HTML API: Return elements pushed and popped rather than tags read.

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

@@ -51,6 +51,18 @@ class WP_HTML_Open_Elements {
 	 */
 	private $has_p_in_button_scope = false;
 
+	private $pop_handler = null;
+
+	private $push_handler = null;
+
+	public function add_pop_handler( Closure $handler ) {
+		$this->pop_handler = $handler;
+	}
+
+	public function add_push_handler( Closure $handler ) {
+		$this->push_handler = $handler;
+	}
+
 	/**
 	 * Reports if a specific node is in the stack of open elements.
 	 *
@@ -429,6 +441,10 @@ public function after_element_push( $item ) {
 				$this->has_p_in_button_scope = true;
 				break;
 		}
+
+		if ( null !== $this->push_handler ) {
+			( $this->push_handler )( $item );
+		}
 	}
 
 	/**
@@ -458,5 +474,9 @@ public function after_element_pop( $item ) {
 				$this->has_p_in_button_scope = $this->has_element_in_button_scope( 'P' );
 				break;
 		}
+
+		if ( null !== $this->pop_handler ) {
+			( $this->pop_handler )( $item );
+		}
 	}
 }

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

@@ -7,6 +7,16 @@
  * @since 6.4.0
  */
 
+class WP_HTML_Element_Operation {
+	public $token;
+	public $operation;
+
+	public function __construct( $token, $operation ) {
+		$this->token     = $token;
+		$this->operation = $operation;
+	}
+}
+
 /**
  * Core class used to safely parse and modify an HTML document.
  *
@@ -201,6 +211,11 @@ class WP_HTML_Processor extends WP_HTML_Tag_Processor {
 	 */
 	private $release_internal_bookmark_on_destruct = null;
 
+	private $element_queue = array();
+
+	/** @var WP_HTML_Element_Operation */
+	private $current_element = null;
+
 	/*
 	 * Public Interface Functions
 	 */
@@ -299,6 +314,14 @@ public function __construct( $html, $use_the_static_create_methods_instead = nul
 
 		$this->state = new WP_HTML_Processor_State();
 
+		$this->state->stack_of_open_elements->add_push_handler( function ( WP_HTML_Token $token ) {
+			$this->element_queue[] = new WP_HTML_Element_Operation( $token, 'open' );
+		} );
+
+		$this->state->stack_of_open_elements->add_pop_handler( function ( WP_HTML_Token $token ) {
+			$this->element_queue[] = new WP_HTML_Element_Operation( $token, 'close' );
+		} );
+
 		/*
 		 * Create this wrapper so that it's possible to pass
 		 * a private method into WP_HTML_Token classes without
@@ -365,7 +388,7 @@ public function next_tag( $query = null ) {
 					continue;
 				}
 
-				if ( ! $this->is_tag_closer() ) {
+				if ( ! parent::is_tag_closer() ) {
 					return true;
 				}
 			}
@@ -392,7 +415,7 @@ public function next_tag( $query = null ) {
 					continue;
 				}
 
-				if ( ! $this->is_tag_closer() ) {
+				if ( ! parent::is_tag_closer() ) {
 					return true;
 				}
 			}
@@ -440,7 +463,22 @@ public function next_tag( $query = null ) {
 	 * @return bool
 	 */
 	public function next_token() {
-		return $this->step();
+		$this->current_element = null;
+
+		if ( 0 === count( $this->element_queue ) && ! $this->step() ) {
+			while ( $this->state->stack_of_open_elements->pop() ) {
+				continue;
+			}
+		}
+
+		$this->current_element = array_shift( $this->element_queue );
+		return null !== $this->current_element;
+	}
+
+	public function is_tag_closer() {
+		return isset( $this->current_element )
+			? ( 'close' === $this->current_element->operation )
+			: parent::is_tag_closer();
 	}
 
 	/**
@@ -629,7 +667,7 @@ public function get_breadcrumbs() {
 	private function step_in_body() {
 		$token_name = $this->get_token_name();
 		$token_type = $this->get_token_type();
-		$op_sigil   = '#tag' === $token_type ? ( $this->is_tag_closer() ? '-' : '+' ) : '';
+		$op_sigil   = '#tag' === $token_type ? ( parent::is_tag_closer() ? '-' : '+' ) : '';
 		$op         = "{$op_sigil}{$token_name}";
 
 		switch ( $op ) {
@@ -1152,7 +1190,7 @@ private function step_in_body() {
 				throw new WP_HTML_Unsupported_Exception( "Cannot process {$token_name} element." );
 		}
 
-		if ( ! $this->is_tag_closer() ) {
+		if ( ! parent::is_tag_closer() ) {
 			/*
 			 * > Any other start tag
 			 */
@@ -1248,6 +1286,10 @@ public function get_tag() {
 			return null;
 		}
 
+		if ( isset( $this->current_element ) ) {
+			return $this->current_element->token->node_name;
+		}
+
 		$tag_name = parent::get_tag();
 
 		switch ( $tag_name ) {
@@ -1263,6 +1305,14 @@ public function get_tag() {
 		}
 	}
 
