Bug #70934 [Opn->Nab]: DOMDocument return wrong tree by tag in script tag

Edit report at https://bugs.php.net/bug.php?id=70934&edit=1 ID: 70934 Updated by: requinix@php.net Reported by: sashott at abv dot bg Summary: DOMDocument return wrong tree by tag in script tag -Status: Open +Status: Not a bug Type: Bug Package: DOM XML related Operating System: Linux PHP Version: 5.5Git-2015-11-18 (snap) Block user comment: N Private report: N New Comment: Actually this is expected behavior. The HTML parser is (still) only compliant with HTML 4, and in that scripts end at the first </ even if it's not a </script>. https://www.w3.org/TR/html4/appendix/notes.html#notes-specifying-data And HTML doesn't support CDATAs so adding one of those won't work. Previous Comments: ------------------------------------------------------------------------ [2018-03-28 16:04:44] luca dot canella at diennea dot com Bug confirmed and I wish to add another caveat. When script tag contains "text/html" the probability to have a broken tree grows exponentially. I also tried adding some <![CDATA[ ... ]]> but this isn't solving the problem. Example: ----------- html.html ----------- <!doctype html> <html lang="en"> <head> <meta charset="utf-8"> <meta http-equiv="X-UA-Compatible" content="IE=edge"> <title>Titolo</title> <meta name="description" content="aaaa"> <meta name="viewport" content="width=device-width, initial-scale=1"> </head> <body> <div class="wrapper"> <header></header> <main> <script type="text/html" id="ciao2"><![CDATA[ <div>DDD</div> ]]></script> </main> <footer></footer> </div> <script type="text/html" id="ciao"> <div>AAAA</div> </script> <script> console.log(document.getElementById('ciao'), document.getElementById('ciao2'), '<div>ZZZZ</div>'); </script> <div>BBBB</div> </body> </html> ----------- test.php ----------- <?php $html = file_get_contents('html.html'); $dom = DOMDocument::loadHtml($html); echo $dom->saveHTML(); -------------------------------- ------------------------------------------------------------------------ [2015-11-18 10:54:30] sashott at abv dot bg Description: ------------ DOMDocument return wrong tree by tag in variable in script tag. As example in DIV, a closing tag DIV written in script tag (in text variable), close the previous DIV. Test script: --------------- <?php $html_content='<!DOCTYPE html> <html> <body> <div id="something"> <script> var somevar=\'<div></div>\'; </script> <div></div> <div></div> </div> </body> </html> '; $oldSetting=libxml_use_internal_errors(true); libxml_clear_errors(); $html=new DOMDocument(); $html->loadHtml($html_content); $node=$html->getElementById('something'); echo "<pre>"; foreach($node->childNodes as $child_node){ echo $child_node->nodeName."\n"; } echo "</pre>"; libxml_clear_errors(); libxml_use_internal_errors($oldSetting); //Result: //#text //script //Error appear from: var somevar=\'<div></div>\'; ?> Expected result: ---------------- script div div //Can have some #text between them. Actual result: -------------- #text script ------------------------------------------------------------------------ -- Edit this bug report at https://bugs.php.net/bug.php?id=70934&edit=1