Make forms work for transitional doctypes.

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

NEWS

@@ -23,6 +23,7 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 ! Implement iframes, and allow them to be used in untrusted mode with
   %HTML.SafeIframe and %URI.SafeIframeRegexp.  Thanks Bradley M. Froehle
   <brad.froehle@gmail.com> for submitting an initial version of the patch.
+! The Forms module now works properly for transitional doctypes.
 - Color keywords are now case insensitive.  Thanks Yzmir Ramirez
   <yramirez-htmlpurifier@adicio.com> for reporting.
 - Explicitly initialize anonModule variable to null.

library/HTMLPurifier.includes.php

@@ -73,6 +73,7 @@
 require 'HTMLPurifier/VarParser.php';
 require 'HTMLPurifier/VarParserException.php';
 require 'HTMLPurifier/AttrDef/CSS.php';
+require 'HTMLPurifier/AttrDef/Clone.php';
 require 'HTMLPurifier/AttrDef/Enum.php';
 require 'HTMLPurifier/AttrDef/Integer.php';
 require 'HTMLPurifier/AttrDef/Lang.php';

library/HTMLPurifier.safe-includes.php

@@ -67,6 +67,7 @@
 require_once $__dir . '/HTMLPurifier/VarParser.php';
 require_once $__dir . '/HTMLPurifier/VarParserException.php';
 require_once $__dir . '/HTMLPurifier/AttrDef/CSS.php';
+require_once $__dir . '/HTMLPurifier/AttrDef/Clone.php';
 require_once $__dir . '/HTMLPurifier/AttrDef/Enum.php';
 require_once $__dir . '/HTMLPurifier/AttrDef/Integer.php';
 require_once $__dir . '/HTMLPurifier/AttrDef/Lang.php';

library/HTMLPurifier/AttrDef/Clone.php

@@ -0,0 +1,28 @@
+<?php
+
+/**
+ * Dummy AttrDef that mimics another AttrDef, BUT it generates clones
+ * with make.
+ */
+class HTMLPurifier_AttrDef_Clone extends HTMLPurifier_AttrDef
+{
+    /**
+     * What we're cloning
+     */
+    protected $clone;
+
+    public function __construct($clone) {
+        $this->clone = $clone;
+    }
+
+    public function validate($v, $config, $context) {
+        return $this->clone->validate($v, $config, $context);
+    }
+
+    public function make($string) {
+        return clone $this->clone;
+    }
+
+}
+
+// vim: et sw=4 sts=4

library/HTMLPurifier/AttrTypes.php

@@ -15,6 +15,13 @@ class HTMLPurifier_AttrTypes
      * types.
      */
     public function __construct() {
+        // XXX This is kind of poor, since we don't actually /clone/
+        // instances; instead, we use the supplied make() attribute. So,
+        // the underlying class must know how to deal with arguments.
+        // With the old implementation of Enum, that ignored its
+        // arguments when handling a make dispatch, the IAlign
+        // definition wouldn't work.
+
         // pseudo-types, must be instantiated via shorthand
         $this->info['Enum']    = new HTMLPurifier_AttrDef_Enum();
         $this->info['Bool']    = new HTMLPurifier_AttrDef_HTML_Bool();
@@ -29,6 +36,9 @@ public function __construct() {
         $this->info['URI']      = new HTMLPurifier_AttrDef_URI();
         $this->info['LanguageCode'] = new HTMLPurifier_AttrDef_Lang();
         $this->info['Color']    = new HTMLPurifier_AttrDef_HTML_Color();
+        $this->info['IAlign']   = self::makeEnum('top,middle,bottom,left,right');
+        $this->info['LAlign']   = self::makeEnum('top,bottom,left,right');
+        $this->info['FrameTarget'] = new HTMLPurifier_AttrDef_HTML_FrameTarget();
 
         // unimplemented aliases
         $this->info['ContentType'] = new HTMLPurifier_AttrDef_Text();
@@ -44,6 +54,10 @@ public function __construct() {
         $this->info['Number']   = new HTMLPurifier_AttrDef_Integer(false, false, true);
     }
 
+    private static function makeEnum($in) {
+        return new HTMLPurifier_AttrDef_Clone(new HTMLPurifier_AttrDef_Enum(explode(',', $in)));
+    }
+
     /**
      * Retrieves a type
      * @param $type String type name

library/HTMLPurifier/HTMLModule/Forms.php

@@ -84,7 +84,8 @@ public function setup($config) {
         $button->excludes = $this->makeLookup(
             'form', 'fieldset', // Form
             'input', 'select', 'textarea', 'label', 'button', // Formctrl
-            'a' // as per HTML 4.01 spec, this is omitted by modularization
+            'a', // as per HTML 4.01 spec, this is omitted by modularization
+            'isindex', 'iframe' // legacy items
         );
 
