Sitemaps: Revert r63573 from the 7.0 branch.

The sitemap 404 fix merged in r63573 was committed to the 7.0 branch in error; it was intended for the 7.1 branch. This restores the branch to its state as of r63449 and removes the recorded merge of r63570, leaving 7.0 unchanged from what shipped in 7.0.4.

Reverts r63573.

See #65945.


git-svn-id: https://develop.svn.wordpress.org/branches/7.0@63574 602fd350-edb4-49c9-b593-d223f7449a82

Author: westonruter

src/wp-includes/class-wp.php

@@ -746,9 +746,8 @@ public function handle_404() {
 
 		$set_404 = true;
 
-		// Never 404 here for the admin, robots, favicon, or sitemaps.
-		// Sitemap routes send their own status in WP_Sitemaps::render_sitemaps().
-		if ( is_admin() || is_robots() || is_favicon() || is_sitemap() || get_query_var( 'sitemap-stylesheet' ) ) {
+		// Never 404 for the admin, robots, or favicon.
+		if ( is_admin() || is_robots() || is_favicon() ) {
 			$set_404 = false;
 
 			// If posts were found, check for paged content.

src/wp-includes/sitemaps/class-wp-sitemaps.php

@@ -157,45 +157,30 @@ public function register_rewrites() {
 	 * Renders sitemap templates based on rewrite rules.
 	 *
 	 * @since 5.5.0
+	 *
+	 * @global WP_Query $wp_query WordPress Query object.
 	 */
 	public function render_sitemaps() {
-		/*
-		 * Bail early if this isn't a sitemap or stylesheet route.
-		 *
-		 * This runs on every front-end request, so it comes before any
-		 * sanitizing. The raw query vars are tested here, matching
-		 * WP::handle_404(), which exempts sitemap requests from its own 404 on
-		 * the same basis. Testing the sanitized values instead would let a
-		 * request that handle_404() exempted fall through both, leaving it a 200.
-		 */
-		if ( ! get_query_var( 'sitemap' ) && ! get_query_var( 'sitemap-stylesheet' ) ) {
-			return;
-		}
+		global $wp_query;
 
 		$sitemap         = sanitize_text_field( get_query_var( 'sitemap' ) );
 		$object_subtype  = sanitize_text_field( get_query_var( 'sitemap-subtype' ) );
 		$stylesheet_type = sanitize_text_field( get_query_var( 'sitemap-stylesheet' ) );
 		$paged           = absint( get_query_var( 'paged' ) );
 
-		// Force a 404 and bail early if the route did not survive sanitizing.
+		// Bail early if this isn't a sitemap or stylesheet route.
 		if ( ! ( $sitemap || $stylesheet_type ) ) {
-			$this->send_404();
 			return;
 		}
 
 		if ( ! $this->sitemaps_enabled() ) {
-			$this->send_404();
+			$wp_query->set_404();
+			status_header( 404 );
 			return;
 		}
 
 		// Render stylesheet if this is stylesheet route.
 		if ( $stylesheet_type ) {
-			// Force a 404 and bail early if the stylesheet type is not recognized.
-			if ( ! in_array( $stylesheet_type, array( 'sitemap', 'index' ), true ) ) {
-				$this->send_404();
-				return;
-			}
-
 			$stylesheet = new WP_Sitemaps_Stylesheet();
 
 			$stylesheet->render_stylesheet( $stylesheet_type );
@@ -212,9 +197,7 @@ public function render_sitemaps() {
 
 		$provider = $this->registry->get_provider( $sitemap );
 
-		// Force a 404 and bail early if the requested provider is not registered.
 		if ( ! $provider ) {
-			$this->send_404();
 			return;
 		}
 
@@ -226,34 +209,15 @@ public function render_sitemaps() {
 
 		// Force a 404 and bail early if no URLs are present.
 		if ( empty( $url_list ) ) {
-			$this->send_404();
+			$wp_query->set_404();
+			status_header( 404 );
 			return;
 		}
 
 		$this->renderer->render_sitemap( $url_list );
 		exit;
 	}
 
-	/**
-	 * Sends a 404 for a sitemap route that cannot be served.
-	 *
-	 * WP::handle_404() exempts sitemap requests, so every sitemap 404 is issued
-	 * here instead. That includes the no-cache headers handle_404() sends with
-	 * its own 404, so an intermediary does not retain a 404 for a route that
-	 * becomes valid once the site has more content.
-	 *
-	 * @since 7.1.1
-	 *
-	 * @global WP_Query $wp_query WordPress Query object.
-	 */
-	private function send_404(): void {
-		global $wp_query;
-
-		$wp_query->set_404();
-		status_header( 404 );
-		nocache_headers();
-	}
-
 	/**
 	 * Redirects a URL to the wp-sitemap.xml
 	 *

