HTML API: Refactor wp_strip_all_tags()

src/wp-includes/formatting.php

@@ -5517,17 +5517,19 @@ function normalize_whitespace( $str ) {
  * will return 'something'. wp_strip_all_tags will return ''
  *
  * @since 2.9.0
+ * @since 6.5.0 Relies on the HTML API for processing.
+ *
+ * @param string $html          String potentially containing HTML tags, comments, etc...
+ * @param bool   $remove_breaks Optional. Whether to remove left over line breaks and white space chars.
  *
- * @param string $text          String containing HTML tags
- * @param bool   $remove_breaks Optional. Whether to remove left over line breaks and white space chars
  * @return string The processed string.
  */
-function wp_strip_all_tags( $text, $remove_breaks = false ) {
-	if ( is_null( $text ) ) {
+function wp_strip_all_tags( $html, $remove_breaks = false ) {
+	if ( ! isset( $html ) ) {
 		return '';
 	}
 
-	if ( ! is_scalar( $text ) ) {
+	if ( ! is_scalar( $html ) ) {
 		/*
 		 * To maintain consistency with pre-PHP 8 error levels,
 		 * trigger_error() is used to trigger an E_USER_WARNING,
@@ -5541,22 +5543,33 @@ function wp_strip_all_tags( $text, $remove_breaks = false ) {
 				'#1',
 				'$text',
 				'string',
-				gettype( $text )
+				gettype( $html )
 			),
 			E_USER_WARNING
 		);
 
 		return '';
 	}
 
-	$text = preg_replace( '@<(script|style)[^>]*?>.*?</\\1>@si', '', $text );
-	$text = strip_tags( $text );
+	$processor    = new WP_Unsafe_Internal_Tag_Processor( (string) $html );
+	$text_content = '';
+
+	while ( $processor->next_token() ) {
+		switch ( $processor->get_token_name() ) {
+			case '#text':
+			case 'TEXTAREA':
+			case 'TITLE':
+			case 'XMP':
+				$text_content .= $processor->unsafe_get_raw_modifiable_text();
+				break;
+		}
+	}
 
 	if ( $remove_breaks ) {
-		$text = preg_replace( '/[\r\n\t ]+/', ' ', $text );
+		$text_content = preg_replace( '/[\f\r\n\t ]+/', ' ', $text_content );
 	}
 
-	return trim( $text );
+	return trim( $text_content );
 }
 
 /**

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

@@ -2810,7 +2810,7 @@ public function get_modifiable_text() {
 
 		$decoded = html_entity_decode( $text, ENT_QUOTES | ENT_HTML5 | ENT_SUBSTITUTE );
 
-		if ( empty( $decoded ) ) {
+		if ( 0 === strlen( $decoded ) ) {
 			return '';
 		}
 
@@ -2834,6 +2834,40 @@ public function get_modifiable_text() {
 		return $decoded;
 	}
 
+	/**
+	 * Returns the span covering the current token, or null if none matched.
+	 *
+	 * @since 6.6.0
+	 *
+	 * @access private
+	 *
+	 * @return WP_HTML_Span|null Span of the current token if matched, otherwise `null`.
+	 */
+	protected function unsafe_get_token_extents() {
+		if ( null === $this->token_starts_at ) {
+			return null;
+		}
+
+		return new WP_HTML_Span( $this->token_starts_at, $this->token_length );
+	}
+
+	/**
+	 * Returns the span covering the current modifiable text, or null if none matched.
+	 *
+	 * @since 6.6.0
+	 *
+	 * @access private
+	 *
+	 * @return WP_HTML_Span|null Span of the current modifiable text if any, otherwise `null`.
+	 */
+	protected function unsafe_get_modifiable_text_extents() {
+		if ( null === $this->text_starts_at ) {
+			return null;
+		}
+
+		return new WP_HTML_Span( $this->text_starts_at, $this->text_length );
+	}
+
 	/**
 	 * Updates or creates a new attribute on the currently matched tag with the passed value.
 	 *

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

@@ -0,0 +1,19 @@
+<?php
+
+class WP_Unsafe_Internal_Tag_Processor extends WP_HTML_Tag_Processor {
+	public function unsafe_get_raw_modifiable_text() {
+		$text = $this->unsafe_get_modifiable_text_extents();
+
+		return null !== $text
+			? substr( $this->html, $text->start, $text->length )
+			: '';
+	}
+
+	public function unsafe_get_raw_token() {
+		$token = $this->unsafe_get_token_extents();
+
+		return null !== $token
+			? substr( $this->html, $token->start, $token->length )
+			: '';
+	}
+}

src/wp-settings.php

@@ -246,6 +246,7 @@
 require ABSPATH . WPINC . '/html-api/class-wp-html-span.php';
 require ABSPATH . WPINC . '/html-api/class-wp-html-text-replacement.php';
 require ABSPATH . WPINC . '/html-api/class-wp-html-tag-processor.php';
+require ABSPATH . WPINC . '/html-api/class-wp-unsafe-internal-tag-processor.php';
 require ABSPATH . WPINC . '/html-api/class-wp-html-unsupported-exception.php';
 require ABSPATH . WPINC . '/html-api/class-wp-html-active-formatting-elements.php';
 require ABSPATH . WPINC . '/html-api/class-wp-html-open-elements.php';