[2.0.1] Officially add experimental auto-paragraphing and linkification functionality. Rename %Core.DefinitionCache to %Cache.DefinitionImpl. Have AutoParagraph handle even more edge cases. Fix MakeWellFormed bug.

git-svn-id: http://htmlpurifier.org/svnroot/htmlpurifier/trunk@1223 48356398-32a2-884e-a903-53898d9a118a
2024-09-18 18:25:18 +00:00 · 2007-06-24 20:29:50 +00:00 · 2007-06-24 20:29:50 +00:00 · b15cbbb42a
commit b15cbbb42a
parent 5f0663cad7
8 changed files with 297 additions and 69 deletions
--- a/4
+++ b/4
@ -12,6 +12,9 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 2.0.1, unknown release date
 ! Tag auto-closing now based on a ChildDef heuristic rather than a
  manually set auto_close array; some behavior may change
+! Experimental AutoFormat functionality added: auto-paragraph and
+  linkify your HTML input by setting %AutoFormat.AutoParagraph and
+  %AutoFormat.Linkify to true
 - Clean up special case code for <script> tags
 - Reorder includes for DefinitionCache decorators, fixes a possible
  missing class error
@ -26,6 +29,7 @@ NEWS ( CHANGELOG and HISTORY )                                     HTMLPurifier
 . HTMLDefinition printer updated with some of the new attributes
 . DefinitionCache keys reordered to reflect precedence: version number,
  hash, then revision number
+. %Core.DefinitionCache renamed to %Cache.DefinitionImpl

 2.0.0, released 2007-06-20
 # Completely refactored HTMLModuleManager, decentralizing safety
--- a/library/HTMLPurifier/ConfigSchema.php
+++ b/library/HTMLPurifier/ConfigSchema.php
@ -67,6 +67,7 @@ class HTMLPurifier_ConfigSchema {
        $this->defineNamespace('URI', 'Features regarding Uniform Resource Identifiers.');
        $this->defineNamespace('HTML', 'Configuration regarding allowed HTML.');
        $this->defineNamespace('CSS', 'Configuration regarding allowed CSS.');
+        $this->defineNamespace('AutoFormat', 'Configuration regarding auto-formatting functionality such as auto-paragraphing or linkification.');
        $this->defineNamespace('Output', 'Configuration relating to the generation of (X)HTML.');
        $this->defineNamespace('Cache', 'Configuration for DefinitionCache and related subclasses.');
        $this->defineNamespace('Test', 'Developer testing configuration for our unit tests.');
--- a/library/HTMLPurifier/DefinitionCacheFactory.php
+++ b/library/HTMLPurifier/DefinitionCacheFactory.php
@ -3,7 +3,7 @@
 require_once 'HTMLPurifier/DefinitionCache.php';

 HTMLPurifier_ConfigSchema::define(
-    'Core', 'DefinitionCache', 'Serializer', 'string/null', '
+    'Cache', 'DefinitionImpl', 'Serializer', 'string/null', '
 This directive defines which method to use when caching definitions,
 the complex data-type that makes HTML Purifier tick. Set to null
 to disable caching (not recommended, as you will see a definite
@ -11,7 +11,12 @@ performance degradation). This directive has been available since 2.0.0.
 ');

 HTMLPurifier_ConfigSchema::defineAllowedValues(
-    'Core', 'DefinitionCache', array('Serializer')
+    'Cache', 'DefinitionImpl', array('Serializer')
+);
+
+HTMLPurifier_ConfigSchema::defineAlias(
+    'Core', 'DefinitionCache',
+    'Cache', 'DefinitionImpl'
 );


@ -54,7 +59,7 @@ class HTMLPurifier_DefinitionCacheFactory
    function &create($type, $config) {
        // only one implementation as for right now, $config will
        // be used to determine implementation
-        $method = $config->get('Core', 'DefinitionCache');
+        $method = $config->get('Cache', 'DefinitionImpl');
        if ($method === null) {
            $null = new HTMLPurifier_DefinitionCache_Null($type);
            return $null;
--- a/library/HTMLPurifier/Injector/AutoParagraph.php
+++ b/library/HTMLPurifier/Injector/AutoParagraph.php
@ -2,6 +2,24 @@

 require_once 'HTMLPurifier/Injector.php';

