Changeset 34492

Timestamp:

09/24/2015 02:03:05 PM (9 years ago)

Author:

wonderboymusic

Message:

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.

Location:

trunk

Files:

• src/wp-includes/canonical.php (modified) (2 diffs)
• src/wp-includes/class-wp-rewrite.php (modified) (1 diff)
• src/wp-includes/class-wp.php (modified) (2 diffs)
• src/wp-includes/rewrite-functions.php (modified) (1 diff)
• tests/phpunit/tests/canonical.php (modified) (1 diff)
• tests/phpunit/tests/canonical/paged.php (added)

trunk/src/wp-includes/canonical.php

@@ -147 +147 @@
                 $redirect['query'] = _remove_qs_args_if_not_in_url( $redirect['query'], array( 'page', 'feed', 'p', 'page_id', 'attachment_id', 'pagename', 'name', 'post_type' ), $redirect_url );
             }
+        }
+
+        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 );
         }
 
@@ -561 +568 @@
         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

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

@@ -1072 +1072 @@
                     $sub2 .= '?$';
 
-                    //post pagination, e.g. <permalink>/2/
-                    $match = $match . '(/[0-9]+)?/?$';
+                    // Post pagination, e.g. <permalink>/2/
+                    // 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

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

@@ -588 +588 @@
      */
     public function handle_404() {
-        global $wp_query;
+        global $wp_query, $wp;
 
         // If we've already issued a 404, bail.
@@ -597 +597 @@
         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' ) );
                 }
-            }
-
-            status_header( 200 );
-            return;
+
+                // 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;
+            }
         }
 

trunk/src/wp-includes/rewrite-functions.php

@@ -257 +257 @@
         $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;

trunk/tests/phpunit/tests/canonical.php

@@ -97 +97 @@
             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/' ),