tests/phpunit/tests/sitemaps/sitemaps.php

@@ -493,108 +493,4 @@ public function test_empty_url_list_should_return_404() {
 
 		$this->assertTrue( is_404() );
 	}
-
-	/**
-	 * Ensures a paged subtype route with no URLs still 404s, now that
-	 * WP::handle_404() no longer sets a 404 for sitemap requests.
-	 *
-	 * @ticket 65945
-	 */
-	public function test_empty_url_list_for_subtype_should_return_404() {
-		wp_register_sitemap_provider( 'foo', new WP_Sitemaps_Empty_Test_Provider( 'foo' ) );
-
-		$this->go_to( home_url( '/?sitemap=foo&sitemap-subtype=bar&paged=2' ) );
-
-		wp_sitemaps_get_server()->render_sitemaps();
-
-		$this->assertTrue( is_404() );
-	}
-
-	/**
-	 * Ensures a sitemap query var that does not survive sanitizing 404s.
-	 *
-	 * WP::handle_404() exempts these requests on the raw query var, while
-	 * render_sitemaps() acts on the sanitized value. Without a matching bail
-	 * they fall through both and an arbitrary URL is served as a 200.
-	 *
-	 * @ticket 65945
-	 *
-	 * @dataProvider data_unusable_sitemap_query_vars
-	 *
-	 * @param non-falsy-string $query_string Query string to append to a nonexistent URL.
-	 */
-	public function test_unusable_sitemap_query_var_should_return_404( string $query_string ) {
-		$this->set_permalink_structure( '/%postname%/' );
-
-		// Instantiate the server before navigating: registering the sitemap
-		// rewrite tags is what adds the query vars to `$wp->public_query_vars`.
-		$sitemaps = wp_sitemaps_get_server();
-
-		$this->go_to( home_url( '/this-page-does-not-exist/' . $query_string ) );
-
-		$this->assertFalse( is_404(), 'WP::handle_404() should not have set a 404.' );
-
-		$sitemaps->render_sitemaps();
-
-		$this->assertTrue( is_404(), 'render_sitemaps() should have set a 404.' );
-	}
-
-	/**
-	 * Data provider.
-	 *
-	 * @return array<non-falsy-string, array{ non-falsy-string }>
-	 */
-	public function data_unusable_sitemap_query_vars(): array {
-		return array(
-			'value stripped by sanitizing' => array( '?sitemap=<>' ),
-			'array sitemap value'          => array( '?sitemap[]=index' ),
-			'array stylesheet value'       => array( '?sitemap-stylesheet[]=sitemap' ),
-		);
-	}
-
-	/**
-	 * Ensures an unrecognized stylesheet type 404s from render_sitemaps().
-	 *
-	 * WP::handle_404() exempts any request carrying a `sitemap-stylesheet`
-	 * query var, and WP_Sitemaps_Stylesheet::render_stylesheet() echoes nothing
-	 * for a type other than 'sitemap' or 'index', so this route would otherwise
-	 * be served as a 200 with an empty body.
-	 *
-	 * @ticket 65945
-	 */
-	public function test_unrecognized_stylesheet_type_should_return_404() {
-		// Instantiate the server before navigating: registering the sitemap rewrite
-		// tags is what adds `sitemap-stylesheet` to `$wp->public_query_vars`.
-		$sitemaps = wp_sitemaps_get_server();
-
-		$this->go_to( home_url( '/?sitemap-stylesheet=this-is-not-a-stylesheet' ) );
-
-		$this->assertFalse( is_404(), 'WP::handle_404() should not have set a 404.' );
-
-		$sitemaps->render_sitemaps();
-
-		$this->assertTrue( is_404(), 'render_sitemaps() should have set a 404.' );
-	}
-
-	/**
-	 * Ensures an unregistered provider 404s from render_sitemaps().
-	 *
-	 * WP::handle_404() exempts every sitemap request, so this route would
-	 * otherwise be served with a 200.
-	 *
-	 * @ticket 65945
-	 */
-	public function test_unregistered_provider_should_return_404() {
-		// Instantiate the server before navigating: registering the sitemap
-		// rewrite tags is what adds `sitemap` to `$wp->public_query_vars`.
-		$sitemaps = wp_sitemaps_get_server();
-
-		$this->go_to( home_url( '/?sitemap=this-provider-does-not-exist' ) );
-
-		$this->assertFalse( is_404(), 'WP::handle_404() should not have set a 404.' );
-
-		$sitemaps->render_sitemaps();
-
-		$this->assertTrue( is_404(), 'render_sitemaps() should have set a 404.' );
-	}
 }