HTML API: Add spawn fragment parser method

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

@@ -422,6 +422,61 @@ function ( WP_HTML_Token $token ): void {
 		};
 	}
 
+	/**
+	 * Creates a fragment processor with the current node as its context element.
+	 *
+	 * @see https://html.spec.whatwg.org/multipage/parsing.html#html-fragment-parsing-algorithm
+	 *
+	 * @param string $html     Input HTML fragment to process.
+	 * @return static|null     The created processor if successful, otherwise null.
+	 */
+	public function spawn_fragment_parser( string $html ): ?self {
+		if ( $this->get_token_type() !== '#tag' ) {
+			return null;
+		}
+
+		$namespace = $this->get_namespace();
+
+		/*
+		 * Prevent creating fragments at "self-contained" nodes.
+		 *
+		 * @see https://github.com/WordPress/wordpress-develop/pull/7141
+		 * @see https://github.com/WordPress/wordpress-develop/pull/7198
+		 */
+		if (
+			'html' === $namespace &&
+			in_array( $this->get_tag(), array( 'IFRAME', 'NOEMBED', 'NOFRAMES', 'SCRIPT', 'STYLE', 'TEXTAREA', 'TITLE', 'XMP' ), true )
+		) {
+			return null;
+		}
+
+		$fragment_processor              = self::create_fragment( $html );
+		$fragment_processor->compat_mode = $this->compat_mode;
+
+		$fragment_processor->context_node                = clone $this->state->current_token;
+		$fragment_processor->context_node->bookmark_name = 'context-node';
+		$fragment_processor->context_node->on_destroy    = null;
+
+		$context_element = array( $fragment_processor->context_node->node_name, array() );
+		foreach ( $this->get_attribute_names_with_prefix( '' ) as $name => $value ) {
+			$context_element[1][ $name ] = $value;
+		}
+
+		$fragment_processor->breadcrumbs = array();
+
+		if ( 'TEMPLATE' === $context_element[0] ) {
+			$fragment_processor->state->stack_of_template_insertion_modes[] = WP_HTML_Processor_State::INSERTION_MODE_IN_TEMPLATE;
+		}
+
+		$fragment_processor->reset_insertion_mode_appropriately();
+
+		// @todo Set the parser's form element pointer.
+
+		$fragment_processor->state->encoding_confidence = 'irrelevant';
+
+		return $fragment_processor;
+	}
+
 	/**
 	 * Stops the parser and terminates its execution when encountering unsupported markup.
 	 *
@@ -4501,7 +4556,7 @@ private function step_in_foreign_content(): bool {
 
 					$this->state->stack_of_open_elements->pop();
 				}
-				return $this->step( self::REPROCESS_CURRENT_NODE );
+				goto in_foreign_content_process_in_current_insertion_mode;
 		}
 
 		/*
@@ -4577,6 +4632,7 @@ private function step_in_foreign_content(): bool {
 				goto in_foreign_content_end_tag_loop;
 			}
 
+			in_foreign_content_process_in_current_insertion_mode:
 			switch ( $this->state->insertion_mode ) {
 				case WP_HTML_Processor_State::INSERTION_MODE_INITIAL:
 					return $this->step_initial();

tests/phpunit/tests/html-api/wpHtmlProcessorHtml5lib.php

@@ -21,6 +21,8 @@
  * @group html-api-html5lib-tests
  */
 class Tests_HtmlApi_Html5lib extends WP_UnitTestCase {
+	const TREE_INDENT = '  ';
+
 	/**
 	 * Skip specific tests that may not be supported or have known issues.
 	 */
@@ -139,10 +141,6 @@ public function data_external_html5lib_tests() {
 	 * @return bool True if the test case should be skipped. False otherwise.
 	 */
 	private static function should_skip_test( ?string $test_context_element, string $test_name ): bool {
-		if ( null !== $test_context_element && 'body' !== $test_context_element ) {
-			return true;
-		}
-
 		if ( array_key_exists( $test_name, self::SKIP_TESTS ) ) {
 			return true;
 		}
@@ -158,21 +156,86 @@ private static function should_skip_test( ?string $test_context_element, string
 	 * @return string|null Tree structure of parsed HTML, if supported, else null.
 	 */
 	private static function build_tree_representation( ?string $fragment_context, string $html ) {
-		$processor = $fragment_context
-			? WP_HTML_Processor::create_fragment( $html, "<{$fragment_context}>" )
-			: WP_HTML_Processor::create_full_parser( $html );
-		if ( null === $processor ) {
-			throw new WP_HTML_Unsupported_Exception( "Could not create a parser with the given fragment context: {$fragment_context}.", '', 0, '', array(), array() );
+		$processor = null;
+		if ( $fragment_context ) {
+			if ( 'body' === $fragment_context ) {
+				$processor = WP_HTML_Processor::create_fragment( $html );
+			} else {
+
+				/*
+				 * If the string of characters starts with "svg ", the context
+				 * element is in the SVG namespace and the substring after
+				 * "svg " is the local name. If the string of characters starts
+				 * with "math ", the context element is in the MathML namespace
+				 * and the substring after "math " is the local name.
+				 * Otherwise, the context element is in the HTML namespace and
+				 * the string is the local name.
+				 */
+				if ( str_starts_with( $fragment_context, 'svg ' ) ) {
+					$tag_name = substr( $fragment_context, 4 );
+					if ( 'svg' === $tag_name ) {
+						$parent_processor = WP_HTML_Processor::create_full_parser( '<!DOCTYPE html><svg>' );
+					} else {
+						$parent_processor = WP_HTML_Processor::create_full_parser( "<!DOCTYPE html><svg><{$tag_name}>" );
+					}
+					$parent_processor->next_tag( $tag_name );
+				} elseif ( str_starts_with( $fragment_context, 'math ' ) ) {
+					$tag_name = substr( $fragment_context, 5 );
+					if ( 'math' === $tag_name ) {
+						$parent_processor = WP_HTML_Processor::create_full_parser( '<!DOCTYPE html><math>' );
+					} else {
+						$parent_processor = WP_HTML_Processor::create_full_parser( "<!DOCTYPE html><math><{$tag_name}>" );
+					}
+					$parent_processor->next_tag( $tag_name );
+				} else {
+					if ( in_array(
+						$fragment_context,
+						array(
+							'caption',
+							'col',
+							'colgroup',
+							'tbody',
+							'td',
+							'tfoot',
+							'th',
+							'thead',
+							'tr',
+						),
+						true
+					) ) {
+						$parent_processor = WP_HTML_Processor::create_full_parser( "<!DOCTYPE html><table><{$fragment_context}>" );
+						$parent_processor->next_tag();
+					} else {
+						$parent_processor = WP_HTML_Processor::create_full_parser( "<!DOCTYPE html><{$fragment_context}>" );
+					}
+					$parent_processor->next_tag( $fragment_context );
+				}
+				if ( null !== $parent_processor->get_unsupported_exception() ) {
+					throw $parent_processor->get_unsupported_exception();
+				}
+				if ( null !== $parent_processor->get_last_error() ) {
+					throw new Exception( $parent_processor->get_last_error() );
+				}
+				$processor = $parent_processor->spawn_fragment_parser( $html );
+			}
+
+			if ( null === $processor ) {
+				throw new WP_HTML_Unsupported_Exception( "Could not create a parser with the given fragment context: {$fragment_context}.", '', 0, '', array(), array() );
+			}
+		} else {
+			$processor = WP_HTML_Processor::create_full_parser( $html );
+			if ( null === $processor ) {
+				throw new Exception( 'Could not create a full parser.' );
+			}
 		}
 
