From 763ce8802121e41cfd981439186ebdf7d092c581 Mon Sep 17 00:00:00 2001
From: Ed Sanders <esanders@wikimedia.org>
Date: Mon, 16 Nov 2020 00:13:55 +0000
Subject: [PATCH] Base TreeWalker implementation on PHPDOM

For consistency with other DOM implementations.

Change-Id: I20447d880ccd3b70b6694b36ea2f63dd0c42fa84
---
 includes/TreeWalker.php | 166 +++++++++++++++++++++-------------------
 1 file changed, 89 insertions(+), 77 deletions(-)

diff --git a/includes/TreeWalker.php b/includes/TreeWalker.php
index e546c9b81..b10984a80 100644
--- a/includes/TreeWalker.php
+++ b/includes/TreeWalker.php
@@ -4,6 +4,7 @@ namespace MediaWiki\Extension\DiscussionTools;
 
 use DOMNode;
 use Exception;
+use Throwable;
 
 /**
  * Partial implementation of W3 DOM4 TreeWalker interface.
@@ -11,7 +12,7 @@ use Exception;
  * See also:
  * - https://dom.spec.whatwg.org/#interface-treewalker
  *
- * Adapted from https://github.com/Krinkle/dom-TreeWalker-polyfill/blob/master/src/TreeWalker-polyfill.js
+ * Ported from https://github.com/TRowbotham/PHPDOM (MIT)
  */
 class TreeWalker {
 
@@ -20,83 +21,24 @@ class TreeWalker {
 	public $currentNode;
 	public $filter;
 
-	/**
-	 * See https://dom.spec.whatwg.org/#concept-node-filter
-	 *
-	 * @param TreeWalker $tw
-	 * @param DOMNode $node
-	 * @return int Constant NodeFilter::FILTER_ACCEPT,
-	 *  NodeFilter::FILTER_REJECT or NodeFilter::FILTER_SKIP.
-	 */
-	private function nodeFilter( TreeWalker $tw, DOMNode $node ) {
-		// Maps nodeType to whatToShow
-		if ( !( ( ( 1 << ( $node->nodeType - 1 ) ) & $tw->whatToShow ) ) ) {
-			return NodeFilter::FILTER_SKIP;
-		}
-
-		if ( $tw->filter === null ) {
-			return NodeFilter::FILTER_ACCEPT;
-		}
-
-		return $tw->filter->acceptNode( $node );
-	}
-
-	/**
-	 * Based on WebKit's NodeTraversal::nextSkippingChildren
-	 * https://trac.webkit.org/browser/trunk/Source/WebCore/dom/NodeTraversal.h?rev=137221#L103
-	 *
-	 * @param DOMNode $node
-	 * @param DOMNode $stayWithin
-	 * @return DOMNode|null
-	 */
-	private function nextSkippingChildren( DOMNode $node, DOMNode $stayWithin ) : ?DOMNode {
-		if ( $node === $stayWithin ) {
-			return null;
-		}
-		if ( $node->nextSibling !== null ) {
-			return $node->nextSibling;
-		}
-
-		/**
-		 * Based on WebKit's NodeTraversal::nextAncestorSibling
-		 * https://trac.webkit.org/browser/trunk/Source/WebCore/dom/NodeTraversal.cpp?rev=137221#L43
-		 */
-		while ( $node->parentNode !== null ) {
-			$node = $node->parentNode;
-			if ( $node === $stayWithin ) {
-				return null;
-			}
-			if ( $node->nextSibling !== null ) {
-				return $node->nextSibling;
-			}
-		}
-		return null;
-	}
+	private $isActive = false;
 
