[1.7.0] Fix bug in HTMLPurifier_Harness that causes certain aspects of $input to change after parsing

- Add makeLookup() convenience function to HTMLModule
- Relocate SGML exclusion comment
- Add preliminary Bdo module test

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1049 48356398-32a2-884e-a903-53898d9a118a

library/HTMLPurifier/ElementDef.php

@@ -83,6 +83,14 @@ class HTMLPurifier_ElementDef
     
     /**
      * Lookup table of tags excluded from all descendants of this tag.
+     * @note SGML permits exclusions for all descendants, but this is
+     *       not possible with DTDs or XML Schemas. W3C has elected to
+     *       use complicated compositions of content_models to simulate
+     *       exclusion for children, but we go the simpler, SGML-style
+     *       route of flat-out exclusions, which correctly apply to
+     *       all descendants and not just children. Note that the XHTML
+     *       Modularization Abstract Modules are blithely unaware of such
+     *       distinctions.
      * @public
      */
     var $excludes = array();

library/HTMLPurifier/HTMLModule.php

@@ -223,6 +223,24 @@ class HTMLPurifier_HTMLModule
         }
         $attr[0] = $attr_includes;
     }
+    
+    /**
+     * Convenience function that generates a lookup table with boolean
+     * true as value.
+     * @param $list List of values to turn into a lookup
+     * @note You can also pass an arbitrary number of arguments in
+     *       place of the regular argument
+     * @return Lookup array equivalent of list
+     */
+    function makeLookup($list) {
+        if (is_string($list)) $list = func_get_args();
+        $ret = array();
+        foreach ($list as $value) {
+            if (is_null($value)) continue;
+            $ret[$value] = true;
+        }
+        return $ret;
+    }
 }
 
 ?>

library/HTMLPurifier/HTMLModule/Text.php

@@ -10,65 +10,61 @@ require_once 'HTMLPurifier/HTMLModule.php';
  *          - Block Structural (div, p)
  *          - Inline Phrasal (abbr, acronym, cite, code, dfn, em, kbd, q, samp, strong, var)
  *          - Inline Structural (br, span)
- *       We have elected not to follow suite, but this may change.
+ *       This module, functionally, does not distinguish between these
+ *       sub-modules, but the code is internally structured to reflect
+ *       these distinctions.
  */
 class HTMLPurifier_HTMLModule_Text extends HTMLPurifier_HTMLModule
 {
     
     var $name = 'Text';
-    
-    var $elements = array('abbr', 'acronym', 'address', 'blockquote',
-        'br', 'cite', 'code', 'dfn', 'div', 'em', 'h1', 'h2', 'h3',
-        'h4', 'h5', 'h6', 'kbd', 'p', 'pre', 'q', 'samp', 'span', 'strong',
-        'var');
-    
     var $content_sets = array(
-        'Heading' => 'h1 | h2 | h3 | h4 | h5 | h6',
-        'Block' => 'address | blockquote | div | p | pre',
-        'Inline' => 'abbr | acronym | br | cite | code | dfn | em | kbd | q | samp | span | strong | var',
         'Flow' => 'Heading | Block | Inline'
     );
     
     function HTMLPurifier_HTMLModule_Text() {
-        foreach ($this->elements as $element) {
+        
-            $this->info[$element] = new HTMLPurifier_ElementDef();
+        // Inline Phrasal -------------------------------------------------
-            // attributes
+        $this->addElement('abbr',    true, 'Inline', 'Inline', 'Common');
-            if ($element == 'br') {
+        $this->addElement('acronym', true, 'Inline', 'Inline', 'Common');
-                $this->info[$element]->attr = array(0 => array('Core'));
+        $this->addElement('cite',    true, 'Inline', 'Inline', 'Common');
-            } elseif ($element == 'blockquote' || $element == 'q') {
+        $this->addElement('code',    true, 'Inline', 'Inline', 'Common');
-                $this->info[$element]->attr = array(0 => array('Common'), 'cite' => 'URI');
+        $this->addElement('dfn',     true, 'Inline', 'Inline', 'Common');
-            } else {
+        $this->addElement('em',      true, 'Inline', 'Inline', 'Common');
-                $this->info[$element]->attr = array(0 => array('Common'));
+        $this->addElement('kbd',     true, 'Inline', 'Inline', 'Common');
-            }
+        $this->addElement('q',       true, 'Inline', 'Inline', 'Common', array('cite' => 'URI'));
-            // content models
+        $this->addElement('samp',    true, 'Inline', 'Inline', 'Common');
-            if ($element == 'br') {
+        $this->addElement('strong',  true, 'Inline', 'Inline', 'Common');
-                $this->info[$element]->content_model_type = 'empty';
+        $this->addElement('var',     true, 'Inline', 'Inline', 'Common');
-            } elseif ($element == 'blockquote') {
+        
-                $this->info[$element]->content_model = 'Heading | Block | List';
+        // Inline Structural ----------------------------------------------
-                $this->info[$element]->content_model_type = 'optional';
+        $this->addElement('span', true, 'Inline', 'Inline', 'Common');
-            } elseif ($element == 'div') {
+        $this->addElement('br',   true, 'Inline', 'Empty',  'Core');
-                $this->info[$element]->content_model = '#PCDATA | Flow';
+        
-                $this->info[$element]->content_model_type = 'optional';
+        // Block Phrasal --------------------------------------------------
-            } else {
+        $this->addElement('address', true, 'Block',  'Inline', 'Common');
-                $this->info[$element]->content_model = '#PCDATA | Inline';
+        $this->addElement('blockquote', true, 'Block',
-                $this->info[$element]->content_model_type = 'optional';
+            'Optional: Heading | Block | List', 'Common', array('cite' => 'URI') );
-            }
+        $pre =& $this->addElement('pre', true, 'Block', 'Inline', 'Common');
-        }
+        $pre->excludes = $this->makeLookup(
-        // SGML permits exclusions for all descendants, but this is
+            'img', 'big', 'small', 'object', 'applet', 'font', 'basefont' );
-        // not possible with DTDs or XML Schemas. W3C has elected to
+        $this->addElement('h1', true, 'Heading', 'Inline', 'Common');
-        // use complicated compositions of content_models to simulate
+        $this->addElement('h2', true, 'Heading', 'Inline', 'Common');
-        // exclusion for children, but we go the simpler, SGML-style
+        $this->addElement('h3', true, 'Heading', 'Inline', 'Common');
-        // route of flat-out exclusions. Note that the Abstract Module
+        $this->addElement('h4', true, 'Heading', 'Inline', 'Common');
-        // is blithely unaware of such distinctions.
+        $this->addElement('h5', true, 'Heading', 'Inline', 'Common');
-        $this->info['pre']->excludes = array_flip(array(
+        $this->addElement('h6', true, 'Heading', 'Inline', 'Common');
-            'img', 'big', 'small',
+        
-            'object', 'applet', 'font', 'basefont' // generally not allowed
+        // Block Structural -----------------------------------------------
-        ));
+        $p =& $this->addElement('p', true, 'Block', 'Inline', 'Common');
-        $this->info['p']->auto_close = array_flip(array(
+        // this seems really ad hoc: implementing some general
+        // heuristics would probably be better
+        $p->auto_close = $this->makeLookup(
             'address', 'blockquote', 'dd', 'dir', 'div', 'dl', 'dt',
             'h1', 'h2', 'h3', 'h4', 'h5', 'h6', 'hr', 'ol', 'p', 'pre',
-            'table', 'ul'
+            'table', 'ul' );
-        ));
+        $this->addElement('div', true, 'Block',  'Flow', 'Common');
+        
     }
     
