[1.7.0] Implement and hook-in Tidy module setup.

- CommonAttributes factored into XMLCommonAttributes and NonXMLCommonAttributes
- Tidy abstract module was completely refactored in interest of usability
- Add friendly error message if module does not have name

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1070 48356398-32a2-884e-a903-53898d9a118a

NEWS

@@ -12,6 +12,8 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 1.7.0, unknown release date
 # Completely refactored HTMLModuleManager, decentralizing safety
   information
+# Transform modules changed to Tidy modules, which offer more flexibility
+  and better modularization
 . Unit test for ElementDef created, ElementDef behavior modified to
   be more flexible
 . Added convenience functions for HTMLModule constructors

TODO

@@ -10,10 +10,9 @@ TODO List
 1.7 release [Advanced API]
  # Complete advanced API, and fully document it
     - Add framework for unsafe attributes
-    - Wire in modes (configuration, module and manager wise)
+    - Document Tidy modules
     - Reorganize configuration directives
-    - Determine handling for complex/cascading configuration directives
+    - Set up doctype object inside configuration object
-    - Reorganize transformation modules
     - Set up anonymous module management by HTMLDefinition
  # Implement HTMLDefinition caching using serialize
  # Implement all deprecated tags and attributes

library/HTMLPurifier/HTMLModule/CommonAttributes.php

@@ -1,5 +1,7 @@
 <?php
 
+require_once 'HTMLPurifier/HTMLModule.php';
+
 class HTMLPurifier_HTMLModule_CommonAttributes extends HTMLPurifier_HTMLModule
 {
     var $name = 'CommonAttributes';
@@ -12,9 +14,7 @@ class HTMLPurifier_HTMLModule_CommonAttributes extends HTMLPurifier_HTMLModule
             'id' => 'ID',
             'title' => 'CDATA',
         ),
-        'Lang' => array(
+        'Lang' => array(),
-            'xml:lang' => 'LanguageCode',
-        ),
         'I18N' => array(
             0 => array('Lang'), // proprietary, for xml:lang/lang
         ),

library/HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php

@@ -0,0 +1,16 @@
+<?php
+
+require_once 'HTMLPurifier/HTMLModule.php';
+
+class HTMLPurifier_HTMLModule_NonXMLCommonAttributes extends HTMLPurifier_HTMLModule
+{
+    var $name = 'NonXMLCommonAttributes';
+    
+    var $attr_collections = array(
+        'Lang' => array(
+            'lang' => 'LanguageCode',
+        )
+    );
+}
+
+?>

library/HTMLPurifier/HTMLModule/Tidy.php

@@ -25,13 +25,13 @@ HTMLPurifier_ConfigSchema::defineAllowedValues(
 );
 
 HTMLPurifier_ConfigSchema::define(
-    'HTML', 'TidyAdd', array(), 'list', '
+    'HTML', 'TidyAdd', array(), 'lookup', '
 Fixes to add to the default set of Tidy fixes as per your level. This
 directive has been available since 1.7.0.
 ' );
 
 HTMLPurifier_ConfigSchema::define(
-    'HTML', 'TidyRemove', array(), 'list', '
+    'HTML', 'TidyRemove', array(), 'lookup', '
 Fixes to remove from the default set of Tidy fixes as per your level. This
 directive has been available since 1.7.0.
 ' );
@@ -49,6 +49,11 @@ class HTMLPurifier_HTMLModule_Tidy extends HTMLPurifier_HTMLModule
      */
     var $levels = array(0 => 'none', 'light', 'medium', 'heavy');
     
+    /**
+     * Default level to place all fixes in. Disabled by default
+     */
+    var $defaultLevel = null;
+    
     /**
      * Lists of fixes used by getFixesForLevel(). Format is:
      *      HTMLModule_Tidy->fixesForLevel[$level] = array('fix-1', 'fix-2');
@@ -66,20 +71,32 @@ class HTMLPurifier_HTMLModule_Tidy extends HTMLPurifier_HTMLModule
      *       subtracted fix has no effect.
      */
     function construct($config) {
+        
+        // create fixes, initialize fixesForLevel
+        $fixes = $this->makeFixes();
+        $this->makeFixesForLevel($fixes);
+        
+        // figure out which fixes to use
         $level = $config->get('HTML', 'TidyLevel');
-        $fixes = $this->getFixesForLevel($level);
+        $fixes_lookup = $this->getFixesForLevel($level);
-        
-        $add_fixes = $config->get('HTML', 'TidyAdd');
-        foreach ($add_fixes as $fix) {
-            $fixes[$fix] = true;
-        }
         
+        // get custom fix declarations: these need namespace processing
+        $add_fixes    = $config->get('HTML', 'TidyAdd');
         $remove_fixes = $config->get('HTML', 'TidyRemove');
-        foreach ($remove_fixes as $fix) {
+        
-            unset($fixes[$fix]);
+        foreach ($fixes as $name => $fix) {
+            // needs to be refactored a little to implement globbing
+            if (
+                isset($remove_fixes[$name]) ||
+                (!isset($add_fixes[$name]) && !isset($fixes_lookup[$name]))
+            ) {
+                unset($fixes[$name]);
+            }
         }
         
