From 398a02039e4c14ee01ad7853c8667c558f8d5f71 Mon Sep 17 00:00:00 2001
From: "Edward Z. Yang" <edwardzyang@thewritingpot.com>
Date: Fri, 20 Mar 2009 19:34:38 -0400
Subject: [PATCH] Implement %HTML.Attr.Name.UseCDATA which relaxes name
 validation rules.

Sponsored-by: Ian Cook <thinkspill@gmail.com>
Signed-off-by: Edward Z. Yang <edwardzyang@thewritingpot.com>
---
 NEWS                                               |   3 +++
 library/HTMLPurifier/AttrTransform/Name.php        |   2 ++
 library/HTMLPurifier/ConfigSchema/schema.ser       | Bin 11779 -> 11884 bytes
 .../schema/HTML.Attr.Name.UseCDATA.txt             |  11 ++++++++
 library/HTMLPurifier/HTMLModule/Name.php           |   4 ++-
 tests/HTMLPurifier/HTMLModule/NameTest.php         |  30 +++++++++++++++++++++
 6 files changed, 49 insertions(+), 1 deletion(-)
 rewrite library/HTMLPurifier/ConfigSchema/schema.ser (80%)
 create mode 100644 library/HTMLPurifier/ConfigSchema/schema/HTML.Attr.Name.UseCDATA.txt
 create mode 100644 tests/HTMLPurifier/HTMLModule/NameTest.php

diff --git a/NEWS b/NEWS
index fa2a65c3..655f76c1 100644
--- a/NEWS
+++ b/NEWS
@@ -17,6 +17,9 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 ! HTMLPurifier_Config::inherit($config) allows you to inherit one
   configuration, and have changes to that configuration be propagated
   to all of its children.
+! Implement %HTML.Attr.Name.UseCDATA, which relaxes validation rules on
+  the name attribute when set. Use with care. Thanks Ian Cook for
+  sponsoring.
 
 3.3.0, released 2009-02-16
 ! Implement CSS property 'overflow' when %CSS.AllowTricky is true.
diff --git a/library/HTMLPurifier/AttrTransform/Name.php b/library/HTMLPurifier/AttrTransform/Name.php
index e6f93aee..15315bc7 100644
--- a/library/HTMLPurifier/AttrTransform/Name.php
+++ b/library/HTMLPurifier/AttrTransform/Name.php
@@ -7,6 +7,8 @@ class HTMLPurifier_AttrTransform_Name extends HTMLPurifier_AttrTransform
 {
 
     public function transform($attr, $config, $context) {
+        // Abort early if we're using relaxed definition of name
+        if ($config->get('HTML.Attr.Name.UseCDATA')) return $attr;
         if (!isset($attr['name'])) return $attr;
         $id = $this->confiscateAttr($attr, 'name');
         if ( isset($attr['id']))   return $attr;
diff --git a/library/HTMLPurifier/ConfigSchema/schema.ser b/library/HTMLPurifier/ConfigSchema/schema.ser
dissimilarity index 80%
index bd8f501375714cf7256ea2596f6866198c6fb437..b8e5f347463af1469f7d0edf5aaac2eabe05319c 100644
GIT binary patch
delta 147
zcwTFcc@s0ijL~@WZN7UP3P#3ON**D;K6;KNB}ICEiMgqIp~b1rE{-9NO4dnM2Aj8u
jH?mHi%;!#mhE1%KZ}BN_?v{1n#IHQl%6#()HEVVN6%#L#

delta 58
zcwT!G(;PFwZ1Or;<;@EGdpIVWsEcp*=Um5%E|ShVxmMPY(Qxy6*(A=*Ch8XK0Bxlc
ALI3~&

diff --git a/library/HTMLPurifier/ConfigSchema/schema/HTML.Attr.Name.UseCDATA.txt b/library/HTMLPurifier/ConfigSchema/schema/HTML.Attr.Name.UseCDATA.txt
new file mode 100644
index 00000000..151fb7b8
--- /dev/null
+++ b/library/HTMLPurifier/ConfigSchema/schema/HTML.Attr.Name.UseCDATA.txt
@@ -0,0 +1,11 @@
+HTML.Attr.Name.UseCDATA
+TYPE: bool
+DEFAULT: false
+VERSION: 4.0.0
+--DESCRIPTION--
+The W3C specification DTD defines the name attribute to be CDATA, not ID, due
+to limitations of DTD.  In certain documents, this relaxed behavior is desired,
+whether it is to specify duplicate names, or to specify names that would be
+illegal IDs (for example, names that begin with a digit.) Set this configuration
+directive to true to use the relaxed parsing rules.
+--# vim: et sw=4 sts=4
diff --git a/library/HTMLPurifier/HTMLModule/Name.php b/library/HTMLPurifier/HTMLModule/Name.php
index af2b1195..05694b45 100644
--- a/library/HTMLPurifier/HTMLModule/Name.php
+++ b/library/HTMLPurifier/HTMLModule/Name.php
@@ -10,7 +10,9 @@ class HTMLPurifier_HTMLModule_Name extends HTMLPurifier_HTMLModule
         foreach ($elements as $name) {
             $element = $this->addBlankElement($name);
             $element->attr['name'] = 'CDATA';
-            $element->attr_transform_post['NameSync'] = new HTMLPurifier_AttrTransform_NameSync();
+            if (!$config->get('HTML.Attr.Name.UseCDATA')) {
+                $element->attr_transform_post['NameSync'] = new HTMLPurifier_AttrTransform_NameSync();
+            }
         }
     }
 
diff --git a/tests/HTMLPurifier/HTMLModule/NameTest.php b/tests/HTMLPurifier/HTMLModule/NameTest.php
new file mode 100644
index 00000000..55180b2e
--- /dev/null
+++ b/tests/HTMLPurifier/HTMLModule/NameTest.php
@@ -0,0 +1,30 @@
+<?php
+
+class HTMLPurifier_HTMLModule_NameTest extends HTMLPurifier_HTMLModuleHarness
+{
+
+    function setUp() {
+        parent::setUp();
+    }
+
+    function testBasicUse() {
+        $this->config->set('Attr.EnableID', true);
+        $this->assertResult(
+            '<a name="foo">bar</a>'
+        );
+    }
+
+    function testCDATA() {
+        $this->config->set('HTML.Attr.Name.UseCDATA', true);
+        $this->assertResult(
+            '<a name="2">Baz</a><a name="2">Bar</a>'
+        );
+    }
+
+    function testCDATAWithHeavyTidy() {
+        $this->config->set('HTML.Attr.Name.UseCDATA', true);
+        $this->config->set('HTML.TidyLevel', 'heavy');
+        $this->assertResult('<a name="2">Baz</a>');
+    }
+
+}
-- 
2.11.4.GIT