Ticket #35022: 35022.4.patch

src/wp-includes/formatting.php

@@ -25 +25 @@
  * @since 0.71
  *
  * @global array $wp_cockneyreplace Array of formatted entities for certain common phrases
- * @global array $shortcode_tags
  * @staticvar array $static_characters
  * @staticvar array $static_replacements
  * @staticvar array $dynamic_characters
@@ -39 +38 @@
  * @return string The string replaced with html entities
  */
 function wptexturize( $text, $reset = false ) {
-        global $wp_cockneyreplace, $shortcode_tags;
+        global $wp_cockneyreplace;
         static $static_characters = null,
                 $static_replacements = null,
                 $dynamic_characters = null,
@@ -216 +215 @@
 
         // Look for shortcodes and HTML elements.
 
-        preg_match_all( '@\[/?([^<>&/\[\]\x00-\x20=]++)@', $text, $matches );
-        $tagnames = array_intersect( array_keys( $shortcode_tags ), $matches[1] );
+        $tagnames = get_shortcode_tagnames( $text );
         $found_shortcodes = ! empty( $tagnames );
         $shortcode_regex = $found_shortcodes ? _get_wptexturize_shortcode_regex( $tagnames ) : '';
         $regex = _get_wptexturize_split_regex( $shortcode_regex );
@@ -696 +694 @@
  */
 function _get_wptexturize_shortcode_regex( $tagnames ) {
         $tagregexp = join( '|', array_map( 'preg_quote', $tagnames ) );
-        $tagregexp = "(?:$tagregexp)(?=[\\s\\]\\/])"; // Excerpt of get_shortcode_regex().
+        $tagregexp = "(?:$tagregexp)(?=" . shortcode_name_terminators() . ")"; // Excerpt of get_shortcode_regex().
         $regex =
                   '\['              // Find start of shortcode.
                 . '[\/\[]?'         // Shortcodes may begin with [/ or [[
@@ -782 +780 @@
  *
  * @since 2.9.0
  *
- * @global array $shortcode_tags
- *
  * @param string $pee The content.
  * @return string The filtered content.
  */
 function shortcode_unautop( $pee ) {
-        global $shortcode_tags;
 
-        if ( empty( $shortcode_tags ) || !is_array( $shortcode_tags ) ) {
+        $tagnames = get_shortcode_tagnames( $pee );
+        if ( empty( $tagnames ) ) {
                 return $pee;
         }
 
-        $tagregexp = join( '|', array_map( 'preg_quote', array_keys( $shortcode_tags ) ) );
+        $tagregexp = join( '|', array_map( 'preg_quote', $tagnames ) );
         $spaces = wp_spaces_regexp();
 
         $pattern =
@@ -804 +800 @@
                 . '('                                // 1: The shortcode
                 .     '\\['                          // Opening bracket
                 .     "($tagregexp)"                 // 2: Shortcode name
-                .     '(?![\\w-])'                   // Not followed by word character or hyphen
+                .     '(?=' . shortcode_name_terminators() . ')' // Followed by a shortcode name terminator.
                                                      // Unroll the loop: Inside the opening shortcode tag
                 .     '[^\\]\\/]*'                   // Not a closing bracket or forward slash
                 .     '(?:'

src/wp-includes/js/shortcode.js

@@ -102 +102 @@
                 // 6. The closing tag.
                 // 7. An extra `]` to allow for escaping shortcodes with double `[[]]`
                 regexp: _.memoize( function( tag ) {
-                        return new RegExp( '\\[(\\[?)(' + tag + ')(?![\\w-])([^\\]\\/]*(?:\\/(?!\\])[^\\]\\/]*)*?)(?:(\\/)\\]|\\](?:([^\\[]*(?:\\[(?!\\/\\2\\])[^\\[]*)*)(\\[\\/\\2\\]))?)(\\]?)', 'g' );
+                        var shortcode_name_terminators = '[<>&\\/\\[\\]\x00-\x20=\u00a0]';
+                        return new RegExp( '\\[(\\[?)(' + tag + ')(?=' + shortcode_name_terminators + ')([^\\]\\/]*(?:\\/(?!\\])[^\\]\\/]*)*?)(?:(\\/)\\]|\\](?:([^\\[]*(?:\\[(?!\\/\\2\\])[^\\[]*)*)(\\[\\/\\2\\]))?)(\\]?)', 'g' );
                 }),
 