+        // populate this module with necessary fixes
         $this->populate($fixes);
+        
     }
     
     /**
@@ -113,13 +130,111 @@ class HTMLPurifier_HTMLModule_Tidy extends HTMLPurifier_HTMLModule
         return $ret;
     }
     
+    /**
+     * Dynamically populates the $fixesForLevel member variable using
+     * the fixes array. It may be custom overloaded, used in conjunction
+     * with $defaultLevel, or not used at all.
+     */
+    function makeFixesForLevel($fixes) {
+        if (!isset($this->defaultLevel)) return;
+        if (!isset($this->fixesForLevel[$this->defaultLevel])) {
+            trigger_error(
+                'Default level ' . $this->defaultLevel . ' does not exist',
+                E_USER_ERROR
+            );
+            return;
+        }
+        $this->fixesForLevel[$this->defaultLevel] = array_keys($fixes);
+    }
+    
     /**
      * Populates the module with transforms and other special-case code
      * based on a list of fixes passed to it
-     * @abstract
      * @param $lookup Lookup table of fixes to activate
      */
-    function populate($lookup) {}
+    function populate($fixes) {
+        foreach ($fixes as $name => $fix) {
+            // determine what the fix is for
+            list($type, $params) = $this->getFixType($name);
+            switch ($type) {
+                case 'attr_transform_pre':
+                case 'attr_transform_post':
+                    $attr = $params['attr'];
+                    if (isset($params['element'])) {
+                        $element = $params['element'];
+                        if (empty($this->info[$element])) {
+                            $e =& $this->addBlankElement($element);
+                        } else {
+                            $e =& $this->info[$element];
+                        }
+                    } else {
+                        $type = "info_$type";
+                        $e =& $this;
+                    }
+                    $f =& $e->$type;
+                    $f[$attr] = $fix;
+                    break;
+                case 'tag_transform':
+                    $this->info_tag_transform[$params['element']] = $fix;
+                    break;
+                case 'child':
+                case 'content_model_type':
+                    $element = $params['element'];
+                    if (empty($this->info[$element])) {
+                        $e =& $this->addBlankElement($element);
+                    } else {
+                        $e =& $this->info[$element];
+                    }
+                    $e->$type = $fix;
+                    break;
+                default:
+                    trigger_error("Fix type $type not supported", E_USER_ERROR);
+                    break;
+            }
+        }
+    }
+    
+    /**
+     * Parses a fix name and determines what kind of fix it is, as well
+     * as other information defined by the fix
+     * @param $name String name of fix
+     * @return array(string $fix_type, array $fix_parameters)
+     * @note $fix_parameters is type dependant, see populate() for usage
+     *       of these parameters
+     */
+    function getFixType($name) {
+        // parse it
+        $property = $attr = null;
+        if (strpos($name, '#') !== false) list($name, $property) = explode('#', $name);
+        if (strpos($name, '@') !== false) list($name, $attr)     = explode('@', $name);
+        
+        // figure out the parameters
+        $params = array();
+        if ($name !== '')    $params['element'] = $name;
+        if (!is_null($attr)) $params['attr'] = $attr;
+        
+        // special case: attribute transform
+        if (!is_null($attr)) {
+            if (is_null($property)) $property = 'pre';
+            $type = 'attr_transform_' . $property;
+            return array($type, $params);
+        }
+        
+        // special case: tag transform
+        if (is_null($property)) {
+            return array('tag_transform', $params);
+        }
+        
+        return array($property, $params);
+        
+    }
+    
+    /**
+     * Defines all fixes the module will perform in a compact
+     * associative array of fix name to fix implementation.
+     * @abstract
+     */
+    function makeFixes() {}
     
 }
 

library/HTMLPurifier/HTMLModule/Tidy/XHTML.php

@@ -0,0 +1,20 @@
+<?php
+
+require_once 'HTMLPurifier/HTMLModule/Tidy.php';
+
+class HTMLPurifier_HTMLModule_Tidy_XHTML extends
+      HTMLPurifier_HTMLModule_Tidy
+{
+    
+    var $name = 'Tidy_XHTML';
+    var $defaultLevel = 'medium';
+    
+    function makeFixes() {
+        $r = array();
+        $r['@lang'] = new HTMLPurifier_AttrTransform_Lang();
+        return $r;
+    }
+    
+}
+
+?>

