URL Whitelist MediaWiki Modification

The following code was created by Daniel Robbins to provide URL Whitelist functionality for MediaWiki. This code allows you to create a page named "URLWhitelist" which contains lists of domains that users can link against. All domains not listed on the URLWhitelist page will result in a Spam Protection error when saving any edits to a page.

On the URLWhitelist page, the wikitext should contain one domain per line, and sub-paths are fine, such as this:

google.com
foo.bar.com/oni

Above, the URLWhitelist will allow http(s)://[domain.]google.com/* and http(s)://[domain.]foo.bar.com/oni[/optional-path]

The following patch will apply against MediaWiki-1.19.1:

--- mediawiki-1.19.1/includes/EditPage.php      2012-06-13 18:22:38.000000000 +0000
+++ public_html/includes/EditPage.php   2012-08-27 04:53:33.073996581 +0000
@@ -1081,7 +1081,9 @@
                # Check for spam
                $match = self::matchSummarySpamRegex( $this->summary );
                if ( $match === false ) {
-                       $match = self::matchSpamRegex( $this->textbox1 );
+                       /* Don't run spam regex matching on the URLWhitelist page itself: */
+                       if ($this->mTitle->getText() != "URLWhitelist")
+                               $match = self::matchSpamRegex( $this->textbox1 );
                }
                if ( $match !== false ) {
                        $result['spam'] = $match;
@@ -1558,7 +1560,7 @@
                global $wgSpamRegex;
                // For back compatibility, $wgSpamRegex may be a single string or an array of regexes.
                $regexes = (array)$wgSpamRegex;
-               return self::matchSpamRegexInternal( $text, $regexes );
+               return self::matchURLWhitelistInternal( $text, $regexes );
        }
 
        /**
@@ -1571,7 +1573,7 @@
        public static function matchSummarySpamRegex( $text ) {
                global $wgSummarySpamRegex;
                $regexes = (array)$wgSummarySpamRegex;
-               return self::matchSpamRegexInternal( $text, $regexes );
+               return self::matchURLWhitelistInternal( $text, $regexes );
        }
 
        /**
@@ -1589,6 +1591,46 @@
                return false;
        }
 
+       protected static function matchURLWhitelistInternal( $text, $regexes ) {
+               global $wgUser;
+               /*if ( in_array('sysop', $wgUser->getGroups())) {
+                       return false;
+               }*/
+               $a = new Article( Title::newFromText('URLWhitelist'));
+               $stuff = $a->getContent();
+               $foo = preg_split('/\s+/', $stuff);
+               /*$httpreg = ",http://[^\s/]*(?!";*/
+               $httpreg = ",https?://(?!";
+               $count = 0;
+               foreach($foo as $reg) {
+                       $reg = str_replace(".","\.",$reg);
+                       $reg = "\S*" . $reg;
+                       if ($count != 0 )
+                               $httpreg .= "|";
+                       $httpreg .= $reg;
+                       $count = $count + 1;
+               }
+               $httpreg .= "),";
+               $hmatches = array();
+               /* URL match with whitelist */
+               if ( preg_match( $httpreg, $text, $hmatches ) ) {
+                       return "non-whitelisted http/https URL";
+               }
+               /*
+               for testing:
+               if ( $wgUser->getName() == "Drobbins" )
+                       return $httpreg;
+               */
+               foreach( $regexes as $regex ) {
+                       $matches = array();
+                       if( preg_match( $regex, $text, $matches ) ) {
+                               return $matches[0];
+                       }
+               }
+               return false;
+       }
+
+
        function setHeaders() {
                global $wgOut, $wgUser;