Move class constant into Tag Processor, match with it, and add tests

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

@@ -106,7 +106,7 @@ public function current_node() {
 	 *
 	 * @see https://html.spec.whatwg.org/#has-an-element-in-the-specific-scope
 	 *
-	 * @param string   $tag_name         Name of tag check, or the class constant HEADING_ELEMENTS to specify H1-H6.
+	 * @param string   $tag_name         Name of tag check, or WP_HTML_Tag_Processor::H1_H6_ELEMENTS for any H1 - H6.
 	 * @param string[] $termination_list List of elements that terminate the search.
 	 * @return bool Whether the element was found in a specific scope.
 	 */
@@ -117,7 +117,7 @@ public function has_element_in_specific_scope( $tag_name, $termination_list ) {
 			}
 
 			if (
-				self::HEADING_ELEMENTS === $tag_name &&
+				WP_HTML_Tag_Processor::H1_H6_ELEMENTS === $tag_name &&
 				in_array( $node->node_name, array( 'H1', 'H2', 'H3', 'H4', 'H5', 'H6' ), true )
 			) {
 				return true;
@@ -271,15 +271,15 @@ public function pop() {
 	 * @see WP_HTML_Open_Elements::pop
 	 *
 	 * @param string $tag_name Name of tag that needs to be popped off of the stack of open elements,
-	 *                         or the class constant HEADING_ELEMENTS to specify any of H1-H6.
+	 *                         or WP_HTML_Tag_Processor::H1_H6_ELEMENTS to specify any of H1 - H6.
 	 * @return bool Whether a tag of the given name was found and popped off of the stack of open elements.
 	 */
 	public function pop_until( $tag_name ) {
 		foreach ( $this->walk_up() as $item ) {
 			$this->pop();
 
 			if (
-				self::HEADING_ELEMENTS === $tag_name &&
+				WP_HTML_Tag_Processor::H1_H6_ELEMENTS === $tag_name &&
 				in_array( $item->node_name, array( 'H1', 'H2', 'H3', 'H4', 'H5', 'H6' ), true )
 			) {
 				return true;
@@ -444,15 +444,4 @@ public function after_element_pop( $item ) {
 				break;
 		}
 	}
-
-	/**
-	 * Represents the collection of H1-H6 elements.
-	 *
-	 * @since 6.5.0
-	 *
-	 * @see has_element_in_scope()
-	 *
-	 * @var string
-	 */
-	const HEADING_ELEMENTS = 'heading-elements';
 }

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

@@ -694,11 +694,7 @@ private function step_in_body() {
 			case '-H4':
 			case '-H5':
 			case '-H6':
-				if (
-					! $this->state->stack_of_open_elements->has_element_in_scope(
-						WP_HTML_Open_Elements::HEADING_ELEMENTS
-					)
-				) {
+				if ( ! $this->state->stack_of_open_elements->has_element_in_scope( WP_HTML_Tag_Processor::H1_H6_ELEMENTS ) ) {
 					/*
 					 * This is a parse error; ignore the token.
 					 *
@@ -713,7 +709,7 @@ private function step_in_body() {
 					// @TODO: Record parse error: this error doesn't impact parsing.
 				}
 
-				$this->state->stack_of_open_elements->pop_until( WP_HTML_Open_Elements::HEADING_ELEMENTS );
+				$this->state->stack_of_open_elements->pop_until( WP_HTML_Tag_Processor::H1_H6_ELEMENTS );
 				return true;
 
 			/*

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

@@ -2401,6 +2401,7 @@ private function parse_query( $query ) {
 	 * Checks whether a given tag and its attributes match the search criteria.
 	 *
 	 * @since 6.2.0
+	 * @since 6.5.0 Allows matching any of the H1 - H6 tag names with the WP_HTML_Tag_Processor::H1_H6_ELEMENTS constant.
 	 *
 	 * @return bool Whether the given tag and its attribute match the search criteria.
 	 */
@@ -2409,8 +2410,31 @@ private function matches() {
 			return false;
 		}
 