+HTMLPurifier_ConfigSchema::define(
+    'AutoFormat', 'AutoParagraph', false, 'bool', '
+<p>
+  This directive turns on auto-paragraphing, where double newlines are
+  converted in to paragraphs whenever possible. Auto-paragraphing
+  applies when:
+</p>
+<ul>
+  <li>There are inline elements or text in the root node</li>
+  <li>There are inline elements or text with double newlines or
+      block elements in nodes that allow paragraph tags</li>
+  <li>There are double newlines in paragraph tags</li>
+</ul>
+<p>
+  This directive has been available since 2.0.1.
+</p>
+');
+
 /**
 * Injector that auto paragraphs text in the root node based on
 * double-spacing.
@ -11,27 +29,94 @@ class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
    
    function handleText(&$token) {
        $text = $token->data;
-        // $token is the focus: if processing is needed, it gets
-        // turned into an array of tokens that will replace the
-        // original token
        if (empty($this->currentNesting)) {
            if (!$this->allowsElement('p')) return;
-            // we're in root node, and the root node allows paragraphs
-            // start a paragraph since we just hit some text
+            // case 1: we're in root node (and it allows paragraphs)
            $token = array(new HTMLPurifier_Token_Start('p'));
            $this->_splitText($text, $token);
        } elseif ($this->currentNesting[count($this->currentNesting)-1]->name == 'p') {
-            // we're not in root node but we're in a paragraph, so don't 
-            // add a paragraph start tag but still perform processing
+            // case 2: we're in a paragraph
            $token = array();
            $this->_splitText($text, $token);
+        } elseif ($this->allowsElement('p')) {
+            // case 3: we're in an element that allows paragraphs
+            if (strpos($text, PHP_EOL . PHP_EOL) !== false) {
+                // case 3.1: this text node has a double-newline
+                $token = array(new HTMLPurifier_Token_Start('p'));
+                $this->_splitText($text, $token);
+            } else {
+                $ok = false;
+                // test if up-coming tokens are either block or have
+                // a double newline in them
+                for ($i = $this->inputIndex + 1; isset($this->inputTokens[$i]); $i++) {
+                    if ($this->inputTokens[$i]->type == 'start'){
+                        if (!$this->_isInline($this->inputTokens[$i])) {
+                            $ok = true;
+                        }
+                        break;
+                    }
+                    if ($this->inputTokens[$i]->type == 'end') break;
+                    if ($this->inputTokens[$i]->type == 'text') {
+                        if (strpos($this->inputTokens[$i]->data, PHP_EOL . PHP_EOL) !== false) {
+                            $ok = true;
+                        }
+                        if (!$this->inputTokens[$i]->is_whitespace) break;
+                    }
+                }
+                if ($ok) {
+                    // case 3.2: this text node is next to another node
+                    // that will start a paragraph
+                    $token = array(new HTMLPurifier_Token_Start('p'), $token);
+                }
+            }
        }
+        
    }
    
    function handleStart(&$token) {
-        // check if we're inside a tag already, if so, don't add
-        // paragraph tags
-        if (!empty($this->currentNesting)) return;
+        // check if we're inside a tag already
+        if (!empty($this->currentNesting)) {
+            if ($this->allowsElement('p')) {
+                // special case: we're in an element that allows paragraphs
+                
+                // this token is already paragraph, abort
+                if ($token->name == 'p') return;
+                
+                // check if this token is adjacent to the parent
+                if ($this->inputTokens[$this->inputIndex - 1]->type != 'start') {
+                    // not adjacent, we can abort early
+                    // add lead paragraph tag if our token is inline
+                    if ($this->_isInline($token)) {
+                        $token = array(new HTMLPurifier_Token_Start('p'), $token);
+                    }
+                    return;
+                }
+                
+                // this token is the first child of the element that allows
+                // paragraph. We have to peek ahead and see whether or not
+                // there is anything inside that suggests that a paragraph
+                // will be needed
+                $ok = false;
+                // maintain a mini-nesting counter, this lets us bail out
+                // early if possible
+                $j = 2; // current nesting, is two due to parent and this start
+                for ($i = $this->inputIndex + 1; isset($this->inputTokens[$i]); $i++) {
+                    if ($this->inputTokens[$i]->type == 'start') $j++;
+                    if ($this->inputTokens[$i]->type == 'end') $j--;
+                    if ($this->inputTokens[$i]->type == 'text') {
+                        if (strpos($this->inputTokens[$i]->data, PHP_EOL . PHP_EOL) !== false) {
+                            $ok = true;
+                            break;
+                        }
+                    }
+                    if ($j <= 0) break;
+                }
+                if ($ok) {
+                    $token = array(new HTMLPurifier_Token_Start('p'), $token);
+                }
+            }
+            return;
+        }
        
        // check if the start tag counts as a "block" element
        if (!$this->_isInline($token)) return;
@ -56,55 +141,79 @@ class HTMLPurifier_Injector_AutoParagraph extends HTMLPurifier_Injector
        