 

src/wp-includes/shortcodes.php

@@ -95 +95 @@
                 return;
         }
 
-        if ( 0 !== preg_match( '@[<>&/\[\]\x00-\x20=]@', $tag ) ) {
+        if ( 0 !== preg_match( '@' . shortcode_name_terminators() . '@', $tag ) ) {
                 /* translators: 1: shortcode name, 2: space separated list of reserved characters */
                 $message = sprintf( __( 'Invalid shortcode name: %1$s. Do not use spaces or reserved characters: %2$s' ), $tag, '& / < > [ ] =' );
                 _doing_it_wrong( __FUNCTION__, $message, '4.4.0' );
@@ -157 +157 @@
  *
  * @since 3.6.0
  *
- * @global array $shortcode_tags
- *
  * @param string $content Content to search for shortcodes.
  * @param string $tag     Shortcode tag to check.
  * @return bool Whether the passed content contains the given shortcode.
@@ -193 +191 @@
  *
  * @since 2.5.0
  *
- * @global array $shortcode_tags List of shortcode tags and their callback hooks.
- *
  * @param string $content Content to search for shortcodes.
  * @param bool $ignore_html When true, shortcodes inside HTML elements will be skipped.
  * @return string Content with shortcodes filtered out.
  */
 function do_shortcode( $content, $ignore_html = false ) {
-        global $shortcode_tags;
 
-        if ( false === strpos( $content, '[' ) ) {
-                return $content;
-        }
-
-        if (empty($shortcode_tags) || !is_array($shortcode_tags))
-                return $content;
-
         // Find all registered tag names in $content.
-        preg_match_all( '@\[([^<>&/\[\]\x00-\x20=]++)@', $content, $matches );
-        $tagnames = array_intersect( array_keys( $shortcode_tags ), $matches[1] );
+        $tagnames = get_shortcode_tagnames( $content );
 
         if ( empty( $tagnames ) ) {
                 return $content;
@@ -245 +232 @@
  *
  * @since 2.5.0
  *
- * @global array $shortcode_tags
- *
  * @param array $tagnames List of shortcodes to find. Optional. Defaults to all registered shortcodes.
  * @return string The shortcode search regular expression
  */
 function get_shortcode_regex( $tagnames = null ) {
-        global $shortcode_tags;
 
         if ( empty( $tagnames ) ) {
-                $tagnames = array_keys( $shortcode_tags );
+                $tagnames = get_shortcode_tagnames();
         }
         $tagregexp = join( '|', array_map('preg_quote', $tagnames) );
 
@@ -264 +248 @@
                   '\\['                              // Opening bracket
                 . '(\\[?)'                           // 1: Optional second opening bracket for escaping shortcodes: [[tag]]
                 . "($tagregexp)"                     // 2: Shortcode name
-                . '(?![\\w-])'                       // Not followed by word character or hyphen
+                . '(?=' . shortcode_name_terminators() . ')' // Followed by a shortcode name terminator.
                 . '('                                // 3: Unroll the loop: Inside the opening shortcode tag
                 .     '[^\\]\\/]*'                   // Not a closing bracket or forward slash
                 .     '(?:'
@@ -292 +276 @@
 }
 
 /**
+ * Return regular expression of characters that terminate the name of a shortcode.
+ * Dependent on charset of blog.
+ *
+ * @since 4.7.0
+ *
+ * @return string Regular expression of terminating characters.
+ */
+function shortcode_name_terminators() {
+        // Needs to be synced with get_shortcode_tagnames() expression below.
+        if ( 'UTF-8' === _canonical_charset( get_option( 'blog_charset' ) ) ) {
+                return '(?:[<>&\/\[\]\x00-\x20=]|\xc2\xa0)';
+        }
+        return '[<>&\/\[\]\x00-\x20=]';
+}
+
+/**
+ * Return registered shortcode tagnames, reduced to those in $text if given.
+ *
+ * @since 4.7.0
+ *
+ * @global array $shortcode_tags
+ *
+ * @param string $text Optional. If given, only those registered shortcodes that appear in $text are returned. 
+ * @return array Array of shortcode names.
+ */
+function get_shortcode_tagnames( $text = null ) {
+        global $shortcode_tags;
+
+        if ( ! $shortcode_tags || ! is_array( $shortcode_tags ) ) {
+                return array();
+        }
+        if ( null === $text ) {
+                return array_keys( $shortcode_tags );
+        }
+        if ( false === strpos( $text, '[' ) ) {
+                return array();
+        }
+
+        // Grouped expressions with unlimited repetition cause seg faults or match failures in PCRE <= 8.12 so use simple character class first and then post-process the matches.
+        if ( ! preg_match_all( '/\[(?!\[)\/?\K[^<>&\/\[\]\x00-\x20=]+/', $text, $matches ) ) { // Needs to be synced with shortcode_name_terminators() expression above.
+                return array();
+        }
+        if ( false !== strpos( $text, "\xc2\xa0" ) && 'UTF-8' === _canonical_charset( get_option( 'blog_charset' ) ) ) {
+                $matches[0] = array_map( '_shortcode_name_terminator_chop_cb', $matches[0] );
+        }
+
+        return array_values( array_intersect( array_keys( $shortcode_tags ), $matches[0] ) );
+}
+
+/**
+ * Callback for array_map in {@see get_shortcode_tagnames()} to chop off UTF-8 no-break space from potential shortcode name.
+ *
+ * @since 4.7.0
+ * @access private
+ *
+ * @param string $entry The potential shortcode name match.
+ * @return string The potential shortcode name chopped down if necessary.
+ */
+function _shortcode_name_terminator_chop_cb( $entry ) {
+        if ( false !== ( $pos = strpos( $entry, "\xc2\xa0" ) ) ) {
+                $entry = substr( $entry, 0, $pos );
+        }
+        return $entry;
+}
+
+/**
  * Regular Expression callable for do_shortcode() for calling shortcode hook.
  * @see get_shortcode_regex for details of the match array contents.
  *
@@ -580 +630 @@
  *
  * @since 2.5.0
  *
- * @global array $shortcode_tags
- *
  * @param string $content Content to remove shortcode tags.
  * @return string Content without shortcode tags.
  */
 function strip_shortcodes( $content ) {
-        global $shortcode_tags;
 