library/HTMLPurifier/HTMLModule/Tidy/XHTMLAndHTML4.php

@@ -0,0 +1,179 @@
+<?php
+
+require_once 'HTMLPurifier/HTMLModule/Tidy.php';
+
+class HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4 extends
+      HTMLPurifier_HTMLModule_Tidy
+{
+    
+    function makeFixes() {
+        
+        $r = array();
+        
+        // == deprecated tag transforms ===================================
+        
+        $r['font']   = new HTMLPurifier_TagTransform_Font();
+        $r['menu']   = new HTMLPurifier_TagTransform_Simple('ul');
+        $r['dir']    = new HTMLPurifier_TagTransform_Simple('ul');
+        $r['center'] = new HTMLPurifier_TagTransform_Center();
+        
+        // == deprecated attribute transforms =============================
+        
+        $r['caption@align'] = 
+            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
+                // we're following IE's behavior, not Firefox's, due
+                // to the fact that no one supports caption-side:right,
+                // W3C included (with CSS 2.1). This is a slightly
+                // unreasonable attribute!
+                'left'   => 'text-align:left;',
+                'right'  => 'text-align:right;',
+                'top'    => 'caption-side:top;',
+                'bottom' => 'caption-side:bottom;' // not supported by IE
+            ));
+        
+        // @align for img -------------------------------------------------
+        $r['img@align'] =
+            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
+                'left'   => 'float:left;',
+                'right'  => 'float:right;',
+                'top'    => 'vertical-align:top;',
+                'middle' => 'vertical-align:middle;',
+                'bottom' => 'vertical-align:baseline;',
+            ));
+        
+        // @align for table -----------------------------------------------
+        $r['table@align'] =
+            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
+                'left'   => 'float:left;',
+                'center' => 'margin-left:auto;margin-right:auto;',
+                'right'  => 'float:right;'
+            ));
+        
+        // @align for hr -----------------------------------------------
+        $r['hr@align'] =
+            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
+                // we use both text-align and margin because these work
+                // for different browsers (IE and Firefox, respectively)
+                // and the melange makes for a pretty cross-compatible
+                // solution
+                'left'   => 'margin-left:0;margin-right:auto;text-align:left;',
+                'center' => 'margin-left:auto;margin-right:auto;text-align:center;',
+                'right'  => 'margin-left:auto;margin-right:0;text-align:right;'
+            ));
+        
+        // @align for h1, h2, h3, h4, h5, h6, p, div ----------------------
+        // {{{
+            $align_lookup = array();
+            $align_values = array('left', 'right', 'center', 'justify');
+            foreach ($align_values as $v) $align_lookup[$v] = "text-align:$v;";
+        // }}}
+        $r['h1@align'] =
+        $r['h2@align'] =
+        $r['h3@align'] =
+        $r['h4@align'] =
+        $r['h5@align'] =
+        $r['h6@align'] =
+        $r['p@align']  =
+        $r['div@align'] =
+            new HTMLPurifier_AttrTransform_EnumToCSS('align', $align_lookup);
+        
+        // @bgcolor for table, tr, td, th ---------------------------------
+        $r['table@bgcolor'] =
+        $r['td@bgcolor'] =
+        $r['th@bgcolor'] =
+            new HTMLPurifier_AttrTransform_BgColor();
+        
+        // @border for img ------------------------------------------------
+        $r['img@border'] = new HTMLPurifier_AttrTransform_Border();
+        
+        // @clear for br --------------------------------------------------
+        $r['br@clear'] =
+            new HTMLPurifier_AttrTransform_EnumToCSS('clear', array(
+                'left'  => 'clear:left;',
+                'right' => 'clear:right;',
+                'all'   => 'clear:both;',
+                'none'  => 'clear:none;',
+            ));
+        
+        // @height for td, th ---------------------------------------------
+        $r['td@height'] = 
+        $r['th@height'] =
+            new HTMLPurifier_AttrTransform_Length('height');
+        
+        // @hspace for img ------------------------------------------------
+        $r['img@hspace'] = new HTMLPurifier_AttrTransform_ImgSpace('hspace');
+        
+        // @name for img, a -----------------------------------------------
+        $r['img@name'] = 
+        $r['a@name'] = new HTMLPurifier_AttrTransform_Name();
+        
+        // @noshade for hr ------------------------------------------------
+        // this transformation is not precise but often good enough.
+        // different browsers use different styles to designate noshade
+        $r['hr@noshade'] =
+            new HTMLPurifier_AttrTransform_BoolToCSS(
+                'noshade',
+                'color:#808080;background-color:#808080;border:0;'
+            );
+        
+        // @nowrap for td, th ---------------------------------------------
+        $r['td@nowrap'] = 
+        $r['th@nowrap'] =
+            new HTMLPurifier_AttrTransform_BoolToCSS(
+                'nowrap',
+                'white-space:nowrap;'
+            );
+        
+        // @size for hr  --------------------------------------------------
+        $r['hr@size'] = new HTMLPurifier_AttrTransform_Length('size', 'height');
+        
+        // @type for li, ol, ul -------------------------------------------
+        // {{{
+            $ul_types = array(
+                'disc'   => 'list-style-type:disc;',
+                'square' => 'list-style-type:square;',
+                'circle' => 'list-style-type:circle;'
+            );
+            $ol_types = array(
+                '1'   => 'list-style-type:decimal;',
+                'i'   => 'list-style-type:lower-roman;',
+                'I'   => 'list-style-type:upper-roman;',
+                'a'   => 'list-style-type:lower-alpha;',
+                'A'   => 'list-style-type:upper-alpha;'
+            );
+            $li_types = $ul_types + $ol_types;
+        // }}}
+        
+        $r['ul@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ul_types);
+        $r['ol@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ol_types, true);
+        $r['li@type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $li_types, true);
+        
+        // @vspace for img ------------------------------------------------
+        $r['img@vspace'] = new HTMLPurifier_AttrTransform_ImgSpace('vspace');
+        
+        // @width for hr, td, th ------------------------------------------
+        $r['td@width'] =
+        $r['th@width'] = 
+        $r['hr@width'] = new HTMLPurifier_AttrTransform_Length('width');
+        
+        return $r;
+        
+    }
+    
+}
+
+class HTMLPurifier_HTMLModule_Tidy_Transitional extends
+      HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4
+{
+    var $name = 'Tidy_Transitional';
+    var $defaultLevel = 'light'; // switch this to heavy once we implement legacy fully
+}
+
+class HTMLPurifier_HTMLModule_Tidy_Strict extends
+      HTMLPurifier_HTMLModule_Tidy_XHTMLAndHTML4
+{
+    var $name = 'Tidy_Strict';
+    var $defaultLevel = 'light';
+}
+
+?>

