Canonical/Rewrite: sanity check posts that are paged with <!--nextpage-->. Page numbers past the max number of pages are returning the last page of content and causing infinite duplicate content.

Awesome rewrite bug: the `page` query var was being set to `'/4'` in `$wp`. When cast to `int`, it returns `0` (Bless you, PHP). `WP_Query` calls `trim( $page, '/' )` when setting its own query var. The few places that were checking `page`	before posts were queried now have sanity checks, so that these changes work without flushing rewrites.	

Adds/updates unit tests.

Props wonderboymusic, dd32.
See #11694.


git-svn-id: https://develop.svn.wordpress.org/trunk@34492 602fd350-edb4-49c9-b593-d223f7449a82

src/wp-includes/canonical.php

@@ -148,6 +148,13 @@ function redirect_canonical( $requested_url = null, $do_redirect = true ) {
 			}
 		}
 
+		if ( get_query_var( 'page' ) && $wp_query->post &&
+			false !== strpos( $wp_query->post->post_content, '<!--nextpage-->' ) ) {
+			$redirect['path'] = rtrim( $redirect['path'], (int) get_query_var( 'page' ) . '/' );
+			$redirect['query'] = remove_query_arg( 'page', $redirect['query'] );
+			$redirect_url = get_permalink( $wp_query->post->ID );
+		}
+
 	} elseif ( is_object($wp_rewrite) && $wp_rewrite->using_permalinks() ) {
 		// rewriting of old ?p=X, ?m=2004, ?m=200401, ?m=20040101
 		if ( is_attachment() && ! $redirect_url ) {
@@ -560,7 +567,7 @@ function redirect_guess_404_permalink() {
 			return false;
 		if ( get_query_var( 'feed' ) )
 			return get_post_comments_feed_link( $post_id, get_query_var( 'feed' ) );
-		elseif ( get_query_var( 'page' ) )
+		elseif ( get_query_var( 'page' ) && 1 < get_query_var( 'page' ) )
 			return trailingslashit( get_permalink( $post_id ) ) . user_trailingslashit( get_query_var( 'page' ), 'single_paged' );
 		else
 			return get_permalink( $post_id );

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

@@ -1071,8 +1071,10 @@ class WP_Rewrite {
 					$sub1 .= '?$';
 					$sub2 .= '?$';
 
-					//post pagination, e.g. <permalink>/2/
+					// Post pagination, e.g. <permalink>/2/
-					$match = $match . '(/[0-9]+)?/?$';
+					// Previously: '(/[0-9]+)?/?$', which produced '/2' for page.
+					// When cast to int, returned 0.
+					$match = $match . '(?:/([0-9]+))?/?$';
 					$query = $index . '?' . $query . '&page=' . $this->preg_index($num_toks + 1);
 				} else { //not matching a permalink so this is a lot simpler
 					//close the match and finalise the query

src/wp-includes/class-wp.php

@@ -587,7 +587,7 @@ class WP {
 	 * @global WP_Query $wp_query
  	 */
 	public function handle_404() {
-		global $wp_query;
+		global $wp_query, $wp;
 
 		// If we've already issued a 404, bail.
 		if ( is_404() )
@@ -596,17 +596,27 @@ class WP {
 		// Never 404 for the admin, robots, or if we found posts.
 		if ( is_admin() || is_robots() || $wp_query->posts ) {
 
-			// Only set X-Pingback for single posts.
+			$success = true;
 			if ( is_singular() ) {
 				$p = clone $wp_query->post;
+				// Only set X-Pingback for single posts that allow pings.
 				if ( $p && pings_open( $p ) ) {
 					@header( 'X-Pingback: ' . get_bloginfo( 'pingback_url' ) );
 				}
+
+				// check for paged content that exceeds the max number of pages
+				$next = '<!--nextpage-->';
+				if ( $p && false !== strpos( $p->post_content, $next ) && ! empty( $wp->query_vars['page'] ) ) {
+					$page = trim( $wp->query_vars['page'], '/' );
+					$success = (int) $page <= ( substr_count( $p->post_content, $next ) + 1 );
+				}
 			}
 
+			if ( $success ) {
 				status_header( 200 );
 				return;
 			}
+		}
 
 		// We will 404 for paged queries, as no posts were found.
 		if ( ! is_paged() ) {

src/wp-includes/rewrite-functions.php

@@ -256,6 +256,8 @@ function wp_resolve_numeric_slug_conflicts( $query_vars = array() ) {
 	} elseif ( 'monthnum' === $compare && isset( $query_vars['day'] ) ) {
 		$maybe_page = $query_vars['day'];
 	}
+	// Bug found in #11694 - 'page' was returning '/4'
+	$maybe_page = (int) trim( $maybe_page, '/' );
 
 	$post_page_count = substr_count( $post->post_content, '<!--nextpage-->' ) + 1;
 

tests/phpunit/tests/canonical.php

@@ -96,7 +96,7 @@ class Tests_Canonical extends WP_Canonical_UnitTestCase {
 			array( '/2010/post-format-test-au/', '/2008/06/02/post-format-test-audio/'), // A Year the post is not in
 			array( '/post-format-test-au/', '/2008/06/02/post-format-test-audio/'),
 
-			array( '/2008/09/03/images-test/3/', array( 'url' => '/2008/09/03/images-test/3/', 'qv' => array( 'name' => 'images-test', 'year' => '2008', 'monthnum' => '09', 'day' => '03', 'page' => '/3' ) ) ), // page = /3 ?!
+			array( '/2008/09/03/images-test/3/', array( 'url' => '/2008/09/03/images-test/3/', 'qv' => array( 'name' => 'images-test', 'year' => '2008', 'monthnum' => '09', 'day' => '03', 'page' => '3' ) ) ),
 			array( '/2008/09/03/images-test/?page=3', '/2008/09/03/images-test/3/' ),
 			array( '/2008/09/03/images-te?page=3', '/2008/09/03/images-test/3/' ),
 

tests/phpunit/tests/canonical/paged.php

@@ -0,0 +1,25 @@
+<?php
+/**
+ * @group canonical
+ * @group rewrite
+ * @group query
+ */
+class Tests_Canonical_Paged extends WP_Canonical_UnitTestCase {
+
+	function test_nextpage() {
+		$para = 'This is a paragraph.
+			This is a paragraph.
+			This is a paragraph.';
+		$next = '<!--nextpage-->';
+
+		$post_id = $this->factory->post->create( array(
+			'post_status' => 'publish',
+			'post_content' => "{$para}{$next}{$para}{$next}{$para}"
+		) );
+
+		$link = parse_url( get_permalink( $post_id ), PHP_URL_PATH );
+		$paged = $link . '4/';
+
+		$this->assertCanonical( $paged, $link );
+	}
+}