-        if ( false === strpos( $content, '[' ) ) {
-                return $content;
-        }
-
-        if (empty($shortcode_tags) || !is_array($shortcode_tags))
-                return $content;
-
         // Find all registered tag names in $content.
-        preg_match_all( '@\[([^<>&/\[\]\x00-\x20=]++)@', $content, $matches );
-        $tagnames = array_intersect( array_keys( $shortcode_tags ), $matches[1] );
+        $tagnames = get_shortcode_tagnames( $content );
 
         if ( empty( $tagnames ) ) {
                 return $content;

tests/phpunit/tests/shortcode.php

@@ -332 +332 @@
         }
 
         /**
+         * @ticket 35022
+         */
+        function test_no_break_space_following_shortcode_tag() {
+                do_shortcode( "[test-shortcode-tag\xC2\xA0foo=\"bar\"]" );
+                $this->assertSame( 'test-shortcode-tag', $this->tagname );
+                $this->assertSame( array( 'foo' => 'bar' ), $this->atts );
+        }
+
+        /**
+         * @ticket 35022
+         */
+        function test_non_no_break_space_following_shortcode_tag() {
+                do_shortcode( "[test-shortcode-tag\xC2\xA1foo=\"bar\"]" );
+                $this->assertNull( $this->tagname );
+                $this->assertNull( $this->atts );
+        }
+
+        function _charset_iso_8859_1() {
+                return 'iso-8859-1';
+        }
+
+        /**
+         * @ticket 35022
+         */
+        function test_no_break_space_following_shortcode_tag_iso_8859_1() {
+                add_filter( 'pre_option_blog_charset', array( $this, '_charset_iso_8859_1' ) );
+
+                do_shortcode( "[test-shortcode-tag foo=\"bar\"]" );
+                $this->assertSame( 'test-shortcode-tag', $this->tagname );
+                $this->assertSame( array( 'foo' => 'bar' ), $this->atts );
+
+                $this->atts = $this->content = $this->tagname = null;
+                do_shortcode( "[test-shortcode-tag\xA0foo=\"bar\"]" ); // ISO-8859-1 no-break space not supported currently.
+                $this->assertNull( $this->tagname );
+                $this->assertNull( $this->atts );
+
+                $this->atts = $this->content = $this->tagname = null;
+                do_shortcode( "[test-shortcode-tag\xC2\xA0foo=\"bar\"]" );
+                $this->assertNull( $this->tagname );
+                $this->assertNull( $this->atts );
+
+                remove_filter( 'pre_option_blog_charset', array( $this, '_charset_iso_8859_1' ) );
+        }
+
+        /**
          * @ticket 14050
          */
         function test_shortcode_unautop() {
@@ -608 +653 @@
                                 '',
                                 false,
                         ),
+                        array(
+                                "bad\xc2\xa0space",
+                                false,
+                        ),
                 );
         }
 
@@ -625 +674 @@
                                 'unreserved!#$%()*+,-.;?@^_{|}~chars',
                                 true,
                         ),
+                        array(
+                                "unreservedutf8\xc2\xa1\xe0\xa0\x80\xf0\xa0\x80\x80\xc2\xa2chars",
+                                true,
+                        ),
                 );
         }
 

tests/qunit/wp-includes/js/shortcode.js

@@ -96 +96 @@
                 equal( result, undefined, 'stub does not trigger match' );
         });
 
+        test( 'next() should find the shortcode delimited by no-break space (U+00A0)', function() {
+                var result;
+
+
+                result = wp.shortcode.next( 'foo', 'this has a no-break space delimited [foo\u00a0param="foo"] shortcode' );
+                equal( result.index, 36, 'no-break space delimited foo shortcode found at index 36' );
+        });
+
         test( 'replace() should replace the shortcode', function() {
                 var result;