HTML API: Queue virtual closers after non-current removals

Author: sirreal

src/wp-includes/html-api/class-wp-html-open-elements.php

@@ -588,7 +588,7 @@ public function remove_node( WP_HTML_Token $token ): bool {
 
 			$position_from_start = $this->count() - $position_from_end - 1;
 			array_splice( $this->stack, $position_from_start, 1 );
-			$this->after_element_pop( $item );
+			$this->after_element_pop( $item, 0 === $position_from_end );
 			return true;
 		}
 
@@ -731,9 +731,10 @@ public function after_element_push( WP_HTML_Token $item ): void {
 	 *
 	 * @since 6.4.0
 	 *
-	 * @param WP_HTML_Token $item Element that was removed from the stack of open elements.
+	 * @param WP_HTML_Token $item               Element that was removed from the stack of open elements.
+	 * @param bool          $invoke_pop_handler Whether to call the pop handler.
 	 */
-	public function after_element_pop( WP_HTML_Token $item ): void {
+	public function after_element_pop( WP_HTML_Token $item, bool $invoke_pop_handler = true ): void {
 		/*
 		 * When adding support for new elements, expand this switch to trap
 		 * cases where the precalculated value needs to change.
@@ -767,7 +768,7 @@ public function after_element_pop( WP_HTML_Token $item ): void {
 				break;
 		}
 
-		if ( null !== $this->pop_handler ) {
+		if ( $invoke_pop_handler && null !== $this->pop_handler ) {
 			call_user_func( $this->pop_handler, $item );
 		}
 	}

src/wp-includes/html-api/class-wp-html-processor.php

@@ -251,6 +251,15 @@ class WP_HTML_Processor extends WP_HTML_Tag_Processor {
 	 */
 	private $current_element = null;
 
+	/**
+	 * Elements removed from the stack of open elements without a normal pop event.
+	 *
+	 * @since 7.1.0
+	 *
+	 * @var array[]
+	 */
+	private $non_lifo_breadcrumb_removals = array();
+
 	/**
 	 * Context node if created as a fragment parser.
 	 *
@@ -814,6 +823,10 @@ private function next_visitable_token(): bool {
 		 *       tokens works in the meantime and isn't obviously wrong.
 		 */
 		if ( empty( $this->element_queue ) ) {
+			if ( $this->queue_virtual_closer_after_non_lifo_removal() ) {
+				return $this->next_visitable_token();
+			}
+
 			if ( $this->step() ) {
 				return $this->next_visitable_token();
 			}
@@ -823,6 +836,10 @@ private function next_visitable_token(): bool {
 			}
 		}
 
+		if ( $this->queue_virtual_closer_after_non_lifo_removal() ) {
+			return $this->next_visitable_token();
+		}
+
 		// Process the next event on the queue.
 		$this->current_element = array_shift( $this->element_queue );
 		if ( ! isset( $this->current_element ) ) {
@@ -860,6 +877,61 @@ private function next_visitable_token(): bool {
 		return true;
 	}
 
+	/**
+	 * Queues a virtual closer for a removed node once its subtree closes.
+	 *
+	 * Non-LIFO removals from the stack of open elements do not emit a normal
+	 * pop event because those events blindly pop the current breadcrumb. The
+	 * removed node remains an ancestor of the currently open subtree, but must
+	 * be reported as a virtual closer before visiting the next token after
+	 * that subtree closes.
+	 *
+	 * @since 7.1.0
+	 *
+	 * @return bool Whether a virtual closer was queued.
+	 */
+	private function queue_virtual_closer_after_non_lifo_removal(): bool {
+		if ( empty( $this->non_lifo_breadcrumb_removals ) ) {
+			return false;
+		}
+
+		$removed_node     = end( $this->non_lifo_breadcrumb_removals );
+		$removed_token    = $removed_node['token'];
+		$breadcrumb_depth = $removed_node['breadcrumb_depth'];
+
+		if (
+			count( $this->breadcrumbs ) !== $breadcrumb_depth ||
+			empty( $this->breadcrumbs ) ||
+			end( $this->breadcrumbs ) !== $removed_token->node_name
+		) {
+			return false;
+		}
+
+		// At EOF, normal stack pops may be queued and processed after the stack is empty.
+		$adjusted_current_node = $this->get_adjusted_current_node();
+
+		if ( isset( $adjusted_current_node ) && end( $this->breadcrumbs ) === $adjusted_current_node->node_name ) {
+			return false;
+		}
+
+		$next_event = reset( $this->element_queue );
+		if (
+			false !== $next_event &&
+			WP_HTML_Stack_Event::POP === $next_event->operation &&
+			$next_event->token !== $removed_token &&
+			$next_event->token->node_name === $removed_token->node_name
+		) {
+			return false;
+		}
+
+		array_pop( $this->non_lifo_breadcrumb_removals );
+		array_unshift(
+			$this->element_queue,
+			new WP_HTML_Stack_Event( $removed_token, WP_HTML_Stack_Event::POP, 'virtual' )
+		);
+		return true;
+	}
+
 	/**
 	 * Indicates if the current tag token is a tag closer.
 	 *
@@ -2848,7 +2920,18 @@ private function step_in_body(): bool {
 						case 'A':
 							$this->run_adoption_agency_algorithm();
 							$this->state->active_formatting_elements->remove_node( $item );
-							$this->state->stack_of_open_elements->remove_node( $item );
+							$is_current_node = $item === $this->state->stack_of_open_elements->current_node();
+							if ( $this->state->stack_of_open_elements->remove_node( $item ) && ! $is_current_node ) {
+								$breadcrumb_depth = count( $this->breadcrumbs );
+								while ( 0 < $breadcrumb_depth && $this->breadcrumbs[ $breadcrumb_depth - 1 ] !== $item->node_name ) {
+									--$breadcrumb_depth;
+								}
+
+								$this->non_lifo_breadcrumb_removals[] = array(
+									'token'            => $item,
+									'breadcrumb_depth' => $breadcrumb_depth,
+								);
+							}
 							break 2;
 					}
 				}
@@ -5675,6 +5758,7 @@ public function seek( $bookmark_name ): bool {
 			$this->state->current_token                     = null;
 			$this->current_element                          = null;
 			$this->element_queue                            = array();
+			$this->non_lifo_breadcrumb_removals             = array();
 
 			/*
 			 * The absence of a context node indicates a full parse.