[+]: try to fix "DOM auto remove HTML closing tag in <script> tag whe…

…n save" https://bugs.php.net/bug.php?id=74628
voku · Mar 3, 2019 · 4877411 · 4877411
1 parent fc06c35
commit 4877411
Show file tree

Hide file tree

Showing 2 changed files with 39 additions and 9 deletions.
diff --git a/src/voku/helper/HtmlDomParser.php b/src/voku/helper/HtmlDomParser.php
@@ -379,12 +379,16 @@ private function createDOMDocument(string $html, $libXMLExtraOptions = null): \D
             $this->isDOMDocumentCreatedWithFakeEndScript = true;
         }
 
-        if (
-            \strpos($html, 'type="text/html"') !== false
-            ||
-            \strpos($html, 'type=\'text/html\'') !== false
-        ) {
-            $this->keepSpecialScriptTags($html);
+        if (\strpos($html, '<script') !== false) {
+            $this->html5FallbackForScriptTags($html);
+
+            if (
+                \strpos($html, 'type="text/html"') !== false
+                ||
+                \strpos($html, 'type=\'text/html\'') !== false
+            ) {
+                $this->keepSpecialScriptTags($html);
+            }
         }
 
         // set error level
@@ -458,20 +462,37 @@ private function createDOMDocument(string $html, $libXMLExtraOptions = null): \D
         return $this->document;
     }
 
+    /**
+     * workaround for bug: https://bugs.php.net/bug.php?id=74628
+     *
+     * @param string $html
+     */
+    protected function html5FallbackForScriptTags(string &$html)
+    {
+        // regEx for e.g.: [<script id="elements-image-2">...<script>]
+        $regExSpecialScript = '/<(script)(?<attr>[^>]*)>(?<content>.*)<\/\1>/isU';
+        $html = \preg_replace_callback($regExSpecialScript, function($scripts) {
+            return '<script' . $scripts['attr'] . '>' . \str_replace('</', '<\/',$scripts['content']) . '</script>';
+        },$html);
+    }
+
     /**
      * @param string $html
      */
     protected function keepSpecialScriptTags(string &$html)
     {
         $specialScripts = [];
-        // regEx for e.g.: [<script id="elements-image-1" type="text/html">...<script>]
+        // regEx for e.g.: [<script id="elements-image-1" type="text/html">...</script>]
         $regExSpecialScript = '/<(script) [^>]*type=(["|\'])text\/html\2([^>]*)>.*<\/\1>/isU';
         \preg_match_all($regExSpecialScript, $html, $specialScripts);
 
         if (isset($specialScripts[0])) {
             foreach ($specialScripts[0] as $specialScript) {
-                $specialNonScript = '<' . self::$domHtmlSpecialScriptHelper . substr($specialScript, strlen('<script'));
-                $specialNonScript = substr($specialNonScript, 0, -strlen('</script>')) . '</' . self::$domHtmlSpecialScriptHelper . '>';
+
+                $specialNonScript = '<' . self::$domHtmlSpecialScriptHelper . \substr($specialScript, \strlen('<script'));
+                $specialNonScript = \substr($specialNonScript, 0, -\strlen('</script>')) . '</' . self::$domHtmlSpecialScriptHelper . '>';
+                // remove the html5 fallback
+                $specialNonScript = \str_replace('<\/', '</', $specialNonScript);
 
                 $html = \str_replace($specialScript, $specialNonScript, $html);
             }

diff --git a/tests/HtmlDomParserTest.php b/tests/HtmlDomParserTest.php
@@ -1384,6 +1384,15 @@ public function testScriptInCommentHtml()
         static::assertSame('<script class="script_1" type="text/javascript">someCode</script>', (string) $script);
     }
 
+    public function testHtmlAndJavaScriptMix()
+    {
+        $htmlAndJs = '<p>Text 1</p><script>$(".second-column-mobile-inner").wrapAll("<div class=\'collapse\' id=\'second-column\'></div>");</script><p>Text 2</p>';
+
+        $dom = HtmlDomParser::str_get_html($htmlAndJs);
+        $script = $dom->find('script');
+        static::assertSame('<script>$(".second-column-mobile-inner").wrapAll("<div class=\'collapse\' id=\'second-column\'><\/div>");</script>', (string) $script);
+    }
+
     public function testSpecialCharsAndPlaintext()
     {
         $file = __DIR__ . '/fixtures/test_page_plaintext.html';