library/HTMLPurifier/HTMLModule/Tidy/XHTMLStrict.php

@@ -0,0 +1,27 @@
+<?php
+
+require_once 'HTMLPurifier/HTMLModule/Tidy.php';
+
+class HTMLPurifier_HTMLModule_Tidy_XHTMLStrict extends
+      HTMLPurifier_HTMLModule_Tidy
+{
+    
+    var $name = 'Tidy_XHTMLStrict';
+    var $defaultLevel = 'light';
+    
+    function makeFixes() {
+        $r = array();
+        $r['blockquote#child'] = false;
+        $r['blockquote#content_model_type'] = 'strictblockquote';
+        return $r;
+    }
+    
+    var $defines_child_def = true;
+    function getChildDef($def) {
+        if ($def->content_model_type != 'strictblockquote') return false;
+        return new HTMLPurifier_ChildDef_StrictBlockquote($def->content_model);
+    }
+    
+}
+
+?>

library/HTMLPurifier/HTMLModule/TransformToStrict.php

@@ -1,216 +0,0 @@
-<?php
-
-require_once 'HTMLPurifier/ChildDef/StrictBlockquote.php';
-
-require_once 'HTMLPurifier/TagTransform/Simple.php';
-require_once 'HTMLPurifier/TagTransform/Center.php';
-require_once 'HTMLPurifier/TagTransform/Font.php';
-
-require_once 'HTMLPurifier/AttrTransform/Lang.php';
-require_once 'HTMLPurifier/AttrTransform/BgColor.php';
-require_once 'HTMLPurifier/AttrTransform/BoolToCSS.php';
-require_once 'HTMLPurifier/AttrTransform/Border.php';
-require_once 'HTMLPurifier/AttrTransform/Name.php';
-require_once 'HTMLPurifier/AttrTransform/Length.php';
-require_once 'HTMLPurifier/AttrTransform/ImgSpace.php';
-require_once 'HTMLPurifier/AttrTransform/EnumToCSS.php';
-
-/**
- * Proprietary module that transforms deprecated elements into Strict
- * HTML (see HTML 4.01 and XHTML 1.0) when possible.
- */
-
-class HTMLPurifier_HTMLModule_TransformToStrict extends HTMLPurifier_HTMLModule
-{
-    
-    var $name = 'TransformToStrict';
-    
-    var $attr_collections = array(
-        'Lang' => array()
-    );
-    
-    function HTMLPurifier_HTMLModule_TransformToStrict() {
-        
-        // behavior with transformations when there's another CSS property
-        // working on it is interesting: the CSS will *always* override
-        // the deprecated attribute, whereas an inline CSS declaration will
-        // override the corresponding declaration in, say, an external
-        // stylesheet. This behavior won't affect most people, but it
-        // does represent an operational difference we CANNOT fix.
-        
-        // == deprecated tag transforms ===================================
-        
-        $this->info_tag_transform['font']   = new HTMLPurifier_TagTransform_Font();
-        $this->info_tag_transform['menu']   = new HTMLPurifier_TagTransform_Simple('ul');
-        $this->info_tag_transform['dir']    = new HTMLPurifier_TagTransform_Simple('ul');
-        $this->info_tag_transform['center'] = new HTMLPurifier_TagTransform_Center();
-        
-        // == miscellaneous transforms ====================================
-        
-        // initialize blank element definitions
-        $elements = array( // alphabetically sorted
-            'a', 'blockquote', 'br', 'caption', 'h1', 'h2', 'h3', 'h4', 
-            'h5', 'h6', 'hr', 'img', 'li', 'ol', 'p',  'table', 'td',
-            'th', 'tr', 'ul' );
-        foreach ($elements as $name) $this->addBlankElement($name);
-        
-        // this should not be applied to XHTML 1.0 Transitional, ONLY
-        // XHTML 1.0 Strict. We may need to put this in another class.
-        $this->info['blockquote']->content_model_type = 'strictblockquote';
-        $this->info['blockquote']->child = false;
-        
-        // == deprecated attribute transforms =============================
-        
-        // this segment will probably need to modularized in some fashion
-        // in order to allow for different "levels" of transformation
-        
-        // @align for caption ---------------------------------------------
-        $this->info['caption']->attr_transform_pre['align'] =
-            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
-                // we're following IE's behavior, not Firefox's, due
-                // to the fact that no one supports caption-side:right,
-                // W3C included (with CSS 2.1). This is a slightly
-                // unreasonable attribute!
-                'left'   => 'text-align:left;',
-                'right'  => 'text-align:right;',
-                'top'    => 'caption-side:top;',
-                'bottom' => 'caption-side:bottom;' // not supported by IE
-            ));
-        
-        // @align for img -------------------------------------------------
-        $this->info['img']->attr_transform_pre['align'] = 
-            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
-                'left'   => 'float:left;',
-                'right'  => 'float:right;',
-                'top'    => 'vertical-align:top;',
-                'middle' => 'vertical-align:middle;',
-                'bottom' => 'vertical-align:baseline;',
-            ));
-        
-        // @align for table -----------------------------------------------
-        $this->info['table']->attr_transform_pre['align'] = 
-            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
-                'left'   => 'float:left;',
-                'center' => 'margin-left:auto;margin-right:auto;',
-                'right'  => 'float:right;'
-            ));
-        
-        // @align for hr -----------------------------------------------
-        $this->info['hr']->attr_transform_pre['align'] = 
-            new HTMLPurifier_AttrTransform_EnumToCSS('align', array(
-                // we use both text-align and margin because these work
-                // for different browsers (IE and Firefox, respectively)
-                // and the melange makes for a pretty cross-compatible
-                // solution