 	/**
 	 * See https://dom.spec.whatwg.org/#interface-treewalker
 	 *
 	 * @param DOMNode $root
-	 * @param int|null $whatToShow
+	 * @param int $whatToShow
 	 * @param callable|null $filter
-	 * @throws Exception
 	 */
-	public function __construct( DOMNode $root, $whatToShow = null, callable $filter = null ) {
-		if ( !$root->nodeType ) {
-			throw new Exception( 'DOMException: NOT_SUPPORTED_ERR' );
-		}
-
-		$this->root = $root;
-		$this->whatToShow = (int)$whatToShow ?: 0;
-
+	public function __construct(
+		DOMNode $root,
+		int $whatToShow = NodeFilter::SHOW_ALL,
+		callable $filter = null
+	) {
 		$this->currentNode = $root;
-
-		if ( !$filter ) {
-			$this->filter = null;
-		} else {
-			$this->filter = new NodeFilter();
-			$this->filter->filter = $filter;
-		}
+		$this->filter = $filter;
+		$this->root = $root;
+		$this->whatToShow = $whatToShow;
 	}
 
 	/**
@@ -111,23 +53,93 @@ class TreeWalker {
 		while ( true ) {
 			while ( $result !== NodeFilter::FILTER_REJECT && $node->firstChild !== null ) {
 				$node = $node->firstChild;
-				$result = $this->nodeFilter( $this, $node );
+				$result = $this->filterNode( $node );
 				if ( $result === NodeFilter::FILTER_ACCEPT ) {
 					$this->currentNode = $node;
 					return $node;
 				}
 			}
-			$following = $this->nextSkippingChildren( $node, $this->root );
-			if ( $following !== null ) {
-				$node = $following;
-			} else {
-				return null;
+
+			$sibling = null;
+			$temp = $node;
+			while ( $temp !== null ) {
+				if ( $temp === $this->root ) {
+					return null;
+				}
+
+				$sibling = $temp->nextSibling;
+
+				if ( $sibling !== null ) {
+					$node = $sibling;
+
+					break;
+				}
+
+				$temp = $temp->parentNode;
 			}
-			$result = $this->nodeFilter( $this, $node );
+
+			$result = $this->filterNode( $node );
+
 			if ( $result === NodeFilter::FILTER_ACCEPT ) {
 				$this->currentNode = $node;
+
 				return $node;
 			}
+
 		}
 	}
+
+	/**
+	 * Filters a node.
+	 *
+	 * @internal
+	 *
+	 * @see https://dom.spec.whatwg.org/#concept-node-filter
+	 *
+	 * @param DOMNode $node The node to check.
+	 * @return int Returns one of NodeFilter's FILTER_* constants.
+	 *     - NodeFilter::FILTER_ACCEPT
+	 *     - NodeFilter::FILTER_REJECT
+	 *     - NodeFilter::FILTER_SKIP
+	 * @throws Exception
+	 */
+	private function filterNode( DOMNode $node ): int {
+		if ( $this->isActive ) {
+			throw new Exception( 'InvalidStateError' );
+		}
+
+		// Let n be node’s nodeType attribute value minus 1.
+		$n = $node->nodeType - 1;
+
+		// If the nth bit (where 0 is the least significant bit) of whatToShow
+		// is not set, return FILTER_SKIP.
+		if ( !( ( 1 << $n ) & $this->whatToShow ) ) {
+			return NodeFilter::FILTER_SKIP;
+		}
+
+		// If filter is null, return FILTER_ACCEPT.
+		if ( !$this->filter ) {
+			return NodeFilter::FILTER_ACCEPT;
+		}
+
+		$this->isActive = true;
+
+		try {
+			// Let $result be the return value of call a user object's operation
+			// with traverser's filter, "acceptNode", and Node. If this throws
+			// an exception, then unset traverser's active flag and rethrow the
+			// exception.
+			$result = $this->filter instanceof NodeFilter
+				? $this->filter->acceptNode( $node )
+				: ( $this->filter )( $node );
+		} catch ( Throwable $e ) {
+			$this->isActive = false;
+
+			throw $e;
+		}
+
+		$this->isActive = false;
+
+		return $result;
+	}
 }