Implement %Attr.AllowedClasses and %Attr.ForbiddenClasses.

Signed-off-by: Edward Z. Yang <edwardzyang@thewritingpot.com>

library/HTMLPurifier/AttrDef/HTML/Class.php

@@ -0,0 +1,22 @@
+<?php
+
+/**
+ * Implements special behavior for class attribute (normally NMTOKENS)
+ */
+class HTMLPurifier_AttrDef_HTML_Class extends HTMLPurifier_AttrDef_HTML_Nmtokens
+{
+    protected function filter($tokens, $config, $context) {
+        $allowed = $config->get('Attr.AllowedClasses');
+        $forbidden = $config->get('Attr.ForbiddenClasses');
+        $ret = array();
+        foreach ($tokens as $token) {
+            if (
+                ($allowed === null || isset($allowed[$token])) &&
+                !isset($forbidden[$token])
+            ) {
+                $ret[] = $token;
+            }
+        }
+        return $ret;
+    }
+}

library/HTMLPurifier/AttrDef/HTML/Nmtokens.php

@@ -2,10 +2,6 @@
 
 /**
  * Validates contents based on NMTOKENS attribute type.
- * @note The only current use for this is the class attribute in HTML
- * @note Could have some functionality factored out into Nmtoken class
- * @warning We cannot assume this class will be used only for 'class'
- *          attributes. Not sure how to hook in magic behavior, then.
  */
 class HTMLPurifier_AttrDef_HTML_Nmtokens extends HTMLPurifier_AttrDef
 {
@@ -17,6 +13,17 @@ class HTMLPurifier_AttrDef_HTML_Nmtokens extends HTMLPurifier_AttrDef
         // early abort: '' and '0' (strings that convert to false) are invalid
         if (!$string) return false;
 
+        $tokens = $this->split($string);
+        $tokens = $this->filter($tokens, $config, $context);
+        if (empty($tokens)) return false;
+        return implode(' ', $tokens);
+
+    }
+
+    /**
+     * Splits a space separated list of tokens into its constituent parts.
+     */
+    protected function split($string) {
         // OPTIMIZABLE!
         // do the preg_match, capture all subpatterns for reformulation
 
@@ -24,23 +31,20 @@ class HTMLPurifier_AttrDef_HTML_Nmtokens extends HTMLPurifier_AttrDef
         // escaping because I don't know how to do that with regexps
         // and plus it would complicate optimization efforts (you never
         // see that anyway).
-        $matches = array();
         $pattern = '/(?:(?<=\s)|\A)'. // look behind for space or string start
                    '((?:--|-?[A-Za-z_])[A-Za-z_\-0-9]*)'.
                    '(?:(?=\s)|\z)/'; // look ahead for space or string end
         preg_match_all($pattern, $string, $matches);
+        return $matches[1];
+    }
 
-        if (empty($matches[1])) return false;
-
-        // reconstruct string
-        $new_string = '';
-        foreach ($matches[1] as $token) {
-            $new_string .= $token . ' ';
-        }
-        $new_string = rtrim($new_string);
-
-        return $new_string;
-
+    /**
+     * Template method for removing certain tokens based on arbitrary criteria.
+     * @note If we wanted to be really functional, we'd do an array_filter
+     *       with a callback. But... we're not.
+     */
+    protected function filter($tokens, $config, $context) {
+        return $tokens;
     }
 
 }