Changeset 30450

Timestamp:

11/20/2014 02:29:03 PM (10 years ago)

Author:

nacin

Message:

Anchor texturize to shortcodes to improve regex efficiency.

For the 4.0 branch; see [30449] for trunk.

props miqrogroove.
see #29557 for segfault issues.

Location:

branches/4.0

Files:

• src/wp-includes/formatting.php (modified) (4 diffs)
• tests/phpunit/tests/formatting/WPTexturize.php (modified) (6 diffs)

branches/4.0/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 <.
+        . '(?:'         // Unroll the loop: Consume everything until --> is found.
+        .     '-(?!->)' // Dash not followed by end of comment.
+        .     '[^\-]*+' // Consume non-dashes.
+        . ')*+'         // Loop possessively.
+        . '-->';        // End of comment.
+
     $regex =  '/('          // Capture the entire match.
         .   '<'     // Find start of element.
         .   '(?(?=!--)' // Is this a comment?
-        .       '.+?--\s*>' // Find end of comment
+        .       $comment_regex  // Find end of comment
         .   '|'
         .       '[^>]+>'    // Find end of element
@@ -215 +227 @@
         . '|'
         .   '\['        // Find start of shortcode.
-        .   '\[?'       // Shortcodes may begin with [[
+        .   '[\/\[]?'   // Shortcodes may begin with [/ or [[
+        .   $tagregexp  // Only match registered shortcodes, because performance.
         .   '(?:'
-        .       '[^\[\]<>]' // Shortcodes do not contain other shortcodes.
+        .       '[^\[\]<>]+'    // Shortcodes do not contain other shortcodes. Quantifier critical.
         .   '|'
-        .       '<[^>]+>'   // HTML elements permitted. Prevents matching ] before >.
-        .   ')++'
+        .       '<[^\[\]>]*>'   // HTML elements permitted. Prevents matching ] before >.
+        .   ')*+'       // Possessive critical.
         .   '\]'        // Find end of shortcode.
         .   '\]?'       // Shortcodes may end with ]]
@@ -242 +255 @@
             continue;
 
-        } elseif ( '[' === $first && 1 === preg_match( '/^\[(?:[^\[\]<>]|<[^>]+>)++\]$/', $curl ) ) {
+        } elseif ( '[' === $first && 1 === preg_match( '/^\[\/?' . $tagregexp . '(?:[^\[\]<>]+|<[^\[\]>]*>)*+\]$/', $curl ) ) {
             // This is a shortcode delimiter.
 
             _wptexturize_pushpop_element( $curl, $no_texturize_shortcodes_stack, $no_texturize_shortcodes );
 
-        } elseif ( '[' === $first && 1 === preg_match( '/^\[\[?(?:[^\[\]<>]|<[^>]+>)++\]\]?$/', $curl ) ) {
+        } elseif ( '[' === $first && 1 === preg_match( '/^\[[\/\[]?' . $tagregexp . '(?:[^\[\]<>]+|<[^\[\]>]*>)*+\]\]?$/', $curl ) ) {
             // This is an escaped shortcode delimiter.
 

branches/4.0/tests/phpunit/tests/formatting/WPTexturize.php

@@ -12 +12 @@
     function test_disable() {
         $this->assertEquals('<pre>---</pre>', wptexturize('<pre>---</pre>'));
-        $this->assertEquals('[a]a&#8211;b[code]---[/code]a&#8211;b[/a]', wptexturize('[a]a--b[code]---[/code]a--b[/a]'));
         $this->assertEquals('<pre><code></code>--</pre>', wptexturize('<pre><code></code>--</pre>'));
 
