Merge PR #51: preserve rawtext contents

# Conflicts:
#	tests/phpunit/tests/html-api/wpHtmlProcessor-serialize.php

Author: sirreal

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

@@ -1495,6 +1495,8 @@ public function serialize(): ?string {
 	 *
 	 * @since 6.7.0
 	 * @since 6.9.0 Converted from protected to public method.
+	 * @since 7.1.0 Contents of IFRAME, NOEMBED, NOFRAMES, and XMP elements are
+	 *              serialized literally instead of being dropped or escaped.
 	 *
 	 * @return string Serialization of token, or empty string if no serialization exists.
 	 */
@@ -1636,17 +1638,38 @@ public function serialize_token(): string {
 			$text = $this->get_modifiable_text();
 
 			switch ( $tag_name ) {
+				/*
+				 * The contents of these elements are emitted literally to preserve
+				 * the document's contents, following the HTML serialization spec:
+				 *
+				 * > If the parent of current node is a style, script, xmp, iframe,
+				 * > noembed, noframes, or plaintext element, or if the parent of
+				 * > current node is a noscript element and scripting is enabled for
+				 * > the node, then append the value of current node's data literally.
+				 *
+				 * This is safe because character references are never decoded in
+				 * their contents. RAWTEXT contents (IFRAME, NOEMBED, NOFRAMES,
+				 * STYLE, XMP) cannot contain their own closing tag, so the closer
+				 * appended below cannot be matched early. SCRIPT data may contain
+				 * escaped closers (e.g. within `<!-- -->`), but re-parsing the
+				 * identical bytes follows the same tokenization rules that produced
+				 * this text, terminating at the appended closer all the same.
+				 *
+				 * @see https://html.spec.whatwg.org/multipage/parsing.html#serialising-html-fragments
+				 */
 				case 'IFRAME':
 				case 'NOEMBED':
 				case 'NOFRAMES':
-					$text = '';
-					break;
-
 				case 'SCRIPT':
 				case 'STYLE':
 				case 'XMP':
 					break;
 
+				/*
+				 * The contents of TEXTAREA and TITLE are parsed as RCDATA, in which
+				 * character references are decoded, so the decoded modifiable text
+				 * must be re-escaped to preserve the document's contents.
+				 */
 				default:
 					$text = self::serialize_decoded_text( $text );
 			}

tests/phpunit/tests/html-api/wpHtmlProcessor-serialize.php

@@ -260,16 +260,115 @@ public function test_style_contents_are_not_escaped() {
 	/**
 	 * Ensures that XMP contents are not escaped, as they are not parsed like text nodes are.
 	 *
-	 * @ticket 62036
+	 * XMP contents are parsed as raw text: character references are never decoded.
+	 * Escaping the contents would change the document, e.g. a "<" would be replaced
+	 * by the literal text "&lt;" after serializing and re-parsing.
+	 *
+	 * @ticket 65372
 	 */
 	public function test_xmp_contents_are_not_escaped() {
 		$this->assertSame(
-			WP_HTML_Processor::normalize( "<xmp>apples > or\x00anges & pears < plums</xmp>" ),
-			"<xmp>apples > or\u{FFFD}anges & pears < plums</xmp>",
+			"<xmp>1 < 2 &amp; apples > or\u{FFFD}anges</xmp>",
+			WP_HTML_Processor::normalize( "<xmp>1 < 2 &amp; apples > or\x00anges</xmp>" ),
 			'Should have preserved text inside an XMP element, except for replacing NULL bytes.'
 		);
 	}
 