         // Extra exclusion: img usemap="" is not permitted within this element.

library/HTMLPurifier/HTMLModule/Legacy.php

@@ -89,7 +89,7 @@ public function setup($config) {
         $hr->attr['width'] = 'Length';
 
         $img = $this->addBlankElement('img');
-        $img->attr['align'] = 'Enum#top,middle,bottom,left,right';
+        $img->attr['align'] = 'IAlign';
         $img->attr['border'] = 'Pixels';
         $img->attr['hspace'] = 'Pixels';
         $img->attr['vspace'] = 'Pixels';
@@ -136,6 +136,22 @@ public function setup($config) {
         $ul->attr['compact'] = 'Bool#compact';
         $ul->attr['type'] = 'Enum#square,disc,circle';
 
+        // "safe" modifications to "unsafe" elements
+        // WARNING: If you want to add support for an unsafe, legacy
+        // attribute, make a new TrustedLegacy module with the trusted
+        // bit set appropriately
+
+        $form = $this->addBlankElement('form');
+        $form->content_model = 'Flow | #PCDATA';
+        $form->content_model_type = 'optional';
+        $form->attr['target'] = 'FrameTarget';
+
+        $input = $this->addBlankElement('input');
+        $input->attr['align'] = 'IAlign';
+
+        $legend = $this->addBlankElement('legend');
+        $legend->attr['align'] = 'LAlign';
+
     }
 
 }

library/HTMLPurifier/HTMLModuleManager.php

@@ -369,6 +369,13 @@ public function getElement($name, $trusted = null) {
                 // :TODO:
                 // non-standalone definitions that don't have a standalone
                 // to merge into could be deferred to the end
+                // HOWEVER, it is perfectly valid for a non-standalone
+                // definition to lack a standalone definition, even
+                // after all processing: this allows us to safely
+                // specify extra attributes for elements that may not be
+                // enabled all in one place.  In particular, this might
+                // be the case for trusted elements.  WARNING: care must
+                // be taken that the /extra/ definitions are all safe.
                 continue;
             }
 

tests/HTMLPurifier/HTMLModule/FormsTest.php

@@ -7,10 +7,10 @@ function setUp() {
         parent::setUp();
         $this->config->set('HTML.Trusted', true);
         $this->config->set('Attr.EnableID', true);
-        $this->config->set('Cache.DefinitionImpl', null);
     }
 
     function testBasicUse() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
         $this->assertResult( // need support for label for later
             '
 <form action="http://somesite.com/prog/adduser" method="post">
@@ -30,6 +30,7 @@ function testBasicUse() {
     }
 
     function testSelectOption() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
         $this->assertResult('
 <form action="http://somesite.com/prog/component-select" method="post">
    <p>
@@ -49,6 +50,7 @@ function testSelectOption() {
     }
 
     function testSelectOptgroup() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
         $this->assertResult('
 <form action="http://somesite.com/prog/someprog" method="post">
 <p>
@@ -74,6 +76,7 @@ function testSelectOptgroup() {
     }
 
     function testTextarea() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
         $this->assertResult('
 <form action="http://somesite.com/prog/text-read" method="post">
    <p>
@@ -90,6 +93,7 @@ function testTextarea() {
     // label tests omitted
 
     function testFieldset() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
         $this->assertResult('
 <form action="..." method="post">
  <fieldset>
@@ -122,17 +126,30 @@ function testFieldset() {
     }
 
     function testInputTransform() {
+        $this->config->set('HTML.Doctype', 'XHTML 1.0 Strict');
         $this->assertResult('<input type="checkbox" />', '<input type="checkbox" value="" />');
     }
 
     function testTextareaTransform() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
         $this->assertResult('<textarea></textarea>', '<textarea cols="22" rows="3"></textarea>');
     }
 
     function testTextInFieldset() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
         $this->assertResult('<fieldset>   <legend></legend>foo</fieldset>');
     }
 
+    function testStrict() {
+        $this->config->set('HTML.Doctype', 'HTML 4.01 Strict');
+        $this->assertResult('<form action=""></form>', '');
+    }
+
+    function testLegacy() {
+        $this->assertResult('<form action=""></form>');
+        $this->assertResult('<form action=""><input align="left" /></form>');
+    }
+
 }
 
 // vim: et sw=4 sts=4