-                'left'   => 'margin-left:0;margin-right:auto;text-align:left;',
-                'center' => 'margin-left:auto;margin-right:auto;text-align:center;',
-                'right'  => 'margin-left:auto;margin-right:0;text-align:right;'
-            ));
-        
-        // @align for h1, h2, h3, h4, h5, h6, p ---------------------------
-        $align_lookup = array();
-        $align_values = array('left', 'right', 'center', 'justify');
-        foreach ($align_values as $v) $align_lookup[$v] = "text-align:$v;";
-        $this->info['h1']->attr_transform_pre['align'] = 
-        $this->info['h2']->attr_transform_pre['align'] = 
-        $this->info['h3']->attr_transform_pre['align'] = 
-        $this->info['h4']->attr_transform_pre['align'] = 
-        $this->info['h5']->attr_transform_pre['align'] = 
-        $this->info['h6']->attr_transform_pre['align'] = 
-        $this->info['p'] ->attr_transform_pre['align'] = 
-          new HTMLPurifier_AttrTransform_EnumToCSS('align', $align_lookup);
-        
-        // @bgcolor for table, tr, td, th ---------------------------------
-        $this->info['table']->attr_transform_pre['bgcolor'] = 
-        $this->info['tr']->attr_transform_pre['bgcolor'] = 
-        $this->info['td']->attr_transform_pre['bgcolor'] = 
-        $this->info['th']->attr_transform_pre['bgcolor'] =
-          new HTMLPurifier_AttrTransform_BgColor();
-        
-        // @border for img ------------------------------------------------
-        $this->info['img']->attr_transform_pre['border'] = new HTMLPurifier_AttrTransform_Border();
-        
-        // @clear for br --------------------------------------------------
-        $this->info['br']->attr_transform_pre['clear'] =
-            new HTMLPurifier_AttrTransform_EnumToCSS('clear', array(
-                'left'  => 'clear:left;',
-                'right' => 'clear:right;',
-                'all'   => 'clear:both;',
-                'none'  => 'clear:none;',
-            ));
-        
-        // @height for td, th ---------------------------------------------
-        $this->info['td']->attr_transform_pre['height'] = 
-        $this->info['th']->attr_transform_pre['height'] =
-          new HTMLPurifier_AttrTransform_Length('height');
-        
-        // @hspace for img ------------------------------------------------
-        $this->info['img']->attr_transform_pre['hspace'] = new HTMLPurifier_AttrTransform_ImgSpace('hspace');
-        
-        // @lang ----------------------------------------------------------
-        // this is overridden in TransformToXHTML11
-        $this->info_attr_transform_post['lang'] = new HTMLPurifier_AttrTransform_Lang();
-        $this->attr_collections['Lang']['lang'] = new HTMLPurifier_AttrDef_Lang();
-        
-        // @name for img, a -----------------------------------------------
-        $this->info['img']->attr_transform_pre['name'] = 
-        $this->info['a']->attr_transform_pre['name'] = new HTMLPurifier_AttrTransform_Name();
-        
-        // @noshade for hr ------------------------------------------------
-        // this transformation is not precise but often good enough.
-        // different browsers use different styles to designate noshade
-        $this->info['hr']->attr_transform_pre['noshade'] =
-            new HTMLPurifier_AttrTransform_BoolToCSS(
-                'noshade',
-                'color:#808080;background-color:#808080;border:0;'
-            );
-        
-        // @nowrap for td, th ---------------------------------------------
-        $this->info['td']->attr_transform_pre['nowrap'] = 
-        $this->info['th']->attr_transform_pre['nowrap'] =
-            new HTMLPurifier_AttrTransform_BoolToCSS(
-                'nowrap',
-                'white-space:nowrap;'
-            );
-        
-        // @size for hr  --------------------------------------------------
-        $this->info['hr']->attr_transform_pre['size'] = new HTMLPurifier_AttrTransform_Length('size', 'height');
-        
-        // @type for li, ol, ul -------------------------------------------
-        $ul_types = array(
-            'disc'   => 'list-style-type:disc;',
-            'square' => 'list-style-type:square;',
-            'circle' => 'list-style-type:circle;'
-        );
-        $ol_types = array(
-            '1'   => 'list-style-type:decimal;',
-            'i'   => 'list-style-type:lower-roman;',
-            'I'   => 'list-style-type:upper-roman;',
-            'a'   => 'list-style-type:lower-alpha;',
-            'A'   => 'list-style-type:upper-alpha;'
-        );
-        $li_types = $ul_types + $ol_types;
-        
-        $this->info['ul']->attr_transform_pre['type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ul_types);
-        $this->info['ol']->attr_transform_pre['type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $ol_types, true);
-        $this->info['li']->attr_transform_pre['type'] = new HTMLPurifier_AttrTransform_EnumToCSS('type', $li_types, true);
-        
-        // @vspace for img ------------------------------------------------
-        $this->info['img']->attr_transform_pre['vspace'] = new HTMLPurifier_AttrTransform_ImgSpace('vspace');
-        
-        // @width for hr, td, th ------------------------------------------
-        $this->info['td']->attr_transform_pre['width'] = 
-        $this->info['th']->attr_transform_pre['width'] = 
-        $this->info['hr']->attr_transform_pre['width'] = new HTMLPurifier_AttrTransform_Length('width');
-        
-    }
-    
-    var $defines_child_def = true;
-    function getChildDef($def) {
-        if ($def->content_model_type != 'strictblockquote') return false;
-        return new HTMLPurifier_ChildDef_StrictBlockquote($def->content_model);
-    }
-    
-}
-
-?>