+	/**
+	 * Ensures that the contents of IFRAME, NOEMBED, and NOFRAMES elements are
+	 * preserved when serializing.
+	 *
+	 * These elements contain raw text which is part of the parsed document.
+	 * Dropping it would change the document's contents across a serialize and
+	 * re-parse cycle.
+	 *
+	 * @ticket 65372
+	 *
+	 * @dataProvider data_rawtext_elements_with_contents
+	 *
+	 * @param string $html Normalized HTML containing a rawtext element with contents.
+	 */
+	public function test_rawtext_element_contents_are_preserved_when_normalizing( string $html ) {
+		$this->assertSame(
+			$html,
+			WP_HTML_Processor::normalize( $html ),
+			'Should have preserved the rawtext element contents.'
+		);
+	}
+
+	/**
+	 * Data provider.
+	 *
+	 * @return array[]
+	 */
+	public static function data_rawtext_elements_with_contents() {
+		return array(
+			'IFRAME with following text'       => array( '<iframe>x</iframe>y' ),
+			'NOEMBED with following text'      => array( '<noembed>x</noembed>y' ),
+			'NOFRAMES with following text'     => array( '<section><noframes>x</noframes>y</section>' ),
+			'NOFRAMES before comment'          => array( '<section><noframes>x</noframes><!----></section>' ),
+			'IFRAME with markup-like contents' => array( '<iframe><div>inert</div></iframe>' ),
+			'NOEMBED with character reference' => array( '<noembed>&amp;</noembed>' ),
+			'IFRAME in foreign content'        => array( '<svg><iframe>1 &lt; 2</iframe></svg>' ),
+		);
+	}
+
+	/**
+	 * Ensures that the contents of IFRAME, NOEMBED, and NOFRAMES elements are
+	 * preserved when serializing full documents, including NOFRAMES elements
+	 * in the HEAD or after a FRAMESET.
+	 *
+	 * @ticket 65372
+	 *
+	 * @dataProvider data_full_documents_with_rawtext_elements
+	 *
+	 * @param string $html     Input HTML document.
+	 * @param string $expected Expected serialization of the full document.
+	 */
+	public function test_rawtext_element_contents_are_preserved_in_full_documents( string $html, string $expected ) {
+		$processor = WP_HTML_Processor::create_full_parser( $html );
+
+		$this->assertSame(
+			$expected,
+			$processor->serialize(),
+			'Should have preserved the rawtext element contents.'
+		);
+	}
+
+	/**
+	 * Data provider.
+	 *
+	 * @return array[]
+	 */
+	public static function data_full_documents_with_rawtext_elements() {
+		return array(
+			'IFRAME in BODY'          => array(
+				'<iframe>x</iframe>y',
+				'<html><head></head><body><iframe>x</iframe>y</body></html>',
+			),
+			'NOEMBED in BODY'         => array(
+				'a<noembed>x</noembed>',
+				'<html><head></head><body>a<noembed>x</noembed></body></html>',
+			),
+			'NOFRAMES in BODY'        => array(
+				'a<noframes>x</noframes>',
+				'<html><head></head><body>a<noframes>x</noframes></body></html>',
+			),
+			'NOFRAMES in HEAD'        => array(
+				'<head><noframes>x</noframes></head>z',
+				'<html><head><noframes>x</noframes></head><body>z</body></html>',
+			),
+			'NOFRAMES in FRAMESET'    => array(
+				'<html><frameset><noframes>x</noframes>',
+				'<html><head></head><frameset><noframes>x</noframes></frameset></html>',
+			),
+			'IFRAME before a comment' => array(
+				'<h3><div><small><dd><iframe>x</iframe><!---->',
+				'<html><head></head><body><h3><div><small><dd><iframe>x</iframe><!----></dd></small></div></h3></body></html>',
+			),
+		);
+	}
+
 	public function test_unexpected_closing_tags_are_removed() {
 		$this->assertSame(
 			WP_HTML_Processor::normalize( 'one</div>two</span>three' ),
@@ -447,6 +546,10 @@ public static function data_tokens_with_null_bytes() {
 			'Foreign content text' => array( "<svg>one\x00two</svg>", "<svg>one\u{FFFD}two</svg>" ),
 			'SCRIPT content'       => array( "<script>alert(\x00)</script>", "<script>alert(\u{FFFD})</script>" ),
 			'STYLE content'        => array( "<style>\x00 {}</style>", "<style>\u{FFFD} {}</style>" ),
+			'IFRAME content'       => array( "<iframe>a\x00b</iframe>", "<iframe>a\u{FFFD}b</iframe>" ),
+			'NOEMBED content'      => array( "<noembed>a\x00b</noembed>", "<noembed>a\u{FFFD}b</noembed>" ),
+			'NOFRAMES content'     => array( "<noframes>a\x00b</noframes>", "<noframes>a\u{FFFD}b</noframes>" ),
+			'XMP content'          => array( "<xmp>a\x00b</xmp>", "<xmp>a\u{FFFD}b</xmp>" ),
 			'Comment text'         => array( "<!-- \x00 -->", "<!-- \u{FFFD} -->" ),
 		);
 	}