Fix hasSingleTagInsideElement method

It would fail for e.g. `<div> <p>foo</p> </div>`.

mozilla/readability uses children for the tag lookup, which return only elements.
PHP does not have children property so b580cf2
mistakenly used `childNodes` instead, but that can return any node type.

Let’s filter the children ourselves.

Also add comments from mozilla/readability’s `_hasSingleTagInsideElement`.

src/Readability.php

@@ -1469,13 +1469,23 @@ private function isPhrasingContent($node): bool
             }, iterator_to_array($node->childNodes)), true));
     }
 
+    /**
+     * Checks if `$node` has only whitespace and a single element with `$tag` for the tag name.
+     * Returns false if `$node` contains non-empty text nodes
+     * or if it contains no element with given tag or more than 1 element.
+     */
     private function hasSingleTagInsideElement(\DOMElement $node, string $tag): bool
     {
-        if (1 !== $node->childNodes->length || $node->childNodes->item(0)->nodeName !== $tag) {
+        $childNodes = iterator_to_array($node->childNodes);
+        $children = array_filter($childNodes, fn ($childNode) => $childNode instanceof \DOMElement);
+
+        // There should be exactly 1 element child with given tag
+        if (1 !== \count($children) || $children[0]->nodeName !== $tag) {
             return false;
         }
 
-        $a = array_filter(iterator_to_array($node->childNodes), function ($childNode) {
+        // And there should be no text nodes with real content
+        $a = array_filter($childNodes, function ($childNode) {
             return $childNode instanceof \DOMText &&
                 preg_match($this->regexps['hasContent'], $this->getInnerText($childNode));
         });