Implement HTML.Nofollow for external links.

Signed-off-by: Edward Z. Yang <ezyang@mit.edu>

NEWS

@@ -10,6 +10,7 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 ==========================
 
 4.2.1, unknown release date
+! Added %HTML.Nofollow to add rel="nofollow" to external links.
 - Make removal of conditional IE comments ungreedy; thanks Bernd
   for reporting.
 

configdoc/usage.xml

@@ -186,15 +186,20 @@
  </directive>
  <directive id="HTML.Proprietary">
   <file name="HTMLPurifier/HTMLModuleManager.php">
-   <line>221</line>
+   <line>220</line>
   </file>
  </directive>
  <directive id="HTML.SafeObject">
   <file name="HTMLPurifier/HTMLModuleManager.php">
-   <line>226</line>
+   <line>223</line>
   </file>
  </directive>
  <directive id="HTML.SafeEmbed">
+  <file name="HTMLPurifier/HTMLModuleManager.php">
+   <line>226</line>
+  </file>
+ </directive>
+ <directive id="HTML.Nofollow">
   <file name="HTMLPurifier/HTMLModuleManager.php">
    <line>229</line>
   </file>

library/HTMLPurifier.includes.php

@@ -125,6 +125,7 @@ require 'HTMLPurifier/AttrTransform/Lang.php';
 require 'HTMLPurifier/AttrTransform/Length.php';
 require 'HTMLPurifier/AttrTransform/Name.php';
 require 'HTMLPurifier/AttrTransform/NameSync.php';
+require 'HTMLPurifier/AttrTransform/Nofollow.php';
 require 'HTMLPurifier/AttrTransform/SafeEmbed.php';
 require 'HTMLPurifier/AttrTransform/SafeObject.php';
 require 'HTMLPurifier/AttrTransform/SafeParam.php';
@@ -151,6 +152,7 @@ require 'HTMLPurifier/HTMLModule/Image.php';
 require 'HTMLPurifier/HTMLModule/Legacy.php';
 require 'HTMLPurifier/HTMLModule/List.php';
 require 'HTMLPurifier/HTMLModule/Name.php';
+require 'HTMLPurifier/HTMLModule/Nofollow.php';
 require 'HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php';
 require 'HTMLPurifier/HTMLModule/Object.php';
 require 'HTMLPurifier/HTMLModule/Presentation.php';

library/HTMLPurifier.safe-includes.php

@@ -119,6 +119,7 @@ require_once $__dir . '/HTMLPurifier/AttrTransform/Lang.php';
 require_once $__dir . '/HTMLPurifier/AttrTransform/Length.php';
 require_once $__dir . '/HTMLPurifier/AttrTransform/Name.php';
 require_once $__dir . '/HTMLPurifier/AttrTransform/NameSync.php';
+require_once $__dir . '/HTMLPurifier/AttrTransform/Nofollow.php';
 require_once $__dir . '/HTMLPurifier/AttrTransform/SafeEmbed.php';
 require_once $__dir . '/HTMLPurifier/AttrTransform/SafeObject.php';
 require_once $__dir . '/HTMLPurifier/AttrTransform/SafeParam.php';
@@ -145,6 +146,7 @@ require_once $__dir . '/HTMLPurifier/HTMLModule/Image.php';
 require_once $__dir . '/HTMLPurifier/HTMLModule/Legacy.php';
 require_once $__dir . '/HTMLPurifier/HTMLModule/List.php';
 require_once $__dir . '/HTMLPurifier/HTMLModule/Name.php';
+require_once $__dir . '/HTMLPurifier/HTMLModule/Nofollow.php';
 require_once $__dir . '/HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php';
 require_once $__dir . '/HTMLPurifier/HTMLModule/Object.php';
 require_once $__dir . '/HTMLPurifier/HTMLModule/Presentation.php';

library/HTMLPurifier/AttrTransform/Nofollow.php

@@ -0,0 +1,41 @@
+<?php
+
+// must be called POST validation
+
+/**
+ * Adds rel="nofollow" to all outbound links.  This transform is
+ * only attached if Attr.Nofollow is TRUE.
+ */
+class HTMLPurifier_AttrTransform_Nofollow extends HTMLPurifier_AttrTransform
+{
+    private $parser;
+
+    public function __construct() {
+        $this->parser = new HTMLPurifier_URIParser();
+    }
+
+    public function transform($attr, $config, $context) {
+
+        if (!isset($attr['href'])) {
+            return $attr;
+        }
+
+        // XXX Kind of inefficient
+        $url = $this->parser->parse($attr['href']);
+        $scheme = $url->getSchemeObj($config, $context);
+
+        if (!is_null($url->host) && $scheme !== false && $scheme->browsable) {
+            if (isset($attr['rel'])) {
+                $attr['rel'] .= ' nofollow';
+            } else {
+                $attr['rel'] = 'nofollow';
+            }
+        }
+
+        return $attr;
+
+    }
+
+}
+
+// vim: et sw=4 sts=4

library/HTMLPurifier/ConfigSchema/schema/HTML.Nofollow.txt

@@ -0,0 +1,7 @@
+HTML.Nofollow
+TYPE: bool
+VERSION: 4.2.1
+DEFAULT: FALSE
+--DESCRIPTION--
+If enabled, nofollow rel attributes are added to all outgoing links.
+--# vim: et sw=4 sts=4

library/HTMLPurifier/HTMLModule/Nofollow.php

@@ -0,0 +1,19 @@
+<?php
+
+/**
+ * Module adds the nofollow attribute transformation to a tags.  It
+ * is enabled by HTML.Nofollow
+ */
+class HTMLPurifier_HTMLModule_Nofollow extends HTMLPurifier_HTMLModule
+{
+
+    public $name = 'Nofollow';
+
+    public function setup($config) {
+        $a = $this->addBlankElement('a');
+        $a->attr_transform_post[] = new HTMLPurifier_AttrTransform_Nofollow();
+    }
+
+}
+
+// vim: et sw=4 sts=4

library/HTMLPurifier/HTMLModuleManager.php

@@ -216,19 +216,19 @@ class HTMLPurifier_HTMLModuleManager
             }
         }
 
-        // add proprietary module (this gets special treatment because
-        // it is completely removed from doctypes, etc.)
+        // custom modules
         if ($config->get('HTML.Proprietary')) {
             $modules[] = 'Proprietary';
         }
-
-        // add SafeObject/Safeembed modules
         if ($config->get('HTML.SafeObject')) {
             $modules[] = 'SafeObject';
         }
         if ($config->get('HTML.SafeEmbed')) {
             $modules[] = 'SafeEmbed';
         }
+        if ($config->get('HTML.Nofollow')) {
+            $modules[] = 'Nofollow';
+        }
 
         // merge in custom modules
         $modules = array_merge($modules, $this->userModules);

tests/HTMLPurifier/HTMLModule/NofollowTest.php

@@ -0,0 +1,20 @@
+<?php
+
+class HTMLPurifier_HTMLModule_NofollowTest extends HTMLPurifier_HTMLModuleHarness
+{
+
+    function setUp() {
+        parent::setUp();
+        $this->config->set('HTML.Nofollow', true);
+    }
+
+    function testNofollow() {
+        $this->assertResult(
+            '<a href="http://google.com">a</a><a href="/local">b</a><a href="mailto:foo@example.com">c</a>',
+            '<a href="http://google.com" rel="nofollow">a</a><a href="/local">b</a><a href="mailto:foo@example.com">c</a>'
+        );
+    }
+
+}
+
+// vim: et sw=4 sts=4