From f80516dfabca785d5cd3a746bdbb17543e77cd56 Mon Sep 17 00:00:00 2001
From: Dennis Snell <dmsnell@git.wordpress.org>
Date: Fri, 2 Feb 2024 22:55:27 +0000
Subject: [PATCH] HTML API: Reset parser state after seeking to bookmark.

When parser states were introduced, nothing in the `seek()` method reset the
parser state. This is problematic because it could leave the parser in the
wrong state.

In this patch the parser state is reset so that it's properly adjusted on
the successive call to `next_token()`.

Developed in https://github.com/WordPress/wordpress-develop/pull/6021
Discussed in https://core.trac.wordpress.org/ticket/60428

Follow-up to [57211]

Props dmsnell, kevin940726
Fixes #60428



git-svn-id: https://develop.svn.wordpress.org/trunk@57527 602fd350-edb4-49c9-b593-d223f7449a82
---
 .../html-api/class-wp-html-tag-processor.php  |  1 +
 .../html-api/wpHtmlTagProcessor-bookmark.php  | 45 +++++++++++++++++++
 2 files changed, 46 insertions(+)

diff --git a/src/wp-includes/html-api/class-wp-html-tag-processor.php b/src/wp-includes/html-api/class-wp-html-tag-processor.php
index 1b4db41bce..b437595bd9 100644
--- a/src/wp-includes/html-api/class-wp-html-tag-processor.php
+++ b/src/wp-includes/html-api/class-wp-html-tag-processor.php
@@ -2335,6 +2335,7 @@ class WP_HTML_Tag_Processor {
 
 		// Point this tag processor before the sought tag opener and consume it.
 		$this->bytes_already_parsed = $this->bookmarks[ $bookmark_name ]->start;
+		$this->parser_state         = self::STATE_READY;
 		return $this->next_token();
 	}
 
diff --git a/tests/phpunit/tests/html-api/wpHtmlTagProcessor-bookmark.php b/tests/phpunit/tests/html-api/wpHtmlTagProcessor-bookmark.php
index 1041dabd43..a0a3b2aa44 100644
--- a/tests/phpunit/tests/html-api/wpHtmlTagProcessor-bookmark.php
+++ b/tests/phpunit/tests/html-api/wpHtmlTagProcessor-bookmark.php
@@ -435,4 +435,49 @@ HTML;
 		$this->setExpectedIncorrectUsage( 'WP_HTML_Tag_Processor::seek' );
 		$this->assertFalse( $processor->seek( 'bookmark' ), "$i-th seek() to the bookmark succeeded, even though it should exceed the allowed limit" );
 	}
+
+	/**
+	 * Ensures that it's possible to seek to an earlier location in a document even
+	 * after reaching the end of a document, when most functionality shuts down.
+	 *
+	 * @ticket 60428
+	 *
+	 * @dataProvider data_incomplete_html_with_target_nodes_for_seeking
+	 *
+	 * @param string $html_with_target_element HTML string containing a tag with a `target` attribute.
+	 */
+	public function test_can_seek_after_document_ends( $html_with_target_element ) {
+		$processor = new WP_HTML_Tag_Processor( $html_with_target_element );
+
+		$sought_tag_name = null;
+		while ( $processor->next_tag() ) {
+			if ( null !== $processor->get_attribute( 'target' ) ) {
+				$processor->set_bookmark( 'target' );
+				$sought_tag_name = $processor->get_tag();
+			}
+		}
+
+		$this->assertTrue(
+			$processor->seek( 'target' ),
+			'Should have been able to seek to the target bookmark after reaching the end of the document.'
+		);
+
+		$this->assertSame(
+			$sought_tag_name,
+			$processor->get_tag(),
+			"Should have found original target node instead of {$processor->get_tag()}."
+		);
+	}
+
+	/**
+	 * Data provider.
+	 *
+	 * @return array[].
+	 */
+	public static function data_incomplete_html_with_target_nodes_for_seeking() {
+		return array(
+			'Compete document'    => array( '<div><img target></div>' ),
+			'Incomplete document' => array( '<div><img target></div' ),
+		);
+	}
 }