Add special case for end tag </br>. Fixes #185

IMSoP · IMSoP · commit 8696866ef257 · 2020-06-24T13:01:34.000+01:00
Normally, an end tag for a void element would simply be discarded,
but the spec includes a special rule as follows:

&gt; An end tag whose tag name is "br"
&gt; Parse error. Drop the attributes from the token, and act as
&gt; described in the next entry; i.e. act as if this was a "br"
&gt; start tag token with no attributes, rather than the end tag
&gt; token that it actually is.
diff --git a/src/HTML5/Parser/DOMTreeBuilder.php b/src/HTML5/Parser/DOMTreeBuilder.php
@@ -474,8 +474,14 @@ public function endTag($name)
     {
         $lname = $this->normalizeTagName($name);
 
-        // Ignore closing tags for unary elements.
-        if (Elements::isA($name, Elements::VOID_TAG)) {
+        // Special case within 12.2.6.4.7: An end tag whose tag name is "br" should be treated as an opening tag
+        if ($name === 'br') {
+            $this->parseError('Closing tag encountered for void element br.');
+
+            $this->startTag('br');
+        }
+        // Ignore closing tags for other unary elements.
+        elseif (Elements::isA($name, Elements::VOID_TAG)) {
             return;
         }
 
diff --git a/test/HTML5/Parser/DOMTreeBuilderTest.php b/test/HTML5/Parser/DOMTreeBuilderTest.php
@@ -715,4 +715,29 @@ public function testIAudioInParagraph()
         $this->assertSame('p', $audio->parentNode->nodeName);
         $this->assertSame(3, $audio->childNodes->length);
     }
+
+    public function testClosingBr()
+    {
+                $html = <<<EOM
+<!DOCTYPE html>
+<html>
+    <head>
+        <title>testClosingBr</title>
+    </head>
+    <body>
+    <p>
+        This line ends with a normal line break <br class="attribute-should-be-retained">
+        This line ends with a line break marked up as a closing tag </br class="attribute-should-be-discarded">
+     </p>
+    </body>
+</html>>
+</html>
+EOM;
+
+        $dom = $this->parse($html);
+
+        $this->assertSame(2, $dom->getElementsByTagName('br')->length);
+        $this->assertSame(1, $dom->getElementsByTagName('br')->item(0)->attributes->length);
+        $this->assertSame(0, $dom->getElementsByTagName('br')->item(1)->attributes->length);
+    }
 }