+
 		// Does the tag name match the requested tag name in a case-insensitive manner?
-		if ( null !== $this->sought_tag_name ) {
+		if ( self::H1_H6_ELEMENTS === $this->sought_tag_name ) {
+			/*
+			 * H1 through H6 are special because they act like one tag
+			 * name but are distinct. It's common enough to want to stop
+			 * at any of them without knowing in advance which one to
+			 * look for; the class constant aids this by representing
+			 * the entire set of six elements: H1, H2, H3, H4, H5, H6.
+			 */
+
+			if ( 2 !== $this->tag_name_length ) {
+				return false;
+			}
+
+			$c = $this->html[ $this->tag_name_starts_at ];
+			if ( 'h' !== $c && 'H' !== $c ) {
+				return false;
+			}
+
+			$c = $this->html[ $this->tag_name_starts_at + 1 ];
+			if ( '1' !== $c && '2' !== $c && '3' !== $c && '4' !== $c && '5' !== $c && '6' !== $c ) {
+				return false;
+			}
+		} elseif ( null !== $this->sought_tag_name ) {
 			/*
 			 * String (byte) length lookup is fast. If they aren't the
 			 * same length then they can't be the same string values.
@@ -2447,4 +2471,17 @@ private function matches() {
 
 		return true;
 	}
+
+	// Class constants that would otherwise be distracting if found at the top of the document.
+
+	/**
+	 * Represents the collection of H1-H6 elements.
+	 *
+	 * @since 6.5.0
+	 *
+	 * @see has_element_in_scope()
+	 *
+	 * @var string
+	 */
+	const H1_H6_ELEMENTS = 'Match on any of H1, H2, H3, H4, H5, H6';
 }

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

@@ -509,6 +509,77 @@ public function test_next_tag_matches_decoded_class_names() {
 		$this->assertTrue( $p->next_tag( array( 'class_name' => '<egg>' ) ), 'Failed to find tag with HTML-encoded class name.' );
 	}
 
+	/**
+	 * Ensures that the H1_H6_ELEMENTS constant leads to matches for H1 through H6 elements.
+	 *
+	 * @ticket {TICKET_NUMBER}
+	 *
+	 * @covers WP_HTML_Tag_Processor::next_tag
+	 *
+	 * @dataProvider data_h_tag_names
+	 *
+	 * @param string $h_tag_name One of H1 through H6, case-insensitive.
+	 */
+	public function test_next_tag_matches_h1_through_h6_with_the_class_constant( $h_tag_name ) {
+		$p = new WP_HTML_Tag_Processor( "<div><img><{$h_tag_name}></div>" );
+
+		$this->assertTrue( $p->next_tag( WP_HTML_Tag_Processor::H1_H6_ELEMENTS ), "Failed to find {$h_tag_name} tag opener." );
+	}
+
+	/**
+	 * Data provider.
+	 *
+	 * @return array[]
+	 */
+	public function data_h_tag_names() {
+		return array(
+			'H1' => array( 'H1' ),
+			'H2' => array( 'H2' ),
+			'H3' => array( 'H3' ),
+			'H4' => array( 'H4' ),
+			'H5' => array( 'H5' ),
+			'H6' => array( 'H6' ),
+		);
+	}
+
+	/**
+	 * Ensures that the H1_H6_ELEMENTS constant doesn't lead to matches on
+	 * tag names that look similar to H1 - H6 but aren't those elements.
+	 *
+	 * @ticket {TICKET_NUMBER}
+	 *
+	 * @covers WP_HTML_Tag_Processor::next_tag
+	 *
+	 * @dataProvider data_invalid_h_tag_names
+	 *
+	 * @param string $invalid_h_tag_name Tag names that look like H1 through H6 but are not those tag names.
+	 */
+	public function test_next_tag_does_not_match_invalid_h_elements_with_the_class_constant( $invalid_h_tag_name ) {
+		$p = new WP_HTML_Tag_Processor( "<div><img><{$invalid_h_tag_name}></div>" );
+
+		$this->assertFalse( $p->next_tag( WP_HTML_Tag_Processor::H1_H6_ELEMENTS ), "Found {$p->get_tag()} when looking for {$invalid_h_tag_name} element and should have found nothing." );
+	}
+
+	/**
+	 * Data provider.
+	 *
+	 * @return array[].
+	 */
+	public function data_invalid_h_tag_names() {
+		return array(
+			'H0'        => array( 'H0' ),
+			'H7'        => array( 'H7' ),
+			'H13'       => array( 'H13' ),
+
+			/*
+			 * Preserve the FULLWIDTH DIGIT SIX key because PHPUnit interprets '6' as
+			 * a numeric array item and reports "data set 0" instead of "6".
+			 */
+			'６'         => array( '6' ),
+			'H4-CUSTOM' => array( 'H4-CUSTOM' ),
+		);
+	}
+
 	/**
 	 * @ticket 56299
 	 * @ticket 57852