('', '', $html);
}
$html = str_replace('html5-dom-document-internal-content', '', $html);
- $matches = [];
- preg_match_all('/html5-dom-document-internal-entity1-(.*?)-end/', $html, $matches);
- foreach ($matches[0] as $i => $match) {
- $html = str_replace($match, '&' . $matches[1][$i] . ';', $html);
- }
- $matches = [];
- preg_match_all('/html5-dom-document-internal-entity2-(.*?)-end/', $html, $matches);
- foreach ($matches[0] as $i => $match) {
- $html = str_replace($match, '&#' . $matches[1][$i] . ';', $html);
+ if (strpos($html, 'html5-dom-document-internal-entity') !== false) {
+ $html = preg_replace('/html5-dom-document-internal-entity1-(.*?)-end/', '&$1;', $html);
+ $html = preg_replace('/html5-dom-document-internal-entity2-(.*?)-end/', '&#$1;', $html);
}
if ($removeHtmlElement) {
$html = str_replace('', '', $html);
}
- $voidElementsList = ['area', 'base', 'br', 'col', 'command', 'embed', 'hr', 'img', 'input', 'keygen', 'link', 'meta', 'param', 'source', 'track', 'wbr'];
- foreach ($voidElementsList as $elementName) {
- $html = str_replace('', '', $html);
- }
+ $html = str_replace(['', '', '', '', '', '', '', '', '', '', '', '', '', '', '', ''], '', $html);
// Remove the whitespace between the doctype and html tag
$html = preg_replace('/\>\s\ $match) {
- $value = str_replace($match, html_entity_decode('&' . $matches[1][$i] . ';'), $value);
- }
- $matches = [];
- preg_match_all('/html5-dom-document-internal-entity2-(.*?)-end/', $value, $matches);
- foreach ($matches[0] as $i => $match) {
- $value = str_replace($match, html_entity_decode('&#' . $matches[1][$i] . ';'), $value);
+ if (strpos($value, 'html5-dom-document-internal-entity') !== false) {
+ $matches = [];
+ preg_match_all('/html5-dom-document-internal-entity1-(.*?)-end/', $value, $matches);
+ foreach ($matches[0] as $i => $match) {
+ $value = str_replace($match, html_entity_decode('&' . $matches[1][$i] . ';'), $value);
+ }
+ $matches = [];
+ preg_match_all('/html5-dom-document-internal-entity2-(.*?)-end/', $value, $matches);
+ foreach ($matches[0] as $i => $match) {
+ $value = str_replace($match, html_entity_decode('&#' . $matches[1][$i] . ';'), $value);
+ }
}
return $value;
}