library/HTMLPurifier/HTMLModule/TransformToXHTML11.php

@@ -1,36 +0,0 @@
-<?php
-
-require_once 'HTMLPurifier/AttrTransform/Lang.php';
-
-/**
- * Proprietary module that transforms XHTML 1.0 deprecated aspects into
- * XHTML 1.1 compliant ones, when possible. For maximum effectiveness,
- * HTMLPurifier_HTMLModule_TransformToStrict must also be loaded
- * (otherwise, elements that were deprecated from Transitional to Strict
- * will not be transformed).
- * 
- * XHTML 1.1 compliant document are automatically XHTML 1.0 compliant too,
- * although they may not be as friendly to legacy browsers.
- */
-
-class HTMLPurifier_HTMLModule_TransformToXHTML11 extends HTMLPurifier_HTMLModule
-{
-    
-    var $name = 'TransformToXHTML11';
-    var $attr_collections = array(
-        'Lang' => array(
-            'lang' => false // remove it
-        )
-    );
-    
-    var $info_attr_transform_post = array(
-        'lang' => false // remove it
-    );
-    
-    function HTMLPurifier_HTMLModule_TransformToXHTML11() {
-        $this->info_attr_transform_pre['lang'] = new HTMLPurifier_AttrTransform_Lang();
-    }
-    
-}
-
-?>

library/HTMLPurifier/HTMLModule/XMLCommonAttributes.php

@@ -0,0 +1,16 @@
+<?php
+
+require_once 'HTMLPurifier/HTMLModule.php';
+
+class HTMLPurifier_HTMLModule_XMLCommonAttributes extends HTMLPurifier_HTMLModule
+{
+    var $name = 'XMLCommonAttributes';
+    
+    var $attr_collections = array(
+        'Lang' => array(
+            'xml:lang' => 'LanguageCode',
+        )
+    );
+}
+
+?>

library/HTMLPurifier/HTMLModuleManager.php

@@ -26,10 +26,14 @@ require_once 'HTMLPurifier/HTMLModule/StyleAttribute.php';
 require_once 'HTMLPurifier/HTMLModule/Legacy.php';
 require_once 'HTMLPurifier/HTMLModule/Target.php';
 require_once 'HTMLPurifier/HTMLModule/Scripting.php';
+require_once 'HTMLPurifier/HTMLModule/XMLCommonAttributes.php';
+require_once 'HTMLPurifier/HTMLModule/NonXMLCommonAttributes.php';
 
-// proprietary modules
+// tidy modules
-require_once 'HTMLPurifier/HTMLModule/TransformToStrict.php';
+require_once 'HTMLPurifier/HTMLModule/Tidy.php';
-require_once 'HTMLPurifier/HTMLModule/TransformToXHTML11.php';
+require_once 'HTMLPurifier/HTMLModule/Tidy/XHTMLAndHTML4.php';
+require_once 'HTMLPurifier/HTMLModule/Tidy/XHTML.php';
+require_once 'HTMLPurifier/HTMLModule/Tidy/XHTMLStrict.php';
 
