Further refactor ConfigDoc, creating HTMLXSLTProcessor. Update NEWS.

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1104 48356398-32a2-884e-a903-53898d9a118a

NEWS

@@ -20,11 +20,14 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
   off, but it's not recommended.
 # New compact syntax for AttrDef objects that can be used to instantiate
   new objects via make()
+# Definitions (esp. HTMLDefinition) are now cached for a significant
+  performance boost
 ! HTML Purifier now works in PHP 4.3.2.
 ! Configuration form-editing API makes tweaking HTMLPurifier_Config a
   breeze!
 ! Configuration directives that accept hashes now allow new string
   format: key1,value1,key2,value2
+! ConfigDoc now factored into OOP design
 - Deprecated and removed EnableRedundantUTF8Cleaning. It didn't even work!
 . Unit test for ElementDef created, ElementDef behavior modified to
   be more flexible

configdoc/generate.php

@@ -2,6 +2,7 @@
 
 /**
  * Generates XML and HTML documents describing configuration.
+ * @note PHP 5 only!
  */
 
 /*
@@ -23,7 +24,7 @@ TODO:
 // Check and configure environment
 
 if (version_compare('5', PHP_VERSION, '>')) exit('Requires PHP 5 or higher.');
-error_reporting(E_ALL);
+error_reporting(E_ALL); // probably not possible to use E_STRICT
 
 
 // ---------------------------------------------------------------------------
@@ -32,22 +33,6 @@ error_reporting(E_ALL);
 require_once '../library/HTMLPurifier.auto.php';
 require_once 'library/ConfigDoc.auto.php';
 
-// ---------------------------------------------------------------------------
-// Setup convenience functions
-
-function appendHTMLDiv($document, $node, $html) {
-    global $purifier;
-    $html = $purifier->purify($html);
-    $dom_html = $document->createDocumentFragment();
-    $dom_html->appendXML($html);
-    
-    $dom_div = $document->createElement('div');
-    $dom_div->setAttribute('xmlns', 'http://www.w3.org/1999/xhtml');
-    $dom_div->appendChild($dom_html);
-    
-    $node->appendChild($dom_div);
-}
-
 
 // ---------------------------------------------------------------------------
 // Load copies of HTMLPurifier_ConfigDef and HTMLPurifier
@@ -83,41 +68,20 @@ $dom_document->save('configdoc.xml');
 // ---------------------------------------------------------------------------
 // Generate final output using XSLT
 
-// load the stylesheet
+// determine stylesheet name
 $xsl_stylesheet_name = 'plain';
 $xsl_stylesheet = "styles/$xsl_stylesheet_name.xsl";
-$xsl_dom_stylesheet = new DOMDocument();
-$xsl_dom_stylesheet->load($xsl_stylesheet);
 
-// setup the XSLT processor
+// transform
-$xsl_processor = new XSLTProcessor();
+$xslt_processor = new ConfigDoc_HTMLXSLTProcessor();
-
+$xslt_processor->importStylesheet($xsl_stylesheet);
-// perform the transformation
+$html_output = $xslt_processor->transformToHTML($dom_document);
-$xsl_processor->importStylesheet($xsl_dom_stylesheet);
-$html_output = $xsl_processor->transformToXML($dom_document);
-
-// some slight fudges to preserve backwards compatibility
-$html_output = str_replace('/>', ' />', $html_output); // <br /> not <br/>
-$html_output = str_replace(' xmlns=""', '', $html_output); // rm unnecessary xmlns
-
-if (class_exists('Tidy')) {
-    // cleanup output
-    $config = array(
-        'indent'        => true,
-        'output-xhtml'  => true,
-        'wrap'          => 80
-    );
-    $tidy = new Tidy;
-    $tidy->parseString($html_output, $config, 'utf8');
-    $tidy->cleanRepair();
-    $html_output = (string) $tidy;
-}
 
 // hack
 if (!defined('HTMLPURIFIER_CUSTOM_SCHEMA')) {
     // write it to a file (todo: parse into seperate pages)
     file_put_contents("$xsl_stylesheet_name.html", $html_output);
-} else {
+} elseif (defined('HTMLPURIFIER_SCRIPT_LOCATION')) {
     $html_output = str_replace('styles/plain.css', HTMLPURIFIER_SCRIPT_LOCATION . 'styles/plain.css', $html_output);
 }
 

configdoc/library/ConfigDoc.php

@@ -1,5 +1,6 @@
 <?php
 
+require_once 'ConfigDoc/HTMLXSLTProcessor.php';
 require_once 'ConfigDoc/XMLSerializer/Types.php';
 require_once 'ConfigDoc/XMLSerializer/ConfigSchema.php';
 

