From 773d2bc4e54afcb1d2464ec61290e352b27cb881 Mon Sep 17 00:00:00 2001
From: Arthur Schiwon <blizzz@arthur-schiwon.de>
Date: Sat, 11 Jan 2025 00:39:10 +0100
Subject: [PATCH] fix: allow DOMProcessingInstruction in Readability and
 NodeUtility

Signed-off-by: Arthur Schiwon <blizzz@arthur-schiwon.de>
---
 src/Nodes/NodeUtility.php | 3 ++-
 src/Readability.php       | 9 +++++----
 2 files changed, 7 insertions(+), 5 deletions(-)

diff --git a/src/Nodes/NodeUtility.php b/src/Nodes/NodeUtility.php
index 2f094f1..859bb43 100644
--- a/src/Nodes/NodeUtility.php
+++ b/src/Nodes/NodeUtility.php
@@ -5,6 +5,7 @@
 use fivefilters\Readability\Nodes\DOM\DOMDocument;
 use fivefilters\Readability\Nodes\DOM\DOMElement;
 use fivefilters\Readability\Nodes\DOM\DOMNode;
+use fivefilters\Readability\Nodes\DOM\DOMProcessingInstruction;
 use fivefilters\Readability\Nodes\DOM\DOMText;
 use fivefilters\Readability\Nodes\DOM\DOMComment;
 use fivefilters\Readability\Nodes\DOM\DOMNodeList;
@@ -120,7 +121,7 @@ public static function removeNode(DOMNode|DOMComment|DOMText|DOMElement $node):
      * Returns the next node. First checks for children (if the flag allows it), then for siblings, and finally
      * for parents.
      */
-    public static function getNextNode(DOMNode|DOMComment|DOMText|DOMElement|DOMDocument $originalNode, bool $ignoreSelfAndKids = false): DOMNode|DOMComment|DOMText|DOMElement|DOMDocument|null
+    public static function getNextNode(DOMNode|DOMComment|DOMText|DOMElement|DOMDocument|DOMProcessingInstruction $originalNode, bool $ignoreSelfAndKids = false): DOMNode|DOMComment|DOMText|DOMElement|DOMDocument|DOMProcessingInstruction|null
     {
         /*
          * Traverse the DOM from node to node, starting at the node passed in.
diff --git a/src/Readability.php b/src/Readability.php
index c37c9b6..8f03fa1 100644
--- a/src/Readability.php
+++ b/src/Readability.php
@@ -5,6 +5,7 @@
 use fivefilters\Readability\Nodes\DOM\DOMDocument;
 use fivefilters\Readability\Nodes\DOM\DOMElement;
 use fivefilters\Readability\Nodes\DOM\DOMNode;
+use fivefilters\Readability\Nodes\DOM\DOMProcessingInstruction;
 use fivefilters\Readability\Nodes\DOM\DOMText;
 use fivefilters\Readability\Nodes\DOM\DOMComment;
 use fivefilters\Readability\Nodes\NodeUtility;
@@ -391,7 +392,7 @@ private function getJSONLD(DOMDocument $dom): array
                 return $metadata;
             } catch (\Exception $err) {
                 // The try-catch blocks are from the JS version. Not sure if there's anything
-                // here in the PHP version that would trigger an error or exception, so perhaps we can 
+                // here in the PHP version that would trigger an error or exception, so perhaps we can
                 // remove the try-catch blocks here (or at least translate errors to exceptions for this bit)
                 $this->logger->debug('[JSON-LD] Error parsing: ' . $err->getMessage());
             }
@@ -418,7 +419,7 @@ private function getMetadata(): void
             /* @var DOMNode $meta */
             $elementName = $meta->getAttribute('name');
             $elementProperty = $meta->getAttribute('property');
-            $content = $meta->getAttribute('content'); 
+            $content = $meta->getAttribute('content');
             $matches = null;
             $name = null;
 
@@ -960,7 +961,7 @@ private function textSimilarity(string $textA, string $textB): float
     /**
      * Checks if the node is a byline.
      */
-    private function checkByline(DOMNode|DOMText|DOMElement $node, string $matchString): bool
+    private function checkByline(DOMNode|DOMText|DOMElement|DOMProcessingInstruction $node, string $matchString): bool
     {
         if (!$this->configuration->getArticleByline()) {
             return false;
@@ -2043,7 +2044,7 @@ public function _cleanHeaders(DOMDocument $article): void
      * @param DOMNode the node to check.
      * @return boolean indicating whether this is a title-like header.
      */
-    private function headerDuplicatesTitle(DOMNode|DOMText|DOMElement $node): bool
+    private function headerDuplicatesTitle(DOMNode|DOMText|DOMElement|DOMProcessingInstruction $node): bool
     {
         if ($node->nodeName !== 'h1' && $node->nodeName !== 'h2') {
             return false;