Fix test, adjust HTML Processor.

Author: dmsnell

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

@@ -361,7 +361,7 @@ public function get_last_error() {
 	public function next_tag( $query = null ) {
 		if ( null === $query ) {
 			while ( $this->step() ) {
-				if ( ! $this->is_tag_closer() ) {
+				if ( '#tag' === $this->get_token_type() && ! $this->is_tag_closer() ) {
 					return true;
 				}
 			}
@@ -384,7 +384,7 @@ public function next_tag( $query = null ) {
 
 		if ( ! ( array_key_exists( 'breadcrumbs', $query ) && is_array( $query['breadcrumbs'] ) ) ) {
 			while ( $this->step() ) {
-				if ( ! $this->is_tag_closer() ) {
+				if ( '#tag' === $this->get_token_type() && ! $this->is_tag_closer() ) {
 					return true;
 				}
 			}
@@ -428,13 +428,7 @@ public function next_tag( $query = null ) {
 	 * @return bool
 	 */
 	public function next_token() {
-		$found_a_token = parent::next_token();
-
-		if ( '#tag' === $this->get_token_type() ) {
-			$this->step( self::PROCESS_CURRENT_NODE );
-		}
-
-		return $found_a_token;
+		return $this->step();
 	}
 
 	/**
@@ -535,13 +529,25 @@ 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() ) {
-				continue;
+			/*
+			 * Currently tag and text nodes must be processed. Text nodes may
+			 * trigger active format reconstruction, otherwise they could be
+			 * skipped until the HTML Processor supports visiting text nodes.
+			 */
+			while ( parent::next_token() ) {
+				$token_type = $this->get_token_type();
+
+				if ( '#tag' === $token_type || '#text' === $token_type ) {
+					break;
+				}
 			}
 		}
 
 		// Finish stepping when there are no more tokens in the document.
-		if ( null === $this->get_tag() ) {
+		if (
+			WP_HTML_Tag_Processor::STATE_COMPLETE === $this->parser_state ||
+			WP_HTML_Tag_Processor::STATE_INCOMPLETE_INPUT === $this->parser_state
+		) {
 			return false;
 		}
 
@@ -1194,10 +1200,6 @@ private function step_in_body() {
 	 * @return string|false Name of created bookmark, or false if unable to create.
 	 */
 	private function bookmark_tag() {
-		if ( ! $this->get_tag() ) {
-			return false;
-		}
-
 		if ( ! parent::set_bookmark( ++$this->bookmark_counter ) ) {
 			$this->last_error = self::ERROR_EXCEEDED_MAX_BOOKMARKS;
 			throw new Exception( 'could not allocate bookmark' );

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

@@ -20,18 +20,18 @@ class Tests_HtmlApi_WpHtmlSupportRequiredActiveFormatReconstruction extends WP_U
 	 * @ticket 60455
 	 */
 	public function test_reconstructs_active_formats_on_text_nodes() {
-		$processor = WP_HTML_Processor::create_fragment( '<p><b>One<span><p>Two<an-element>' );
+		$processor = WP_HTML_Processor::create_fragment( '<p><b>One<p><source>Two<source>' );
 
-		$processor->next_tag( 'span' );
-		$this->assertSame(
-			array( 'HTML', 'BODY', 'P', 'B', 'SPAN' ),
-			$processor->get_breadcrumbs(),
-			'Should have identified the stack of open elements for the first text node.'
+		// 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->assertTrue(
-			$processor->next_tag( 'p' ),
-			'Should have found second P element.'
+		$this->assertSame(
+			array( 'HTML', 'BODY', 'P', 'SOURCE' ),
+			$processor->get_breadcrumbs(),
+			'Should have closed formatting element at first P element.'
 		);
 
 		/*
@@ -45,11 +45,18 @@ public function test_reconstructs_active_formats_on_text_nodes() {
 		 * 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( 'AN-ELEMENT' ) ) {
+		if ( $processor->next_tag( 'SOURCE' ) ) {
+			echo "\e[32mSOURCE\e[m\n";
 			$this->assertSame(
-				array( 'HTML', 'BODY', 'P', 'B', 'AN-ELEMENT' ),
+				array( 'HTML', 'BODY', 'P', 'B', 'SOURCE' ),
 				$processor->get_breadcrumbs(),
 				'Should have reconstructed the implicitly-closed B element.'
 			);