HTML API: Abort on text nodes requiring reconstruction

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

@@ -430,8 +430,11 @@ public function next_tag( $query = null ) {
 	public function next_token() {
 		$found_a_token = parent::next_token();
 
-		if ( '#tag' === $this->get_token_type() ) {
-			$this->step( self::PROCESS_CURRENT_NODE );
+		switch ( $this->get_token_type() ) {
+			case '#tag':
+			case '#text':
+				$this->step( self::PROCESS_CURRENT_NODE );
+				break;
 		}
 
 		return $found_a_token;
@@ -536,6 +539,11 @@ public function step( $node_to_process = self::PROCESS_NEXT_NODE ) {
 
 		if ( self::PROCESS_NEXT_NODE === $node_to_process ) {
 			while ( parent::next_token() && '#tag' !== $this->get_token_type() ) {
+				if ( '#text' === $this->get_token_type() && $this->has_active_formats_needing_reconstruction() ) {
+					$this->last_error = self::ERROR_UNSUPPORTED;
+					return false;
+				}
+
 				continue;
 			}
 		}
@@ -1498,6 +1506,44 @@ private function reconstruct_active_formatting_elements() {
 		throw new WP_HTML_Unsupported_Exception( 'Cannot reconstruct active formatting elements when advancing and rewinding is required.' );
 	}
 
+	/**
+	 * Indicates if there are active formatting elements needing reconstruction.
+	 *
+	 * @since 6.5.0
+	 *
+	 * @return bool False if reconstruction would definitely not create any new elements.
+	 */
+	private function has_active_formats_needing_reconstruction() {
+		/*
+		 * > If there are no entries in the list of active formatting elements, then there is nothing
+		 * > to reconstruct; stop this algorithm.
+		 */
+		if ( 0 === $this->state->active_formatting_elements->count() ) {
+			return false;
+		}
+
+		$last_entry = $this->state->active_formatting_elements->current_node();
+		if (
+
+			/*
+			 * > If the last (most recently added) entry in the list of active formatting elements is a marker;
+			 * > stop this algorithm.
+			 */
+			'marker' === $last_entry->node_name ||
+
+			/*
+			 * > If the last (most recently added) entry in the list of active formatting elements is an
+			 * > element that is in the stack of open elements, then there is nothing to reconstruct;
+			 * > stop this algorithm.
+			 */
+			$this->state->stack_of_open_elements->contains_node( $last_entry )
+		) {
+			return false;
+		}
+
+		return true;
+	}
+
 	/**
 	 * Runs the adoption agency algorithm.
 	 *

tests/phpunit/tests/html-api/wpHtmlSupportRequiredActiveFormatReconstruction.php

@@ -0,0 +1,71 @@
+<?php
+/**
+ * Unit tests for the HTML API ensuring proper handling of behaviors related to
+ * active format reconstruction.
+ *
+ * @package WordPress
+ * @subpackage HTML-API
+ *
+ * @since 6.5.0
+ *
+ * @group html-api
+ *
+ * @coversDefaultClass WP_HTML_Processor
+ */
+class Tests_HtmlApi_WpHtmlSupportRequiredActiveFormatReconstruction extends WP_UnitTestCase {
+	/**
+	 * Ensures that active formats are properly reconstructed when visiting text nodes,
+	 * verifying that the proper breadcrumbs are maintained when scanning through HTML.
+	 *
+	 * @ticket 60455
+	 */
+	public function test_reconstructs_active_formats_on_text_nodes() {
+		$processor = WP_HTML_Processor::create_fragment( '<p><b>One<p><source>Two<source>' );
+
+		// The SOURCE element doesn't trigger reconstruction, and this test asserts that.
+		$this->assertTrue(
+			$processor->next_tag( 'SOURCE' ),
+			'Should have found the first custom element.'
+		);
+
+		$this->assertSame(
+			array( 'HTML', 'BODY', 'P', 'SOURCE' ),
+			$processor->get_breadcrumbs(),
+			'Should have closed formatting element at first P element.'
+		);
+
+		/*
+		 * There are two ways this test could fail. One is to appropriately find the
+		 * second text node but fail to reconstruct the implicitly-closed B element.
+		 * The other way is to fail to abort when encountering the second text node
+		 * because the kind of active format reconstruction isn't supported.
+		 *
+		 * At the time of writing this test, the HTML Processor bails whenever it
+		 * needs to reconstruct active formats, unless there are no active formats.
+		 * To ensure that this test properly works once that support is expanded,
+		 * it's written to verify both circumstances. Once support is added, this
+		 * can be simplified to only contain the first clause of the conditional.
+		 *
+		 * The use of the SOURCE element is important here because most elements
+		 * will also trigger reconstruction, which would conflate the test results
+		 * with the text node triggering reconstruction. The SOURCE element won't
+		 * do this, making it neutral. Therefore, the implicitly-closed B element
+		 * will only be reconstructed by the text node.
+		 */
+
+		if ( $processor->next_tag( 'SOURCE' ) ) {
+			echo "\e[32mSOURCE\e[m\n";
+			$this->assertSame(
+				array( 'HTML', 'BODY', 'P', 'B', 'SOURCE' ),
+				$processor->get_breadcrumbs(),
+				'Should have reconstructed the implicitly-closed B element.'
+			);
+		} else {
+			$this->assertSame(
+				WP_HTML_Processor::ERROR_UNSUPPORTED,
+				$processor->get_last_error(),
+				'Should have aborted for incomplete active format reconstruction when encountering the second text node.'
+			);
+		}
+	}
+}