Ticket #27588: prior-work.patch

src/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 );
 
-                /*
-                 * Regex 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 #WP22692, that sequence
-                 * was found to be unreliable due to random inclusion of the A0 byte.
-                 */
-                $spaces = '[\r\n\t ]|\xC2\xA0| ';
+                $spaces = wp_spaces_regexp();
 
 
                 // Pattern-based replacements of characters.
@@ -361 +354 @@
         }
 
         $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
@@ -390 +384 @@
                 .         ')?'
                 .     ')'
                 . ')'
-                . '\\s*+'                            // optional trailing whitespace
+                . '(?:' . $spaces . ')*+'            // optional trailing whitespace
                 . '<\\/p>'                           // closing paragraph
                 . '/s';
 
@@ -3825 +3819 @@
 
         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 4.0.0
+ *
+ * @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;
+}

tests/phpunit/tests/shortcode.php

@@ -373 +373 @@
                 remove_filter( 'shortcode_atts_bartag', array( $this, '_filter_atts2' ), 10, 3 );
         }
 
+        /**
+         * Check that shortcode_unautop() will always recognize spaces around shortcodes.
+         *
+         * @ticket 22692
+         */
+        function test_spaces_around_shortcodes() {
+                $nbsp = "\xC2\xA0";
+
+                $input  = array();
+
+                $input[] = "<p>[gallery ids=\"37,15,11\"]</p>";
+                $input[] = "<p> [gallery ids=\"37,15,11\"] </p>";
+                $input[] = "<p> {$nbsp}[gallery ids=\"37,15,11\"] {$nbsp}</p>";
+                $input[] = "<p> &nbsp;[gallery ids=\"37,15,11\"] &nbsp;</p>";
+
+                $output = "[gallery ids=\"37,15,11\"]";
+
+                foreach($input as $in) {
+                        $this->assertEquals( $output, shortcode_unautop( $in ) );
+                }
+        }
 }