[1.7.0] Implement HTMLDefinition cache (very hacked together, but long unit test times were driving me crazy!)

- Add extra protection in AttrDef_URI against phantom Schemes
- Doctype moved from config to HTMLDefinition
- AttrDef_URITest mocks have more generic object parameters to deal with PHP4's copy-happy behavior

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1089 48356398-32a2-884e-a903-53898d9a118a

NEWS

@@ -27,6 +27,9 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 . AttrTypes now has accessor functions that should be used instead
   of directly manipulating info
 . TagTransform_Center deprecated in favor of generic TagTransform_Simple
+. Add extra protection in AttrDef_URI against phantom Schemes
+. Doctype object added to HTMLDefinition which describes certain aspects
+  of the operational document type
 
 1.6.1, released 2007-05-05
 ! Support for more deprecated attributes via transformations:

TODO

@@ -11,14 +11,15 @@ TODO List
  # Complete advanced API, and fully document it
     - Add framework for unsafe attributes
     - Reorganize configuration directives
-    - Set up doctype object inside configuration object
+    - Set up anonymous module management by HTMLDefinition (Advanced API)
-    - Set up anonymous module management by HTMLDefinition
+    - Get all AttrTypes into string form
- # Implement HTMLDefinition caching using serialize
+ # Clean up HTMLDefinition caching, need easy cache invalidation,
+   versioning of caches, etc.
  # Implement all deprecated tags and attributes
  # Create parsing/standards compliance smoketest
  # Reorganize Unit Tests
- - Parse TinyMCE-style whitelist into our %HTML.Allow* whitelists (possibly
+    - Refactor loop tests (esp. AttrDef_URI)
-   do this earlier)
+ - Parse TinyMCE-style whitelist into our %HTML.Allow* whitelists
  ? HTML interface for tweaking configuration to see changes
 
 1.8 release [Refactor, refactor!]

library/HTMLPurifier/AttrDef/URI.php

@@ -158,6 +158,14 @@ class HTMLPurifier_AttrDef_URI extends HTMLPurifier_AttrDef
             );
         }
         
+        // something funky weird happened in the registry, abort!
+        if (!$scheme_obj) {
+            trigger_error(
+                'Default scheme object "' . $config->get('URI', 'DefaultScheme') . '" was not readable',
+                E_USER_WARNING
+            );
+            return false;
+        }
         
         // the URI we're processing embeds_resource a resource in the page, but the URI
         // it references cannot be located

library/HTMLPurifier/Config.php

@@ -53,11 +53,6 @@ class HTMLPurifier_Config
      */
     var $autoFinalize = true;
     
-    /**
-     * Instance of HTMLPurifier_Doctype, representing current doctype
-     */
-    var $doctype;
-    
     /**
      * @param $definition HTMLPurifier_ConfigSchema that defines what directives
      *                    are allowed.
@@ -201,10 +196,17 @@ class HTMLPurifier_Config
             empty($this->html_definition) || // hasn't ever been setup
             ($raw && $this->html_definition->setup) // requesting new one
         ) {
+            if (!$raw) {
+                $this->html_definition = HTMLPurifier_HTMLDefinition::getCache($this);
+                if ($this->html_definition) return $this->html_definition;
+            }
             $this->html_definition = new HTMLPurifier_HTMLDefinition($this);
             if ($raw) return $this->html_definition; // no setup!
         }
-        if (!$this->html_definition->setup) $this->html_definition->setup();
+        if (!$this->html_definition->setup) {
+            $this->html_definition->setup();
+            $this->html_definition->saveCache($this);
+        }
         return $this->html_definition;
     }
     
@@ -243,16 +245,6 @@ class HTMLPurifier_Config
         }
     }
     
-    /**
-     * Returns the current doctype object
-     */
-    function getDoctype() {
-        if (!$this->doctype) {
-            $this->getHTMLDefinition();
-        }
-        return $this->doctype;
-    }
-    
     /**
      * Loads configuration values from an ini file
      * @param $filename Name of ini file

library/HTMLPurifier/DoctypeRegistry.php

@@ -77,7 +77,8 @@ class HTMLPurifier_DoctypeRegistry
         if (isset($this->aliases[$doctype])) $doctype = $this->aliases[$doctype];
         if (!isset($this->doctypes[$doctype])) {
             trigger_error('Doctype ' . htmlspecialchars($doctype) . ' does not exist');
-            $null = null; return $null;
+            $anon = new HTMLPurifier_Doctype($doctype);
+            return $anon;
         }
         return $this->doctypes[$doctype];
     }
@@ -93,7 +94,6 @@ class HTMLPurifier_DoctypeRegistry
     function make($config) {
         $original_doctype = $this->get($this->getDoctypeFromConfig($config));
         $doctype = $original_doctype->copy();
-        $config->doctype = $doctype;
         return $doctype;
     }
     

library/HTMLPurifier/Generator.php

@@ -71,10 +71,8 @@ class HTMLPurifier_Generator
         if (!$config) $config = HTMLPurifier_Config::createDefault();
         $this->_scriptFix   = $config->get('Output', 'CommentScriptContents');
         
