Changeset 30449

Timestamp:

11/20/2014 02:27:08 PM (12 years ago)

Author:

nacin

Message:

Anchor texturize to shortcodes to improve regex efficiency.

props miqrogroove.
see #29557 for segfault issues.

Location:

trunk

Files:

• src/wp-includes/formatting.php (modified) (3 diffs)
• tests/phpunit/tests/formatting/WPTexturize.php (modified) (2 diffs)

trunk/src/wp-includes/formatting.php

@@ -29 +29 @@
  */
 function wptexturize($text, $reset = false) {
-        global $wp_cockneyreplace;
+        global $wp_cockneyreplace, $shortcode_tags;
         static $static_characters, $static_replacements, $dynamic_characters, $dynamic_replacements,
                 $default_no_texturize_tags, $default_no_texturize_shortcodes, $run_texturize = true;
@@ -206 +206 @@
         // Look for shortcodes and HTML elements.
 
+        $tagnames = array_keys( $shortcode_tags );
+        $tagregexp = join( '|', array_map( 'preg_quote', $tagnames ) );
+        $tagregexp = "(?:$tagregexp)(?![\\w-])"; // Excerpt of get_shortcode_regex().
+
         $comment_regex =
                   '!'           // Start of comment, after the <.
@@ -215 +219 @@
 
         $shortcode_regex =
-                  '\['          // Find start of shortcode.
-                . '[\/\[]?'     // Shortcodes may begin with [/ or [[
-                . '[^\s\/\[\]]' // No whitespace before name.
-                . '[^\[\]]*+'   // Shortcodes do not contain other shortcodes. Possessive critical.
-                . '\]'          // Find end of shortcode.
-                . '\]?';        // Shortcodes may end with ]]
+                  '\['              // Find start of shortcode.
+                . '[\/\[]?'         // Shortcodes may begin with [/ or [[
+                . $tagregexp        // Only match registered shortcodes, because performance.
+                . '(?:'
+                .     '[^\[\]<>]+'  // Shortcodes do not contain other shortcodes. Quantifier critical.
+                . '|'
+                .     '<[^\[\]>]*>' // HTML elements permitted. Prevents matching ] before >.
+                . ')*+'             // Possessive critical.
+                . '\]'              // Find end of shortcode.
+                . '\]?';            // Shortcodes may end with ]]
 
         $regex =

trunk/tests/phpunit/tests/formatting/WPTexturize.php

@@ -1197 +1197 @@
                         ),
                         array(
-                                '[is it wise to <a title="allow user content ] here? hmm"> maybe </a> ]', // HTML corruption is a known bug.  See tickets #12690 and #29557.
-                                '[is it wise to <a title="allow user content ] here? hmm&#8221;> maybe </a> ]',
+                                '[is it wise to <a title="allow user content ] here? hmm"> maybe </a> ]',
+                                '[is it wise to <a title="allow user content ] here? hmm"> maybe </a> ]',
                         ),
                         array(
                                 '[caption - is it wise to <a title="allow user content ] here? hmm"> maybe </a> ]',
-                                '[caption - is it wise to <a title="allow user content ] here? hmm&#8221;> maybe </a> ]',
+                                '[caption &#8211; is it wise to <a title="allow user content ] here? hmm"> maybe </a> ]',
                         ),
                         array(
@@ -1768 +1768 @@
                         array(
                                 '[code ...]...[/code]', // code is not a registered shortcode.
-                                '[code ...]...[/code]',
+                                '[code &#8230;]&#8230;[/code]',
                         ),
                         array(
                                 '[hello ...]...[/hello]', // hello is not a registered shortcode.
-                                '[hello ...]&#8230;[/hello]',
+                                '[hello &#8230;]&#8230;[/hello]',
                         ),
                         array(
                                 '[...]...[/...]', // These are potentially usable shortcodes.
-                                '[...]&#8230;[/...]',
+                                '[&#8230;]&#8230;[/&#8230;]',
                         ),
                         array(
                                 '[gal>ery ...]',
-                                '[gal>ery ...]',
+                                '[gal>ery &#8230;]',
                         ),
                         array(
                                 '[randomthing param="test"]',
-                                '[randomthing param="test"]',
+                                '[randomthing param=&#8221;test&#8221;]',
                         ),
                         array(