 		/*
 		 * The fragment parser will start in 2 levels deep at: html > body > [position]
 		 * and requires adjustment to initial parameters.
 		 * The full parser will not.
 		 */
-		$output       = $fragment_context ? "<html>\n  <head>\n  <body>\n" : '';
-		$indent_level = $fragment_context ? 2 : 0;
-		$indent       = '  ';
+		$output       = '';
+		$indent_level = 0;
 		$was_text     = null;
 		$text_node    = '';
 
@@ -225,7 +288,7 @@ private static function build_tree_representation( ?string $fragment_context, st
 						++$indent_level;
 					}
 
-					$output .= str_repeat( $indent, $tag_indent ) . "<{$tag_name}>\n";
+					$output .= str_repeat( self::TREE_INDENT, $tag_indent ) . "<{$tag_name}>\n";
 
 					$attribute_names = $processor->get_attribute_names_with_prefix( '' );
 					if ( $attribute_names ) {
@@ -278,18 +341,18 @@ static function ( $a, $b ) {
 							if ( true === $val ) {
 								$val = '';
 							}
-							$output .= str_repeat( $indent, $tag_indent + 1 ) . "{$display_name}=\"{$val}\"\n";
+							$output .= str_repeat( self::TREE_INDENT, $tag_indent + 1 ) . "{$display_name}=\"{$val}\"\n";
 						}
 					}
 
 					// Self-contained tags contain their inner contents as modifiable text.
 					$modifiable_text = $processor->get_modifiable_text();
 					if ( '' !== $modifiable_text ) {
-						$output .= str_repeat( $indent, $tag_indent + 1 ) . "\"{$modifiable_text}\"\n";
+						$output .= str_repeat( self::TREE_INDENT, $tag_indent + 1 ) . "\"{$modifiable_text}\"\n";
 					}
 
 					if ( 'html' === $namespace && 'TEMPLATE' === $token_name ) {
-						$output .= str_repeat( $indent, $indent_level ) . "content\n";
+						$output .= str_repeat( self::TREE_INDENT, $indent_level ) . "content\n";
 						++$indent_level;
 					}
 
@@ -303,14 +366,14 @@ static function ( $a, $b ) {
 					}
 					$was_text = true;
 					if ( '' === $text_node ) {
-						$text_node .= str_repeat( $indent, $indent_level ) . '"';
+						$text_node .= str_repeat( self::TREE_INDENT, $indent_level ) . '"';
 					}
 					$text_node .= $text_content;
 					break;
 
 				case '#funky-comment':
 					// Comments must be "<" then "!-- " then the data then " -->".
-					$output .= str_repeat( $indent, $indent_level ) . "<!-- {$processor->get_modifiable_text()} -->\n";
+					$output .= str_repeat( self::TREE_INDENT, $indent_level ) . "<!-- {$processor->get_modifiable_text()} -->\n";
 					break;
 
 				case '#comment':
@@ -333,7 +396,7 @@ static function ( $a, $b ) {
 							throw new Error( "Unhandled comment type for tree construction: {$processor->get_comment_type()}" );
 					}
 					// Comments must be "<" then "!-- " then the data then " -->".
-					$output .= str_repeat( $indent, $indent_level ) . "<!-- {$comment_text_content} -->\n";
+					$output .= str_repeat( self::TREE_INDENT, $indent_level ) . "<!-- {$comment_text_content} -->\n";
 					break;
 
 				default:
@@ -449,7 +512,7 @@ public static function parse_html5_dat_testfile( $filename ) {
 				 * context element as context.
 				 */
 				case 'document-fragment':
-					$test_context_element = explode( ' ', $line )[0];
+					$test_context_element = trim( $line );
 					break;
 
 				/*