-        $doctype = $config->getDoctype();
-        $this->_xhtml = $doctype->xml;
-        
         $this->_def = $config->getHTMLDefinition();
+        $this->_xhtml = $this->_def->doctype->xml;
         
         if (!$tokens) return '';
         for ($i = 0, $size = count($tokens); $i < $size; $i++) {

library/HTMLPurifier/HTMLDefinition.php

@@ -147,6 +147,10 @@ class HTMLPurifier_HTMLDefinition
      */
     var $info_content_sets = array();
     
+    /**
+     * Doctype object
+     */
+    var $doctype;
     
     
     /** PUBLIC BUT INTERNAL VARIABLES */
@@ -160,11 +164,49 @@ class HTMLPurifier_HTMLDefinition
      * Performs low-cost, preliminary initialization.
      * @param $config Instance of HTMLPurifier_Config
      */
-    function HTMLPurifier_HTMLDefinition(&$config) {
+    function HTMLPurifier_HTMLDefinition($config) {
-        $this->config =& $config;
+        $this->config = $config;
         $this->manager = new HTMLPurifier_HTMLModuleManager();
     }
     
+    /**
+     * Retrieve definition object from cache
+     */
+    function getCache($config) {
+        static $cache = array();
+        $file = HTMLPurifier_HTMLDefinition::getCacheFile($config);
+        if (isset($cache[$file])) return $cache[$file]; // unit test optimization
+        if (!file_exists($file)) return false;
+        $cache[$file] = unserialize(file_get_contents($file));
+        return $cache[$file];
+    }
+    
+    /**
+     * Determines a cache key identifier for a particular configuration
+     */
+    function getCacheKey($config) {
+        return md5(serialize(array($config->getBatch('HTML'), $config->getBatch('Attr'))));
+    }
+    
+    /**
+     * Determines file a particular configuration's definition is stored in
+     */
+    function getCacheFile($config) {
+        $key = HTMLPurifier_HTMLDefinition::getCacheKey($config);
+        return dirname(__FILE__) . '/HTMLDefinition/' . $key . '.ser';
+    }
+    
+    /**
+     * Saves HTMLDefinition to cache
+     */
+    function saveCache($config) {
+        $file = $this->getCacheFile($config);
+        $contents = serialize($this);
+        $fh = fopen($file, 'w');
+        fwrite($fh, $contents);
+        fclose($fh);
+    }
+    
     /**
      * Processes internals into form usable by HTMLPurifier internals. 
      * Modifying the definition after calling this function should not
@@ -190,6 +232,7 @@ class HTMLPurifier_HTMLDefinition
     function processModules() {
         
         $this->manager->setup($this->config);
+        $this->doctype = $this->manager->doctype;
         
         foreach ($this->manager->modules as $module) {
             foreach($module->info_tag_transform         as $k => $v) {

library/HTMLPurifier/HTMLModuleManager.php

@@ -99,6 +99,12 @@ class HTMLPurifier_HTMLModuleManager
      */
     var $doctypes;
     
+    /**
+     * Instance of current doctype
+     * @public
+     */
+    var $doctype;
+    
     /**
      * Instance of HTMLPurifier_AttrTypes
      * @public
@@ -288,8 +294,8 @@ class HTMLPurifier_HTMLModuleManager
         $this->trusted = $config->get('HTML', 'Trusted');
         
         // generate
-        $doctype = $this->doctypes->make($config);
+        $this->doctype = $this->doctypes->make($config);
-        $modules = $doctype->modules;
+        $modules = $this->doctype->modules;
         
         // take out the default modules that aren't allowed
         $lookup = $config->get('HTML', 'AllowedModules');
@@ -309,7 +315,7 @@ class HTMLPurifier_HTMLModuleManager
             $this->processModule($module);
         }
         
-        foreach ($doctype->tidyModules as $module) {
+        foreach ($this->doctype->tidyModules as $module) {
             $this->processModule($module);
             if (method_exists($this->modules[$module], 'construct')) {
                 $this->modules[$module]->construct($config);

tests/HTMLPurifier/AttrDef/URITest.php

@@ -209,17 +209,17 @@ class HTMLPurifier_AttrDef_URITest extends HTMLPurifier_AttrDefHarness
         $this->scheme = new HTMLPurifier_URISchemeMock($this);
         
         // here are the schemes we will support with overloaded mocks
-        $registry->setReturnReference('getScheme', $this->scheme, array('http', $this->config, $this->context));
+        $registry->setReturnReference('getScheme', $this->scheme, array('http', '*', '*'));
-        $registry->setReturnReference('getScheme', $this->scheme, array('mailto', $this->config, $this->context));
+        $registry->setReturnReference('getScheme', $this->scheme, array('mailto', '*', '*'));
         
         // default return value is false (meaning no scheme defined: reject)
-        $registry->setReturnValue('getScheme', false, array('*', $this->config, $this->context));
+        $registry->setReturnValue('getScheme', false, array('*', '*', '*'));
         
         if ($this->components === false) {
             $this->scheme->expectNever('validateComponents');
         } else {
-            $this->components[] = $this->config; // append the configuration
+            $this->components[] = '*'; // append the configuration
-            $this->components[] =& $this->context; // append context
+            $this->components[] = '*'; // append context
             $this->scheme->setReturnValue(
                 'validateComponents', $this->return_components, $this->components);
             $this->scheme->expectOnce('validateComponents', $this->components);