 HTMLPurifier_ConfigSchema::define(
     'HTML', 'Doctype', null, 'string/null',
@@ -113,40 +117,37 @@ class HTMLPurifier_HTMLModuleManager
             'StyleAttribute', 'Scripting'
         );
         $transitional = array('Legacy', 'Target');
+        $xml = array('XMLCommonAttributes');
+        $non_xml = array('NonXMLCommonAttributes');
         
         $this->doctypes->register(
             'HTML 4.01 Transitional', false,
-            array_merge($common, $transitional),
+            array_merge($common, $transitional, $non_xml),
-            array('TransformToStrict')
+            array('Tidy_Transitional')
-            // Tidy: Transitional
         );
         
         $this->doctypes->register(
             'HTML 4.01 Strict', false,
-            array_merge($common),
+            array_merge($common, $non_xml),
-            array('TransformToStrict')
+            array('Tidy_Strict')
-            // Tidy: Strict
         );
         
         $this->doctypes->register(
             'XHTML 1.0 Transitional', true,
-            array_merge($common, $transitional),
+            array_merge($common, $transitional, $xml, $non_xml),
-            array('TransformToStrict')
+            array('Tidy_Transitional', 'Tidy_XHTML')
-            // Tidy: Transitional, XHTML
         );
         
         $this->doctypes->register(
             'XHTML 1.0 Strict', true,
-            array_merge($common),
+            array_merge($common, $xml, $non_xml),
-            array('TransformToStrict')
+            array('Tidy_Strict', 'Tidy_XHTML', 'Tidy_XHTMLStrict')
-            // Tidy: Strict, XHTML
         );
         
         $this->doctypes->register(
             'XHTML 1.1', true,
-            array_merge($common),
+            array_merge($common, $xml),
-            array('TransformToStrict', 'TransformToXHTML11')
+            array('Tidy_Strict', 'Tidy_XHTML') // Tidy_XHTML1_1
-            // Tidy: Strict, XHTML1_1
         );
         
     }
@@ -194,6 +195,10 @@ class HTMLPurifier_HTMLModuleManager
             }
             $module = new $module();
         }
+        if (empty($module->name)) {
+            trigger_error('Module instance of ' . get_class($module) . ' must have name');
+            return;
+        }
         $this->registeredModules[$module->name] = $module;
     }
     
@@ -257,7 +262,9 @@ class HTMLPurifier_HTMLModuleManager
         
         foreach ($doctype->tidyModules as $module) {
             $this->processModule($module);
-            // FIXME!!! initialize the tidy modules here
+            if (method_exists($this->modules[$module], 'construct')) {
+                $this->modules[$module]->construct($config);
+            }
         }
         
         // setup lookup table based on all valid modules

tests/HTMLPurifier/HTMLModule/TidyTest.php

@@ -5,7 +5,7 @@ require_once 'HTMLPurifier/HTMLModule/Tidy.php';
 Mock::generatePartial(
     'HTMLPurifier_HTMLModule_Tidy',
     'HTMLPurifier_HTMLModule_Tidy_TestForConstruct',
-    array('populate')
+    array('makeFixes', 'makeFixesForLevel', 'populate')
 );
 
 class HTMLPurifier_HTMLModule_TidyTest extends UnitTestCase
@@ -49,7 +49,17 @@ class HTMLPurifier_HTMLModule_TidyTest extends UnitTestCase
         $module->fixesForLevel['light']  = array('light-fix-1', 'light-fix-2');
         $module->fixesForLevel['medium'] = array('medium-fix-1', 'medium-fix-2');
         $module->fixesForLevel['heavy']  = array('heavy-fix-1', 'heavy-fix-2');
-        // $module->HTMLPurifier_HTMLModule_Tidy(); // constructor
+        
+        $j = 0;
+        $fixes = array(
+            'light-fix-1'  => $lf1 = $j++,
+            'light-fix-2'  => $lf2 = $j++,
+            'medium-fix-1' => $mf1 = $j++,
+            'medium-fix-2' => $mf2 = $j++,
+            'heavy-fix-1'  => $hf1 = $j++,
+            'heavy-fix-2'  => $hf2 = $j++
+        );
+        $module->setReturnValue('makeFixes', $fixes);
         
         $config = HTMLPurifier_Config::create(array(
             'HTML.TidyLevel' => 'none'
@@ -62,13 +72,23 @@ class HTMLPurifier_HTMLModule_TidyTest extends UnitTestCase
         $config = HTMLPurifier_Config::create(array(
             'HTML.TidyLevel' => 'light'
         ));
-        $module->expectAt($i++, 'populate', array($module->getFixesForLevel('light')));
+        $module->expectAt($i++, 'populate', array(array(
+            'light-fix-1' => $lf1,
+            'light-fix-2' => $lf2
+        )));
         $module->construct($config);
         
