HTML API: Preserve PRE/LISTING newline skipping after seek

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

@@ -754,6 +754,14 @@ class WP_HTML_Tag_Processor {
 	 */
 	protected $bookmarks = array();
 
+	/**
+	 * Tracks bookmarks set on text nodes whose leading linefeed is ignored.
+	 *
+	 * @since 6.9.0
+	 * @var bool[]
+	 */
+	private $bookmarks_with_skipped_newline = array();
+
 	const ADD_CLASS    = true;
 	const REMOVE_CLASS = false;
 	const SKIP_CLASS   = null;
@@ -1357,6 +1365,11 @@ public function set_bookmark( $name ): bool {
 		}
 
 		$this->bookmarks[ $name ] = new WP_HTML_Span( $this->token_starts_at, $this->token_length );
+		if ( self::STATE_TEXT_NODE === $this->parser_state && $this->skip_newline_at === $this->token_starts_at ) {
+			$this->bookmarks_with_skipped_newline[ $name ] = true;
+		} else {
+			unset( $this->bookmarks_with_skipped_newline[ $name ] );
+		}
 
 		return true;
 	}
@@ -1377,6 +1390,7 @@ public function release_bookmark( $name ): bool {
 		}
 
 		unset( $this->bookmarks[ $name ] );
+		unset( $this->bookmarks_with_skipped_newline[ $name ] );
 
 		return true;
 	}
@@ -2517,7 +2531,9 @@ private function apply_attributes_updates( int $shift_this_point ): int {
 			return 0;
 		}
 
-		$accumulated_shift_for_given_point = 0;
+		$accumulated_shift_for_given_point  = 0;
+		$accumulated_shift_for_skip_newline = 0;
+		$skip_newline_at                    = $this->skip_newline_at;
 
 		/*
 		 * Attribute updates can be enqueued in any order but updates
@@ -2541,6 +2557,10 @@ private function apply_attributes_updates( int $shift_this_point ): int {
 				$this->bytes_already_parsed += $shift;
 			}
 
+			if ( null !== $skip_newline_at && $diff->start < $skip_newline_at ) {
+				$accumulated_shift_for_skip_newline += $shift;
+			}
+
 			// Accumulate shift of the given pointer within this function call.
 			if ( $diff->start < $shift_this_point ) {
 				$accumulated_shift_for_given_point += $shift;
@@ -2551,6 +2571,10 @@ private function apply_attributes_updates( int $shift_this_point ): int {
 			$bytes_already_copied = $diff->start + $diff->length;
 		}
 
+		if ( null !== $skip_newline_at ) {
+			$this->skip_newline_at += $accumulated_shift_for_skip_newline;
+		}
+
 		$this->html = $output_buffer . substr( $this->html, $bytes_already_copied );
 
 		/*
@@ -2583,8 +2607,11 @@ private function apply_attributes_updates( int $shift_this_point ): int {
 
 				$delta = strlen( $diff->text ) - $diff->length;
 
-				if ( $bookmark->start >= $diff->start ) {
+				if ( $bookmark->start > $diff->start || ( $bookmark->start === $diff->start && 0 === $diff->length ) ) {
 					$head_delta += $delta;
+					if ( $bookmark->start === $diff->start && 0 === $diff->length && '' !== $diff->text ) {
+						unset( $this->bookmarks_with_skipped_newline[ $bookmark_name ] );
+					}
 				}
 
 				if ( $bookmark_end >= $diff_end ) {
@@ -2657,6 +2684,9 @@ public function seek( $bookmark_name ): bool {
 
 		// Point this tag processor before the sought tag opener and consume it.
 		$this->bytes_already_parsed = $this->bookmarks[ $bookmark_name ]->start;
+		$this->skip_newline_at      = isset( $this->bookmarks_with_skipped_newline[ $bookmark_name ] )
+			? $this->bytes_already_parsed
+			: null;
 		$this->parser_state         = self::STATE_READY;
 		return $this->next_token();
 	}
@@ -3636,13 +3666,6 @@ public function subdivide_text_appropriately(): bool {
 	 * that a token has modifiable text, and a token with modifiable text may
 	 * have an empty string (e.g. a comment with no contents).
 	 *
-	 * Limitations:
-	 *
-	 *  - This function will not strip the leading newline appropriately
-	 *    after seeking into a LISTING or PRE element. To ensure that the
-	 *    newline is treated properly, seek to the LISTING or PRE opening
-	 *    tag instead of to the first text node inside the element.
-	 *
 	 * @since 6.5.0
 	 * @since 6.7.0 Replaces NULL bytes (U+0000) and newlines appropriately.
 	 *

