Ticket #21688: 21688-5.patch

wp-includes/default-filters.php

@@ -192 +192 @@
 add_filter( 'editable_slug',            'urldecode'                           );
 add_filter( 'editable_slug',            'esc_textarea'                        );
 add_filter( 'nav_menu_meta_box_object', '_wp_nav_menu_meta_box_object'        );
+add_filter( 'wp_search_stopwords',      'wp_search_stopwords'                 );
 
 // Actions
 add_action( 'wp_head',             'wp_enqueue_scripts',              1     );

wp-includes/functions.php

@@ -3637 +3637 @@
         return true;
 }
 
+/**
+ * Common stopwords (terms) that are excluded from the separate term matching when searching for posts.
+ * The list of English stopwords is the approximate search engines list. MySQL has a much longer default list of full-text stopwords.
+ *
+ * @since 3.5.0
+ *
+ * @param array optional Terms to exclude when searching
+ * @return array
+ */
+function wp_search_stopwords( $in = array() ) {
+        $_words = explode( ',', _x('about,are,com,for,from,how,that,the,this,was,what,when,where,who,will,with,www', 'Comma separated list of common words to exclude when searching.') );
+        $words = array();
+
+        foreach( $_words as $word ) {
+                $word = trim($word, "\r\n\t ");
+                if ( !$word )
+                        continue;
+                $words[] = $word;
+        }
+
+        return array_merge( $in, $words );
+}
+
+/**
+ * Check if the term is suitable for searching.
+ *
+ * @access private
+ * @since 3.5.0
+ *
+ * @param string Term to check
+ * @return string
+ */
+function _check_search_term($term) {
+        static $strtolower_func, $stopwords;
+
+        if ( empty( $term ) )
+                return '';
+
+        // keep spaces when term is for exact match
+        if ( preg_match('/^".+"$/', $term) )
+                $term = trim($term, "\"'");
+        else
+                $term = trim($term, "\"' ");
+
+        if ( !$strtolower_func ) {
+                $strtolower_func = function_exists('mb_strtolower') ? 'mb_strtolower' : 'strtolower';
+                $stopwords = apply_filters( 'wp_search_stopwords', array() );
+        }
+
+        if ( empty( $term{2} ) ) {
+                // $term is too short, remove it if it doesn't contain numbers and is lower case
+                if ( !preg_match( '/[0-9]/', $term ) && $strtolower_func($term) == $term  )
+                        $term = '';
+        }
+
+        if ( $term && in_array( $strtolower_func($term), $stopwords, true ) )
+                $term = '';
+
+        return $term;
+}
+

wp-includes/query.php

@@ -2173 +2173 @@
                         }
                 }
 
-                // If a search pattern is specified, load the posts that match
-                if ( !empty($q['s']) ) {
+                // If a search pattern is specified, load the posts that match.
+                // Sanity check: search string shouldn't be more than 1600 characters.
+                // See ticket #21688 for more info.
+                if ( !empty($q['s']) && strlen($q['s']) < 1600 ) {
                         // added slashes screw with quote grouping when done early, so done later
                         $q['s'] = stripslashes($q['s']);
                         if ( empty( $_GET['s'] ) && $this->is_main_query() )
                                 $q['s'] = urldecode($q['s']);
+                        // there are no line breaks in <input /> fields
+                        $q['s'] = str_replace( array("\r", "\n"), '', $q['s'] );
                         if ( !empty($q['sentence']) ) {
                                 $q['search_terms'] = array($q['s']);
                         } else {
-                                preg_match_all('/".*?("|$)|((?<=[\r\n\t ",+])|^)[^\r\n\t ",+]+/', $q['s'], $matches);
-                                $q['search_terms'] = array_map('_search_terms_tidy', $matches[0]);
+                                if ( preg_match_all('/".*?("|$)|((?<=[\t ",+])|^)[^\t ",+]+/', $q['s'], $matches) ) {
+                                        $q['search_terms'] = array();
+
+                                        if ( !empty($matches[0]) ) {
+                                                foreach ( $matches[0] as $term ) {
+                                                        $term = _check_search_term($term);
+                                                        if ( $term )
+                                                                $q['search_terms'][] = $term;
+                                                }
+                                        }
+
+                                        // if the search string has only short terms or stopwords, or is more than 10 terms long, match it as sentence
+                                        if ( empty($q['search_terms']) || count($q['search_terms']) > 10 )
+                                                $q['search_terms'] = array($q['s']);
+                                } else {
+                                        $q['search_terms'] = array($q['s']);
+                                }
                         }
+
                         $n = !empty($q['exact']) ? '' : '%';
                         $searchand = '';
-                        foreach( (array) $q['search_terms'] as $term ) {
+                        foreach( $q['search_terms'] as $term ) {
                                 $term = esc_sql( like_escape( $term ) );
                                 $search .= "{$searchand}(($wpdb->posts.post_title LIKE '{$n}{$term}{$n}') OR ($wpdb->posts.post_content LIKE '{$n}{$term}{$n}'))";
                                 $searchand = ' AND ';