@@ -1194 +1193 @@
             ),
             array(
+                '[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 &#8211; 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"> ] 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"> maybe </a> ]',
+            ),
+            array(
+                '[caption compare=">"]',
+                '[caption compare=&#8221;>&#8221;]',
+            ),
+            array(
+                '[caption compare="<>"]',
+                '[caption compare="<>"]',
+            ),
+            array(
+                '[caption compare="<" attr2="value" <!-->/]',
+                '[caption compare="<" attr2="value" <!-->/]',
+            ),
+            array(
+                '[caption compare="<"]',
+                '[caption compare=&#8221;<&#8221;]',
+            ),
+            array(
+                '[caption compare="<"]<br />',
+                '[caption compare=&#8221;<"]<br />',
+            ),
+            array(
                 '[ photos by <a href="http://example.com/?a[]=1&a[]=2"> this guy </a> ]',
                 '[ photos by <a href="http://example.com/?a[]=1&#038;a[]=2"> this guy </a> ]',
@@ -1211 +1246 @@
             array(
                 '[/...]', // This would actually be ignored by the shortcode system.  The decision to not texturize it is intentional, if not correct.
-                '[/...]',
+                '[/&#8230;]',
             ),
             array(
                 '[...]...[/...]', // These are potentially usable shortcodes.
-                '[...]&#8230;[/...]',
+                '[&#8230;]&#8230;[/&#8230;]',
             ),
             array(
                 '[[...]]...[[/...]]', // Shortcode parsing will ignore the inner ]...[ part and treat this as a single escaped shortcode.
-                '[[...]]&#8230;[[/...]]',
+                '[[&#8230;]]&#8230;[[/&#8230;]]',
             ),
             array(
                 '[[[...]]]...[[[/...]]]', // Again, shortcode parsing matches, but only the [[...] and [/...]] parts.
-                '[[[...]]]&#8230;[[[/...]]]',
+                '[[[&#8230;]]]&#8230;[[[/&#8230;]]]',
             ),
             array(
@@ -1231 +1266 @@
             array(
                 '[code]...[/code]]...', // These are potentially usable shortcodes.  Unfortunately, the meaning of [/code]] is ambiguous unless we run the entire shortcode regexp.
-                '[code]...[/code]]...', // This test would not pass in 3.9 because the extra brace was always ignored by texturize.
+                '[code]&#8230;[/code]]&#8230;', // This test would not pass in 3.9 because the extra brace was always ignored by texturize.
             ),
             array(
@@ -1346 +1381 @@
             ),
             array(
-                '[Let\'s get crazy<input>[plugin code="<a href=\'?a[]=100\'>hello</a>"]</input>world]',
-                '[Let&#8217;s get crazy<input>[plugin code="<a href=\'?a[]=100\'>hello</a>"]</input>world]',
+                '[Let\'s get crazy<input>[caption code="<a href=\'?a[]=100\'>hello</a>"]</input>world]', // caption shortcode is invalid here because it contains [] chars.
+                '[Let&#8217;s get crazy<input>[caption code=&#8221;<a href=\'?a[]=100\'>hello</a>&#8220;]</input>world]',
             ),
         );
@@ -1716 +1751 @@
             array(
                 '<span>hello[code]---</span>',
-                '<span>hello[code]---</span>',
+                '<span>hello[code]&#8212;</span>',
             ),
             array(
                 '[code]hello<span>---</span>',
-                '[code]hello<span>---</span>',
+                '[code]hello<span>&#8212;</span>',
             ),
             array(
                 '[code]hello</span>---</span>',
-                '[code]hello</span>---</span>',
+                '[code]hello</span>&#8212;</span>',
+            ),
+        );
+    }
+
+    /**
+     * Test disabling shortcode texturization.
+     *
+     * @ticket 29557
+     * @dataProvider data_unregistered_shortcodes
+     */
+    function test_unregistered_shortcodes( $input, $output ) {
+        add_filter( 'no_texturize_shortcodes', array( $this, 'filter_shortcodes' ), 10, 1 );
+    
+        $output = $this->assertEquals( $output, wptexturize( $input ) );
+    
+        remove_filter( 'no_texturize_shortcodes', array( $this, 'filter_shortcodes' ), 10, 1 );
+        return $output;
+    }
+    
+    function filter_shortcodes( $disabled ) {
+        $disabled[] = 'audio';
+        return $disabled;
+    }
+
+    function data_unregistered_shortcodes() {
+        return array(
+            array(
+                '[a]a--b[code]---[/code]a--b[/a]', // code is not a registered shortcode.
+                '[a]a&#8211;b[code]&#8212;[/code]a&#8211;b[/a]',
+            ),
+            array(
+                '[a]a--b[audio]---[/audio]a--b[/a]',
+                '[a]a&#8211;b[audio]---[/audio]a&#8211;b[/a]',
+            ),
+            array(
+                '[code ...]...[/code]', // code is not a registered shortcode.
+                '[code &#8230;]&#8230;[/code]',
+            ),
+            array(
+                '[hello ...]...[/hello]', // hello is not a registered shortcode.
+                '[hello &#8230;]&#8230;[/hello]',
+            ),
+            array(
+                '[...]...[/...]', // These are potentially usable shortcodes.
+                '[&#8230;]&#8230;[/&#8230;]',
+            ),
+            array(
+                '[gal>ery ...]',
+                '[gal>ery &#8230;]',
+            ),
+            array(
+                '[randomthing param="test"]',
+                '[randomthing param=&#8221;test&#8221;]',
+            ),
+            array(
+                '[[audio]...[/audio]...', // These are potentially usable shortcodes.  Unfortunately, the meaning of [[audio] is ambiguous unless we run the entire shortcode regexp.
+                '[[audio]&#8230;[/audio]&#8230;',
+            ),
+            array(
+                '[audio]...[/audio]]...', // These are potentially usable shortcodes.  Unfortunately, the meaning of [/audio]] is ambiguous unless we run the entire shortcode regexp.
+                '[audio]...[/audio]]...', // This test would not pass in 3.9 because the extra brace was always ignored by texturize.
+            ),
+            array(
+                '<span>hello[/audio]---</span>',
+                '<span>hello[/audio]&#8212;</span>',
+            ),
+            array(
+                '[/audio]hello<span>---</span>',
+                '[/audio]hello<span>&#8212;</span>',
+            ),
+            array(
+                '[audio]hello[/audio]---</span>',
+                '[audio]hello[/audio]&#8212;</span>',
+            ),
+            array(
+                '<span>hello</span>---[audio]',
+                '<span>hello</span>&#8212;[audio]',
+            ),
+            array(
+                '<span>hello[audio]---</span>',
+                '<span>hello[audio]---</span>',
+            ),
+            array(
+                '[audio]hello<span>---</span>',
+                '[audio]hello<span>---</span>',
+            ),
+            array(
+                '[audio]hello</span>---</span>',
+                '[audio]hello</span>---</span>',
             ),
         );