configdoc/library/ConfigDoc/HTMLXSLTProcessor.php

@@ -0,0 +1,56 @@
+<?php
+
+/**
+ * Special XSLTProcessor specifically for HTML documents. Loosely
+ * based off of XSLTProcessor, but not really
+ */
+class ConfigDoc_HTMLXSLTProcessor
+{
+    
+    protected $xsltProcessor;
+    
+    public function __construct() {
+        $this->xsltProcessor = new XSLTProcessor();
+    }
+    
+    /**
+     * Imports stylesheet for processor to use
+     * @param $xsl XSLT DOM tree, or filename of the XSL transformation
+     */
+    public function importStylesheet($xsl) {
+        if (is_string($xsl)) {
+            $xsl_file = $xsl;
+            $xsl = new DOMDocument();
+            $xsl->load($xsl_file);
+        }
+        return $this->xsltProcessor->importStylesheet($xsl);
+    }
+    
+    /**
+     * Transforms an XML file into HTML based on the stylesheet
+     * @param $xml XML DOM tree
+     */
+    public function transformToHTML($xml) {
+        $out = $this->xsltProcessor->transformToXML($xml);
+        
+        // fudges for HTML backwards compatibility
+        $out = str_replace('/>', ' />', $out); // <br /> not <br/>
+        $out = str_replace(' xmlns=""', '', $out); // rm unnecessary xmlns
+        if (class_exists('Tidy')) {
+            // cleanup output
+            $config = array(
+                'indent'        => true,
+                'output-xhtml'  => true,
+                'wrap'          => 80
+            );
+            $tidy = new Tidy;
+            $tidy->parseString($out, $config, 'utf8');
+            $tidy->cleanRepair();
+            $out = (string) $tidy;
+        }
+        return $out;
+    }
+    
+}
+
+?>

configdoc/library/ConfigDoc/XMLSerializer.php

@@ -1,9 +1,26 @@
 <?php
 
+/**
+ * The XMLSerializer hierarchy of classes consist of classes that take
+ * objects and serialize them into XML, specifically DOM, form; this
+ * super-class contains convenience functions for those classes.
+ */
 class ConfigDoc_XMLSerializer
 {
     
+    protected function appendHTMLDiv($document, $node, $html) {
+        // oh no, it's a global!
+        global $purifier;
+        $html = $purifier->purify($html);
+        $dom_html = $document->createDocumentFragment();
+        $dom_html->appendXML($html);
         
+        $dom_div = $document->createElement('div');
+        $dom_div->setAttribute('xmlns', 'http://www.w3.org/1999/xhtml');
+        $dom_div->appendChild($dom_html);
+        
+        $node->appendChild($dom_div);
+    }
     
 }
 

configdoc/library/ConfigDoc/XMLSerializer/ConfigSchema.php

@@ -10,7 +10,7 @@ class ConfigDoc_XMLSerializer_ConfigSchema extends ConfigDoc_XMLSerializer
      * @todo Split into sub-serializers
      * @param $schema HTMLPurifier_ConfigSchema to serialize
      */
-    function serialize($schema) {
+    public function serialize($schema) {
         $dom_document = new DOMDocument('1.0', 'UTF-8');
         $dom_root = $dom_document->createElement('configdoc');
         $dom_document->appendChild($dom_root);
@@ -35,7 +35,7 @@ class ConfigDoc_XMLSerializer_ConfigSchema extends ConfigDoc_XMLSerializer
             );
             $dom_namespace_description = $dom_document->createElement('description');
             $dom_namespace->appendChild($dom_namespace_description);
-            appendHTMLDiv($dom_document, $dom_namespace_description,
+            $this->appendHTMLDiv($dom_document, $dom_namespace_description,
                 $schema->info_namespace[$namespace_name]->description);
             
             foreach ($namespace_info as $name => $info) {
@@ -100,7 +100,7 @@ class ConfigDoc_XMLSerializer_ConfigSchema extends ConfigDoc_XMLSerializer
                             $dom_description->setAttribute('file', $file);
                             $dom_description->setAttribute('line', $line);
                         }
-                        appendHTMLDiv($dom_document, $dom_description, $description);
+                        $this->appendHTMLDiv($dom_document, $dom_description, $description);
                         $dom_descriptions->appendChild($dom_description);
                     }
                 }

configdoc/library/ConfigDoc/XMLSerializer/Types.php

@@ -9,7 +9,7 @@ class ConfigDoc_XMLSerializer_Types extends ConfigDoc_XMLSerializer
      * Serializes the types in a schema into DOM form
      * @param $schema HTMLPurifier_ConfigSchema owner of types to serialize
      */
-    function serialize($schema) {
+    public function serialize($schema) {
         $types_document = new DOMDocument('1.0', 'UTF-8');
         $types_root = $types_document->createElement('types');
         $types_document->appendChild($types_root);