         $config = HTMLPurifier_Config::create(array(
             'HTML.TidyLevel' => 'heavy'
         ));
-        $module->expectAt($i++, 'populate', array($module->getFixesForLevel('heavy')));
+        $module->expectAt($i++, 'populate', array(array(
+            'light-fix-1'  => $lf1,
+            'light-fix-2'  => $lf2,
+            'medium-fix-1' => $mf1,
+            'medium-fix-2' => $mf2,
+            'heavy-fix-1'  => $hf1,
+            'heavy-fix-2'  => $hf2
+        )));
         $module->construct($config);
         
         // fine grained tuning
@@ -78,8 +98,8 @@ class HTMLPurifier_HTMLModule_TidyTest extends UnitTestCase
             'HTML.TidyAdd'   => array('light-fix-1', 'medium-fix-1')
         ));
         $module->expectAt($i++, 'populate', array(array(
-            'light-fix-1' => true,
+            'light-fix-1' => $lf1,
-            'medium-fix-1' => true
+            'medium-fix-1' => $mf1
         )));
         $module->construct($config);
         
@@ -88,8 +108,8 @@ class HTMLPurifier_HTMLModule_TidyTest extends UnitTestCase
             'HTML.TidyRemove'   => array('light-fix-1', 'medium-fix-1')
         ));
         $module->expectAt($i++, 'populate', array(array(
-            'light-fix-2' => true,
+            'light-fix-2' => $lf2,
-            'medium-fix-2' => true
+            'medium-fix-2' => $mf2
         )));
         $module->construct($config);
         
@@ -99,6 +119,112 @@ class HTMLPurifier_HTMLModule_TidyTest extends UnitTestCase
         
     }
     
+    function test_makeFixesForLevel() {
+        
+        $module = new HTMLPurifier_HTMLModule_Tidy();
+        $module->defaultLevel = 'heavy';
+        
+        $module->makeFixesForLevel(array(
+            'fix-1' => 0,
+            'fix-2' => 1,
+            'fix-3' => 2
+        ));
+        
+        $this->assertIdentical($module->fixesForLevel['heavy'], array('fix-1', 'fix-2', 'fix-3'));
+        $this->assertIdentical($module->fixesForLevel['medium'], array());
+        $this->assertIdentical($module->fixesForLevel['light'], array());
+        
+    }
+    function test_makeFixesForLevel_undefinedLevel() {
+        
+        $module = new HTMLPurifier_HTMLModule_Tidy();
+        $module->defaultLevel = 'bananas';
+        
+        $this->expectError('Default level bananas does not exist');
+        
+        $module->makeFixesForLevel(array(
+            'fix-1' => 0
+        ));
+        
+    }
+    
+    function test_getFixType() {
+        
+        // syntax needs documenting
+        
+        $module = new HTMLPurifier_HTMLModule_Tidy();
+        
+        $this->assertIdentical(
+            $module->getFixType('a'),
+            array('tag_transform', array('element' => 'a'))
+        );
+        
+        $this->assertIdentical(
+            $module->getFixType('a@href'),
+            $reuse = array('attr_transform_pre', array('element' => 'a', 'attr' => 'href'))
+        );
+        
+        $this->assertIdentical(
+            $module->getFixType('a@href#pre'),
+            $reuse
+        );
+        
+        $this->assertIdentical(
+            $module->getFixType('a@href#post'),
+            array('attr_transform_post', array('element' => 'a', 'attr' => 'href'))
+        );
+        
+        $this->assertIdentical(
+            $module->getFixType('xml:foo@xml:bar'),
+            array('attr_transform_pre', array('element' => 'xml:foo', 'attr' => 'xml:bar'))
+        );
+        
+        $this->assertIdentical(
+            $module->getFixType('blockquote#child'),
+            array('child', array('element' => 'blockquote'))
+        );
+        
+        $this->assertIdentical(
+            $module->getFixType('@lang'),
+            array('attr_transform_pre', array('attr' => 'lang'))
+        );
+        
+        $this->assertIdentical(
+            $module->getFixType('@lang#post'),
+            array('attr_transform_post', array('attr' => 'lang'))
+        );
+        
+    }
+    
+    function test_populate() {
+        
+        $i = 0;
+        
+        $module = new HTMLPurifier_HTMLModule_Tidy();
+        $module->populate(array(
+            'element' => $element = $i++,
+            'element@attr' => $attr = $i++,
+            'element@attr#post' => $attr_post = $i++,
+            'element#child' => $child = $i++,
+            'element#content_model_type' => $content_model_type = $i++,
+            '@attr' => $global_attr = $i++,
+            '@attr#post' => $global_attr_post = $i++
+        ));
+        
+        $module2 = new HTMLPurifier_HTMLModule_Tidy();
+        $e =& $module2->addBlankElement('element');
+        $e->attr_transform_pre['attr'] = $attr;
+        $e->attr_transform_post['attr'] = $attr_post;
+        $e->child = $child;
+        $e->content_model_type = $content_model_type;
+        $module2->info_tag_transform['element'] = $element;
+        $module2->info_attr_transform_pre['attr'] = $global_attr;
+        $module2->info_attr_transform_post['attr'] = $global_attr_post;
+        
+        $this->assertEqual($module, $module2);
+        
+    }
+    
 }
 
 ?>