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

@@ -9,6 +9,37 @@
  * @coversDefaultClass WP_HTML_Processor
  */
 class Tests_HtmlApi_WpHtmlProcessorModifiableText extends WP_UnitTestCase {
+	/**
+	 * Ensures that bookmarked text remains seekable after updating the same text node.
+	 */
+	public function test_seeks_to_bookmarked_text_after_modifiable_text_update(): void {
+		$processor = WP_HTML_Processor::create_fragment( "<pre>\nabc</pre><span>" );
+
+		$this->assertTrue( $processor->next_token(), 'Should have found the PRE opener.' );
+		$this->assertSame( 'PRE', $processor->get_token_name(), 'Should have found the PRE opener: check test setup.' );
+
+		while ( $processor->next_token() && 'abc' !== $processor->get_modifiable_text() ) {
+			continue;
+		}
+
+		$this->assertSame( '#text', $processor->get_token_name(), 'Should have found the PRE text node: check test setup.' );
+		$this->assertSame( 'abc', $processor->get_modifiable_text(), 'Should have stripped the leading newline from the PRE text on first traversal.' );
+		$this->assertTrue( $processor->set_bookmark( 'text' ), 'Should have bookmarked the PRE text node.' );
+
+		$this->assertTrue( $processor->set_modifiable_text( 'xyz' ), 'Should have updated the PRE text node.' );
+		$this->assertTrue( $processor->next_token(), 'Should have advanced away from the bookmarked text node.' );
+		$this->assertSame( 'PRE', $processor->get_token_name(), 'Should have advanced to the PRE closer: check test setup.' );
+		$this->assertSame( "<pre>\nxyz</pre><span>", $processor->get_updated_html(), 'Should have updated the PRE text node.' );
+
+		$this->assertTrue( $processor->seek( 'text' ), 'Should have sought back to the updated PRE text node.' );
+		$this->assertSame( '#text', $processor->get_token_name(), 'Should have sought back to the text node.' );
+		$this->assertSame(
+			'xyz',
+			$processor->get_modifiable_text(),
+			'Should have replayed the updated PRE text node after seeking.'
+		);
+	}
+
 	/**
 	 * TEXTAREA elements ignore the first newline in their content.
 	 * Setting the modifiable text with a leading newline (or carriage return variants)

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

@@ -410,6 +410,47 @@ public function test_updates_bookmark_for_deletions_before_both_sides() {
 		);
 	}
 
+	/**
+	 * Ensures that bookmarks stay attached to their original token when text is inserted at the bookmark start.
+	 *
+	 * @ticket 56299
+	 *
+	 * @covers WP_HTML_Tag_Processor::seek
+	 */
+	public function test_updates_bookmark_for_insertions_at_start() {
+		$processor = new class( '<div></div><span></span>' ) extends WP_HTML_Tag_Processor {
+			/**
+			 * Inserts HTML at the start of the given bookmark.
+			 *
+			 * @param string $bookmark_name Bookmark name.
+			 * @param string $html          HTML to insert.
+			 */
+			public function insert_html_at_bookmark_start( string $bookmark_name, string $html ): void {
+				$this->lexical_updates[] = new WP_HTML_Text_Replacement(
+					$this->bookmarks[ $bookmark_name ]->start,
+					0,
+					$html
+				);
+			}
+		};
+
+		$processor->next_tag( 'SPAN' );
+		$processor->set_bookmark( 'target' );
+		$processor->insert_html_at_bookmark_start( 'target', '<p></p>' );
+
+		$this->assertSame(
+			'<div></div><p></p><span></span>',
+			$processor->get_updated_html(),
+			'Should have inserted the HTML at the bookmark start.'
+		);
+		$this->assertTrue( $processor->seek( 'target' ), 'Should have sought back to the original bookmarked token.' );
+		$this->assertSame(
+			'SPAN',
+			$processor->get_token_name(),
+			'Should have kept the bookmark attached to the original SPAN token after insertion.'
+		);
+	}
+
 	/**
 	 * @ticket 56299
 	 *