Fix #2698: Prevent unnecessary cache purges for hierarchical pages

inc/common/purge.php

@@ -131,11 +131,15 @@ function rocket_get_purge_urls( $post_id, $post ) {
 			$purge_urls[] = $author_url;
 		}
 
-		// Add all parents.
-		$parents = get_post_ancestors( $post_id );
-		if ( (bool) $parents ) {
-			foreach ( $parents as $parent_id ) {
-				$purge_urls[] = get_permalink( $parent_id );
+		$post_name = get_post_field( 'post_name', $post_id );
+		// If the slug has changed, add all children pages with the new slug.
+		if ( $post_name !== $post->post_name ) {
+			// Add all children.
+			$children = rocket_get_all_descendant( $post_id );
+			if ( (bool) $children ) {
+				foreach ( $children as $child_id ) {
+					$purge_urls[] = get_permalink( $child_id );
+				}
 			}
 		}
 
@@ -146,6 +150,27 @@ function rocket_get_purge_urls( $post_id, $post ) {
 	}
 }
 
+/**
+ * Recursively retrieves all descendant page IDs for a given parent page ID.
+ *
+ * @param int $parent_id The ID of the parent page.
+ * @return int[] A flat array with all descendant page IDs (children, grandchildren, etc.).
+ */
+function rocket_get_all_descendant( $parent_id ) {
+	// 'child_of' retrieves all descendants, not just direct children.
+	$all_descendants = get_pages(
+		[ 'child_of' => $parent_id ]
+	);
+
+	if ( empty( $all_descendants ) ) {
+		return [];
+	}
+
+	// Convert the array of page objects to an array of IDs only.
+	return wp_list_pluck( $all_descendants, 'ID' );
+}
+
+
 /**
  * Update cache when a post is updated or commented
  *
@@ -636,6 +661,20 @@ function rocket_clean_post_cache_on_slug_change( $post_id, $post_data ) {
 	if ( empty( $post_name ) ) {
 		return;
 	}
-	rocket_clean_files( get_the_permalink( $post_id ) );
+
+	$purge_urls   = [];
+	$purge_urls[] = get_the_permalink( $post_id );
+
+	// Clear cache for all child pages.
+	$children = rocket_get_all_descendant( $post_id );
+	if ( (bool) $children ) {
+		foreach ( $children as $child_id ) {
+			$purge_urls[] = get_the_permalink( $child_id );
+		}
+	}
+
+	rocket_clean_files( $purge_urls );
+
 }
 add_action( 'pre_post_update', 'rocket_clean_post_cache_on_slug_change', PHP_INT_MAX, 2 );
+

inc/functions/files.php

@@ -530,6 +530,32 @@ function rocket_maybe_find_right_trash_url( array $parsed_url, int $post_id ) {
 	return get_rocket_parse_url( $new_permalink );
 }
 
+
+/**
+ * Count the number of path segments in a URL.
+ *
+ * Parses the given URL, extracts its path component, trims leading/trailing
+ * slashes, and counts the remaining segments separated by '/'.
+ *
+ * Examples:
+ * - "https://example.com/" => 0
+ * - "https://example.com/a/b/c" => 3
+ * - "/a/b/" => 2
+ *
+ * @param string $url The URL (absolute or relative) to analyze.
+ * @return int Number of segments in the URL path (0 if empty or no path).
+ */
+function rocket_count_path_segments( string $url ): int {
+	$path = wp_parse_url( $url, PHP_URL_PATH ) ?? '';
+	$path = trim( $path, '/' );              // "/" -> ""
+
+	if ( '' === $path ) {
+		return 0;
+	}
+
+	return count( explode( '/', $path ) );   // "a/b/c" -> 3
+}
+
 /**
  * Delete one or several cache files.
  *
@@ -563,6 +589,23 @@ function rocket_clean_files( $urls, $filesystem = null, $run_actions = true ) {
 		$filesystem = rocket_direct_filesystem();
 	}
 
+	// Sort: most segments first (deepest URLs first).
+	usort(
+		$urls,
+		static function ( $a, $b ) {
+			$da = rocket_count_path_segments( (string) $a );
+			$db = rocket_count_path_segments( (string) $b );
+
+			// First by depth (descending).
+			if ( $da !== $db ) {
+				return $db <=> $da;
+			}
+
+			// If depth is equal, alphabetically.
+			return strcmp( (string) $a, (string) $b );
+		}
+	);
+
 	if ( $run_actions ) {
 		/**
 		 * Fires before all cache files are deleted.
@@ -575,6 +618,7 @@ function rocket_clean_files( $urls, $filesystem = null, $run_actions = true ) {
 	}
 
 	foreach ( $urls as $url_key => $url ) {
+
 		if ( $run_actions ) {
 			/**
 			 * Fires before the cache file is deleted.
@@ -591,55 +635,85 @@ function rocket_clean_files( $urls, $filesystem = null, $run_actions = true ) {
 
 		$parsed_url = get_rocket_parse_url( $url );
 