 }

tests/HTMLPurifier/HTMLModule/BdoTest.php

@@ -0,0 +1,30 @@
+<?php
+
+require_once 'HTMLPurifier/HTMLModuleHarness.php';
+
+class HTMLPurifier_HTMLModule_BdoTest extends HTMLPurifier_HTMLModuleHarness
+{
+    
+    function test() {
+        
+        $this->assertResult(
+            '<span>
+                 <bdo
+                    id="test-id"
+                    class="class-name"
+                    style="font-weight:bold;"
+                    title="Title of tag"
+                    lang="en"
+                    xml:lang="en"
+                    dir="rtl"
+                 >
+                    #PCDATA <span>Inline</span>
+                 </bdo>
+             </span>', true, array('Attr.EnableID' => true)
+        );
+        
+    }
+    
+}
+
+?>

tests/HTMLPurifier/HTMLModuleHarness.php

@@ -0,0 +1,14 @@
+<?php
+
+require_once 'HTMLPurifier/StrategyHarness.php';
+require_once 'HTMLPurifier/Strategy/Core.php';
+
+class HTMLPurifier_HTMLModuleHarness extends HTMLPurifier_StrategyHarness
+{
+    function setup() {
+        parent::setup();
+        $this->obj = new HTMLPurifier_Strategy_Core();
+    }
+}
+
+?>

tests/HTMLPurifier/HTMLModuleTest.php

@@ -121,6 +121,30 @@ class HTMLPurifier_HTMLModuleTest extends UnitTestCase
         
     }
     
+    function test_makeLookup() {
+        
+        $module = new HTMLPurifier_HTMLModule();
+        
+        $this->assertIdentical(
+            $module->makeLookup('foo'),
+            array('foo' => true)
+        );
+        $this->assertIdentical(
+            $module->makeLookup(array('foo')),
+            array('foo' => true)
+        );
+        
+        $this->assertIdentical(
+            $module->makeLookup('foo', 'two'),
+            array('foo' => true, 'two' => true)
+        );
+        $this->assertIdentical(
+            $module->makeLookup(array('foo', 'two')),
+            array('foo' => true, 'two' => true)
+        );
+        
+    }
+    
 }
 
 ?>

tests/HTMLPurifier/Harness.php

@@ -73,12 +73,17 @@ class HTMLPurifier_Harness extends UnitTestCase
         $context->loadArray($context_array);
         
         if ($this->to_tokens && is_string($input)) {
-            $input = $this->lexer->tokenizeHTML($input, $config, $context);
+            // $func may cause $input to change, so "clone" another copy
+            // to sacrifice
+            $input   = $this->lexer->tokenizeHTML($s = $input, $config, $context);
+            $input_c = $this->lexer->tokenizeHTML($s, $config, $context);
+        } else {
+            $input_c = $input;
         }
         
         // call the function
         $func = $this->func;
-        $result = $this->obj->$func($input, $config, $context);
+        $result = $this->obj->$func($input_c, $config, $context);
         
         // test a bool result
         if (is_bool($result)) {

tests/test_files.php

@@ -62,6 +62,7 @@ $test_files[] = 'EntityParserTest.php';
 $test_files[] = 'GeneratorTest.php';
 $test_files[] = 'HTMLModuleManagerTest.php';
 $test_files[] = 'HTMLModuleTest.php';
+$test_files[] = 'HTMLModule/BdoTest.php';
 $test_files[] = 'IDAccumulatorTest.php';
 $test_files[] = 'LanguageFactoryTest.php';
 $test_files[] = 'LanguageTest.php';