From 6437e87b77eec9870ac042298731c8cf34fc8b74 Mon Sep 17 00:00:00 2001
From: Andres Rey <andreskrey@gmail.com>
Date: Thu, 9 Nov 2017 18:37:40 +0000
Subject: Better detection of empty paragraphs

---
 src/HTMLParser.php  | 11 +++++++++--
 src/Readability.php | 12 ++++++++++++
 2 files changed, 21 insertions(+), 2 deletions(-)

(limited to 'src')

diff --git a/src/HTMLParser.php b/src/HTMLParser.php
index 8f00eda..1345bc1 100644
--- a/src/HTMLParser.php
+++ b/src/HTMLParser.php
@@ -1454,11 +1454,18 @@ class HTMLParser
     private function hasSinglePNode(Readability $node)
     {
         // There should be exactly 1 element child which is a P:
-        // And there should be no text nodes with real content (param true on ->getChildren)
-        if (count($children = $node->getChildren(true)) !== 1 || !$children[0]->tagNameEqualsTo('p')) {
+        if (count($children = $node->getChildren()) !== 1 || !$children[0]->tagNameEqualsTo('p')) {
             return false;
         }
 
+        // And there should be no text nodes with real content (param true on ->getChildren)
+        foreach ($children as $child) {
+            /** @var $child Readability */
+            if ($child->nodeTypeEqualsTo(XML_TEXT_NODE) && !preg_match('/\S$/', $child->getTextContent())) {
+                return false;
+            }
+        }
+
         return true;
     }
 
diff --git a/src/Readability.php b/src/Readability.php
index b0a5830..922e60f 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -77,6 +77,18 @@ class Readability extends Element implements ReadabilityInterface
         return false;
     }
 
+    /**
+     * Checks for the node type.
+     *
+     * @param string $value Type of node to compare to
+     *
+     * @return bool
+     */
+    public function nodeTypeEqualsTo($value)
+    {
+        return $this->node->nodeType === $value;
+    }
+
     /**
      * Get the ancestors of the current node.
      *
-- 
cgit v1.2.3