Ticket #22692: miqro-22692.6.patch

wp-includes/formatting.php

@@ -73 +73 @@
                 $static_characters = array_merge( array( '---', ' -- ', '--', ' - ', 'xn–', '...', '``', '\'\'', ' (tm)' ), $cockney );
                 $static_replacements = array_merge( array( $em_dash, ' ' . $em_dash . ' ', $en_dash, ' ' . $en_dash . ' ', 'xn--', '…', $opening_quote, $closing_quote, ' ™' ), $cockneyreplace );
 
+                // Pattern-based replacements of characters.
                 $dynamic = array();
-                if ( "'" != $apos ) {
-                        $dynamic[ '/\'(\d\d(?:’|\')?s)/' ] = $apos . '$1'; // '99's
-                        $dynamic[ '/\'(\d)/'                   ] = $apos . '$1'; // '99
-                }
+                $spaces = wp_spaces_regexp();
+
+                // '99s (apostrophe)
+                if ( "'" != $apos )
+                        $dynamic[ '/\'(?=\d)/' ] = $apos;
+
+                // Single quote at start, or preceded by (, {, <, [, ", or spaces.
                 if ( "'" != $opening_single_quote )
-                        $dynamic[ '/(\s|\A|[([{<]|")\'/'       ] = '$1' . $opening_single_quote; // opening single quote, even after (, {, <, [
+                        $dynamic[ '/(?<=\A|[([{<"]|' . $spaces . ')\'/' ] = $opening_single_quote;
+
+                // 9" (double prime)
                 if ( '"' != $double_prime )
-                        $dynamic[ '/(\d)"/'                    ] = '$1' . $double_prime; // 9" (double prime)
+                        $dynamic[ '/(?<=\d)"/' ] = $double_prime;
+
+                // 9' (prime)
                 if ( "'" != $prime )
-                        $dynamic[ '/(\d)\'/'                   ] = '$1' . $prime; // 9' (prime)
+                        $dynamic[ '/(?<=\d)\'/' ] = $prime;
+
+                // Apostrophe in a word.  No spaces or double primes.
                 if ( "'" != $apos )
-                        $dynamic[ '/(\S)\'([^\'\s])/'          ] = '$1' . $apos . '$2'; // apostrophe in a word
+                        $dynamic[ '/(?<!' . $spaces . ')\'(?!\'|' . $spaces . ')/' ] = $apos;
+
+                // Double quote at start, or preceded by (, {, <, [, or spaces, and not followed by spaces.
                 if ( '"' != $opening_quote )
-                        $dynamic[ '/(\s|\A|[([{<])"(?!\s)/'    ] = '$1' . $opening_quote . '$2'; // opening double quote, even after (, {, <, [
+                        $dynamic[ '/(?<=\A|[([{<]|' . $spaces . ')"(?!' . $spaces . ')/' ] = $opening_quote;
+
+                // Any remaining double quotes.
                 if ( '"' != $closing_quote )
-                        $dynamic[ '/"(\s|\S|\Z)/'              ] = $closing_quote . '$1'; // closing double quote
+                        $dynamic[ '/"/' ] = $closing_quote;
+
+                // Single quotes followed by spaces.
                 if ( "'" != $closing_single_quote )
-                        $dynamic[ '/\'([\s.]|\Z)/'             ] = $closing_single_quote . '$1'; // closing single quote
+                        $dynamic[ '/\'(?=\Z|' . $spaces . ')/' ] = $closing_single_quote; 
 
-                $dynamic[ '/\b(\d+)x(\d+)\b/'              ] = '$1&#215;$2'; // 9x9 (times)
+                // 9x9 (times)
+                $dynamic[ '/\b(\d+)x(\d+)\b/' ] = '$1&#215;$2';
 
                 $dynamic_characters = array_keys( $dynamic );
                 $dynamic_replacements = array_values( $dynamic );
@@ -318 +335 @@
         }
 
         $tagregexp = join( '|', array_map( 'preg_quote', array_keys( $shortcode_tags ) ) );
+        $spaces = wp_spaces_regexp();
 
         $pattern =
                   '/'
                 . '<p>'                              // Opening paragraph
-                . '\\s*+'                            // Optional leading whitespace
+                . '(?:' . $spaces . ')*+'            // Optional leading whitespace
                 . '('                                // 1: The shortcode
                 .     '\\['                          // Opening bracket
                 .     "($tagregexp)"                 // 2: Shortcode name
@@ -347 +365 @@
                 .         ')?'
                 .     ')'
                 . ')'
-                . '\\s*+'                            // optional trailing whitespace
+                . '(?:' . $spaces . ')*+'            // optional trailing whitespace
                 . '<\\/p>'                           // closing paragraph
                 . '/s';
 
@@ -1865 +1883 @@
          */
         $src_url = apply_filters( 'smilies_src', includes_url( "images/smilies/$img" ), $img, site_url() );
 
-        return sprintf( ' <img src="%s" alt="%s" class="wp-smiley" /> ', esc_url( $src_url ), esc_attr( $smiley ) );
+        return sprintf( '<img src="%s" alt="%s" class="wp-smiley" />', esc_url( $src_url ), esc_attr( $smiley ) );
 }
 
 /**
@@ -3781 +3799 @@
 
         return false;
 }
+
+/**
+ * Returns the regexp for common whitespace characters.
+ *
+ * By default, spaces include new lines, tabs, nbsp entities, and the UTF-8 nbsp.
+ * This is designed to replace the PCRE \s sequence.  In ticket #22692, that
+ * sequence was found to be unreliable due to random inclusion of the A0 byte.
+ *
+ * @since 3.9.0
+ *
+ * @return string The spaces regexp.
+ */
+function wp_spaces_regexp() {
+        static $spaces;
+
+        if (empty($spaces)) {
+                $spaces = apply_filters( 'wp_spaces_regexp', '[\r\n\t ]|\xC2\xA0|&nbsp;' );
+        }
+
+        return $spaces;
+}

wp-includes/functions.php

@@ -2560 +2560 @@
          */
         krsort($wpsmiliestrans);
 
-        $wp_smiliessearch = '/((?:\s|^)';
+        $spaces = wp_spaces_regexp();
 
+        // Begin first "subpattern"
+        $wp_smiliessearch = '/(?<=' . $spaces . '|^)';
+
         $subchar = '';
         foreach ( (array) $wpsmiliestrans as $smiley => $img ) {
                 $firstchar = substr($smiley, 0, 1);
@@ -2570 +2573 @@
                 // new subpattern?
                 if ($firstchar != $subchar) {
                         if ($subchar != '') {
-                                $wp_smiliessearch .= ')(?=\s|$))|((?:\s|^)'; ;
+                                $wp_smiliessearch .= ')(?=' . $spaces . '|$)';  // End previous "subpattern"
+                                $wp_smiliessearch .= '|(?<=' . $spaces . '|^)'; // Begin another "subpattern"
                         }
                         $subchar = $firstchar;
                         $wp_smiliessearch .= preg_quote($firstchar, '/') . '(?:';
@@ -2580 +2584 @@
                 $wp_smiliessearch .= preg_quote($rest, '/');
         }
 
-        $wp_smiliessearch .= ')(?=\s|$))/m';
+        $wp_smiliessearch .= ')(?=' . $spaces . '|$)/m';
 
 }