-		if ( ! empty( $parsed_url['host'] ) ) {
-			foreach ( _rocket_get_cache_dirs( $parsed_url['host'], $cache_path ) as $dir ) {
-				// Decode url path.
-				$url_chunks = explode( '/', $parsed_url['path'] );
-				$matches    = preg_grep( '/%/', $url_chunks );
+		if ( empty( $parsed_url['host'] ) ) {
+			continue;
+		}
 
-				if ( ! empty( $matches ) ) {
-					$parsed_url['path'] = rawurldecode( $parsed_url['path'] );
-				}
+		foreach ( _rocket_get_cache_dirs( $parsed_url['host'], $cache_path ) as $dir ) {
+			// Decode url path.
+			$url_chunks = explode( '/', $parsed_url['path'] );
+			$matches    = preg_grep( '/%/', $url_chunks );
+
+			if ( ! empty( $matches ) ) {
+				$parsed_url['path'] = rawurldecode( $parsed_url['path'] );
+			}
 
-				// Encode Non-latin characters if found in url path.
-				if ( false !== preg_match_all( '/(?<non_latin>[^\x00-\x7F]+)/', $parsed_url['path'], $matches ) ) {
-					$cb_encode_non_latin = function ( $non_latin ) {
-						return strtolower( rawurlencode( $non_latin ) );
-					};
+			// Encode Non-latin characters if found in url path.
+			if ( false !== preg_match_all( '/(?<non_latin>[^\x00-\x7F]+)/', $parsed_url['path'], $matches ) ) {
+				$cb_encode_non_latin = function ( $non_latin ) {
+					return strtolower( rawurlencode( $non_latin ) );
+				};
 
-					$parsed_url['path'] = str_replace( $matches['non_latin'], array_map( $cb_encode_non_latin, $matches['non_latin'] ), $parsed_url['path'] );
-				}
+				$parsed_url['path'] = str_replace( $matches['non_latin'], array_map( $cb_encode_non_latin, $matches['non_latin'] ), $parsed_url['path'] );
+			}
 
-				$entry = $dir . $parsed_url['path'];
-
-				// For regex we use it for file names only, and it should include the * character.
-				if ( str_contains( $entry, '*' ) ) {
-					$regex_part    = basename( $entry );
-					$search_dir    = str_replace( $regex_part, '', $entry );
-					$matched_files = _rocket_get_dir_files_by_regex( $search_dir, '#' . $regex_part . '#i' );
-					foreach ( $matched_files as $item ) {
-						$current_file = $item->getPath() . DIRECTORY_SEPARATOR . $item->getFilename();
-						if ( $filesystem->exists( $current_file ) ) {
-							$filesystem->delete( $current_file );
-						}
+			$entry = $dir . $parsed_url['path'];
+
+			// For regex we use it for file names only, and it should include the * character.
+			if ( str_contains( $entry, '*' ) ) {
+				$regex_part    = basename( $entry );
+				$search_dir    = str_replace( $regex_part, '', $entry );
+				$matched_files = _rocket_get_dir_files_by_regex( $search_dir, '#' . $regex_part . '#i' );
+				foreach ( $matched_files as $item ) {
+					$current_file = $item->getPath() . DIRECTORY_SEPARATOR . $item->getFilename();
+					if ( $filesystem->exists( $current_file ) ) {
+						$filesystem->delete( $current_file );
 					}
-					// Remove the regex part from the url.
-					$url              = str_replace( $regex_part, '', $url );
-					$urls[ $url_key ] = $url;
 				}
+				// Remove the regex part from the url.
+				$url              = str_replace( $regex_part, '', $url );
+				$urls[ $url_key ] = $url;
+			}
+
+			// Skip if the dir/file does not exist.
+			if ( ! $filesystem->exists( $entry ) ) {
+				continue;
+			}
 
-				// Skip if the dir/file does not exist.
-				if ( ! $filesystem->exists( $entry ) ) {
-					continue;
+			if ( ! $filesystem->is_dir( $entry ) ) {
+				$filesystem->delete( $entry );
+				continue;
+			}
+
+			// Check whether the directory contains subfolders (vs only files).
+			$has_subdirs = false;
+			try {
+				foreach ( new FilesystemIterator( $entry, FilesystemIterator::SKIP_DOTS ) as $child ) {
+					if ( $child->isDir() ) {
+						$has_subdirs = true;
+						break;
+					}
 				}
+			} catch ( Exception $e ) {
+				// If we can't inspect the directory, be conservative and only delete files.
+				$has_subdirs = true;
+			}
 
-				if ( $filesystem->is_dir( $entry ) ) {
-					rocket_rrmdir( $entry, [], $filesystem );
-				} else {
-					$filesystem->delete( $entry );
+			if ( ! $has_subdirs ) {
+				// Directory contains only files: remove it entirely.
+				rocket_rrmdir( $entry, [], $filesystem );
+				continue;
+			}
+
+			// Directory contains subfolders: delete only the files in the top-level.
+			foreach ( _rocket_get_dir_files_by_regex( $entry, '#.+#' ) as $child ) {
+				if ( $child->isFile() ) {
+					$filesystem->delete( $child->getPathname() );
 				}
 			}
+
 		}
+
 		if ( $run_actions ) {
 			/**
 			 * Fires after the cache file is deleted.