+	public function get_token_name() {
+		if ( isset( $this->current_element ) ) {
+			return $this->current_element->token->node_name;
+		}
+
+		return parent::get_token_name();
+	}
+
 	/**
 	 * Removes a bookmark that is no longer needed.
 	 *

src/wp-includes/html-api/class-wp-html-to-markdown-converter.php

@@ -0,0 +1,155 @@
+<?php
+
+class WP_HTML_To_Markdown_Converter {
+	public static function convert( $html ) {
+		$processor  = WP_HTML_Processor::create_fragment( $html );
+		$md         = '';
+		$list_items = array();
+		$depth      = 0;
+		echo "\n";
+
+		echo "\e[90mFound these nodes…\e[m\n";
+		$node_count = 8;
+		while ( $processor->next_token() ) {
+			$indent      = str_pad( '', $depth * 2, ' ' );
+			$token_name  = $processor->get_token_name();
+			$breadcrumbs = $processor->get_breadcrumbs();
+
+			$closer = $processor->is_tag_closer() ? '/' : '';
+			if ( 0 === --$node_count ) {
+				$node_count = 8;
+				echo "\n";
+			}
+			echo "\e[36m{$closer}\e[32m{$token_name}\e[m ";
+
+			if ( $processor->is_tag_closer() ) {
+				switch ( $token_name ) {
+					case 'H1':
+					case 'H2':
+					case 'H3':
+					case 'H4':
+					case 'H5':
+					case 'H6':
+						$md .= "\n";
+						break;
+
+					case 'B':
+					case 'STRONG':
+						$md .= '*';
+						break;
+
+					case 'I':
+					case 'EM':
+						$md .= '_';
+						break;
+
+					case 'OL':
+					case 'UL':
+						--$depth;
+						array_pop( $list_items );
+						break;
+				}
+
+				// Proceed to the next token.
+				continue;
+			}
+
+			switch ( $token_name ) {
+				case '#text':
+					$md .= $processor->get_modifiable_text();
+					break;
+
+				case 'P':
+					$md .= "\n";
+					break;
+
+				case 'H1':
+				case 'H2':
+				case 'H3':
+				case 'H4':
+				case 'H5':
+				case 'H6':
+					$hash_count = intval( $token_name[1] );
+					$hashes     = str_pad( '', $hash_count, '#' );
+					$md        .= "\n\n{$hashes} ";
+					break;
+
+				case 'B':
+				case 'STRONG':
+					$md .= '*';
+					break;
+
+				case 'I':
+				case 'EM':
+					$md .= '_';
+					break;
+
+				case 'LI':
+					$list_item = end( $list_items );
+					$md       .= "\n{$indent}{$list_item} ";
+					break;
+
+				case 'OL':
+					++$depth;
+					$list_items[] = '*';
+					break;
+
+				case 'UL':
+					++$depth;
+					$list_items[] = '-';
+					break;
+			}
+
+			$last_breadcrumbs = $breadcrumbs;
+		}
+
+		if ( null !== $processor->get_last_error() ) {
+			die( "Encountered unsupported HTML: failed to convert.\n" );
+		}
+
+		$closed_elements = array();
+		for ( $i = 0; $i < count( $last_breadcrumbs ); $i++ ) {
+			if (
+				isset( $last_breadcrumbs[ $i ], $breadcrumbs[ $i ] ) &&
+				$last_breadcrumbs[ $i ] === $breadcrumbs[ $i ]
+			) {
+				continue;
+			}
+
+			$closed_elements = array_slice( $last_breadcrumbs, $i );
+			break;
+		}
+
+		$closed_elements = array_reverse( $closed_elements );
+		foreach ( $closed_elements as $element ) {
+			switch ( $element ) {
+				case 'H1':
+				case 'H2':
+				case 'H3':
+				case 'H4':
+				case 'H5':
+				case 'H6':
+					$md .= "\n";
+					break;
+
+				case 'B':
+				case 'STRONG':
+					$md .= '*';
+					break;
+
+				case 'I':
+				case 'EM':
+					$md .= '_';
+					break;
+
+				case 'OL':
+				case 'UL':
+					--$depth;
+					array_pop( $list_items );
+					break;
+			}
+		}
+
+		return $md;
+	}
+}

src/wp-settings.php

@@ -258,6 +258,7 @@
 require ABSPATH . WPINC . '/html-api/class-wp-html-token.php';
 require ABSPATH . WPINC . '/html-api/class-wp-html-processor-state.php';
 require ABSPATH . WPINC . '/html-api/class-wp-html-processor.php';
+require ABSPATH . WPINC . '/html-api/class-wp-html-to-markdown-converter.php';
 require ABSPATH . WPINC . '/class-wp-http.php';
 require ABSPATH . WPINC . '/class-wp-http-streams.php';
 require ABSPATH . WPINC . '/class-wp-http-curl.php';