Script related fixes and optimizations.

ivopetkov · Sep 6, 2019 · 9a60716 · 9a60716
1 parent a56766e
commit 9a60716
Show file tree

Hide file tree

Showing 2 changed files with 19 additions and 5 deletions.
diff --git a/src/HTML5DOMDocument.php b/src/HTML5DOMDocument.php
@@ -88,17 +88,30 @@ public function loadHTML($source, $options = 0)
         }
 
         $source = trim($source);
+
+        // Add CDATA around script tags content
         $matches = null;
         preg_match_all('/<script(.*?)>/', $source, $matches);
         if (isset($matches[0])) {
+            $matches[0] = array_unique($matches[0]);
             foreach ($matches[0] as $match) {
                 if (substr($match, -2, 1) !== '/') { // check if ends with />
                     $source = str_replace($match, $match . '<![CDATA[html5-dom-document-internal-cdata', $source);
                 }
             }
         }
         $source = str_replace('</script>', 'html5-dom-document-internal-cdata]]></script>', $source);
-        $source = preg_replace('/(\<!\[CDATA\[html5-dom-document-internal-cdata.*?)\<\/(.*?html5-dom-document-internal-cdata\]\]>)/s', '$1<html5-dom-document-internal-cdata-endtagfix/$2', $source);
+        $source = str_replace('<![CDATA[html5-dom-document-internal-cdatahtml5-dom-document-internal-cdata]]>', '', $source); // clean empty script tags
+        $matches = null;
+        preg_match_all('/\<!\[CDATA\[html5-dom-document-internal-cdata.*?html5-dom-document-internal-cdata\]\]>/s', $source, $matches);
+        if (isset($matches[0])) {
+            $matches[0] = array_unique($matches[0]);
+            foreach ($matches[0] as $match) {
+                if (strpos($match, '</') !== false) { // check if contains </
+                    $source = str_replace($match, str_replace('</', '<html5-dom-document-internal-cdata-endtagfix/', $match), $source);
+                }
+            }
+        }
 
         $autoAddHtmlAndBodyTags = !defined('LIBXML_HTML_NOIMPLIED') || ($options & LIBXML_HTML_NOIMPLIED) === 0;
         $autoAddDoctype = !defined('LIBXML_HTML_NODEFDTD') || ($options & LIBXML_HTML_NODEFDTD) === 0;

diff --git a/tests/Test.php b/tests/Test.php
@@ -1324,14 +1324,15 @@ public function testSpecialCharsInScriptTags()
         $js2 = 'var f2=function(t){
             return t.replace(/</g,"&lt;").replace(/>/g,"&gt;");
         };';
-        $content = '<html><head><script src="url"/><script type="text/javascript">' . $js1 . '</script><script>' . $js2 . '</script></head></html>';
+        $content = '<html><head><script src="url1"/><script src="url2"></script><script type="text/javascript">' . $js1 . '</script><script>' . $js2 . '</script></head></html>';
         $dom = new HTML5DOMDocument();
         $dom->loadHTML($content);
         $scripts = $dom->querySelectorAll('script');
         $this->assertEquals($scripts[0]->innerHTML, '');
-        $this->assertEquals($scripts[1]->innerHTML, $js1);
-        $this->assertEquals($scripts[2]->innerHTML, $js2);
-        $this->assertEquals($dom->saveHTML(), "<!DOCTYPE html>\n" . str_replace('<script src="url"/>', '<script src="url"></script>', $content));
+        $this->assertEquals($scripts[1]->innerHTML, '');
+        $this->assertEquals($scripts[2]->innerHTML, $js1);
+        $this->assertEquals($scripts[3]->innerHTML, $js2);
+        $this->assertEquals($dom->saveHTML(), "<!DOCTYPE html>\n" . str_replace('<script src="url1"/>', '<script src="url1"></script>', $content));
     }
 
     /**