        // remove empty paragraphs
        $paragraphs = array();
+        $needs_start = false;
+        $first = true;
        foreach ($raw_paragraphs as $par) {
-            if (trim($par) !== '') $paragraphs[] = $par;
+            if (trim($par) !== '') {
+               $paragraphs[] = $par;
+            } elseif (empty($result) && $first) {
+                // The empty result indicates that the AutoParagraph
+                // injector did not add any start paragraph tokens.
+                // The fact that the first paragraph is empty indicates
+                // that there was a double-newline at the start of the
+                // data.
+                // Combined together, this means that we are in a paragraph,
+                // and the newline means we should start a new one.
+                $result[] = new HTMLPurifier_Token_End('p');
+                // However, the start token should only be added if 
+                // there is more processing to be done (i.e. there are
+                // real paragraphs in here). If there are none, the
+                // next start paragraph tag will be handled by the
+                // next run-around the injector
+                $needs_start = true;
+            }
+            $first = false;
        }
        
        // check if there are no "real" paragraphs to be processed
-        if (empty($paragraphs) && count($raw_paragraphs) > 1) {
-            $result[] = new HTMLPurifier_Token_End('p');
+        if (empty($paragraphs)) {
            return;
        }
        
+        // add a start tag if an end tag was added while processing
+        // the raw paragraphs (that happens if there's a leading double
+        // newline)
+        if ($needs_start) $result[] = new HTMLPurifier_Token_Start('p');
+        
        // append the paragraphs onto the result
        foreach ($paragraphs as $par) {
            $result[] = new HTMLPurifier_Token_Text($par);
            $result[] = new HTMLPurifier_Token_End('p');
            $result[] = new HTMLPurifier_Token_Start('p');
        }
-        array_pop($result); // remove trailing start token
+        
+        // remove trailing start token, if one is needed, it will
+        // be handled the next time this injector is called
+        array_pop($result);
+        
+        // check the outside to determine whether or not the
+        // end paragraph tag should be removed. It should be removed
+        // unless the next non-whitespace token is a paragraph
+        // or a block element.
+        
+        $remove_paragraph_end = true;
+        // Start of the checks one after the current token's index
+        for ($i = $this->inputIndex + 1; isset($this->inputTokens[$i]); $i++) {
+            if ($this->inputTokens[$i]->type == 'start' || $this->inputTokens[$i]->type == 'empty') {
+                $remove_paragraph_end = $this->_isInline($this->inputTokens[$i]);
+                break;
+            }
+            // check if we can abort early (whitespace means we carry-on!)
+            if ($this->inputTokens[$i]->type == 'text' && !$this->inputTokens[$i]->is_whitespace) break;
+            // end tags will automatically be handled by MakeWellFormed,
+            // so we don't have to worry about them
+            if ($this->inputTokens[$i]->type == 'end') break;
+        }
        
        // check the outside to determine whether or not the
        // end paragraph tag should be removed
-        if ($this->_removeParagraphEnd()) {
+        if ($remove_paragraph_end) {
            array_pop($result);
        }
        
        
    }
    
-    /**
-     * Returns boolean whether or not to remove the paragraph end tag
-     * that was automatically added. The paragraph end tag should be
-     * removed unless the next token is a paragraph or block element.
-     * @private
-     */
-    function _removeParagraphEnd() {
-        $tokens =& $this->inputTokens;
-        $remove_paragraph_end = true;
-        // Start of the checks one after the current token's index
-        for ($i = $this->inputIndex + 1; isset($tokens[$i]); $i++) {
-            if ($tokens[$i]->type == 'start' || $tokens[$i]->type == 'empty') {
-                $remove_paragraph_end = $this->_isInline($tokens[$i]);
-                break;
-            }
-            // check if we can abort early (whitespace means we carry-on!)
-            if ($tokens[$i]->type == 'text' && !$tokens[$i]->is_whitespace) break;
-            if ($tokens[$i]->type == 'end') break; // nonsensical
-        }
-        return $remove_paragraph_end;
-    }
-    
    /**
     * Returns true if passed token is inline (and, ergo, allowed in
     * paragraph tags)
--- a/library/HTMLPurifier/Injector/Linkify.php
+++ b/library/HTMLPurifier/Injector/Linkify.php
@ -2,6 +2,14 @@

 require_once 'HTMLPurifier/Injector.php';

+HTMLPurifier_ConfigSchema::define(
+    'AutoFormat', 'Linkify', false, 'bool', '
+<p>
+  This directive turns on linkification, auto-linking http, ftp and
+  https URLs. This directive has been available since 2.0.1.
+</p>
+');
+
 /**
 * Injector that converts http, https and ftp text URLs to actual links.
 */
--- a/library/HTMLPurifier/Strategy/MakeWellFormed.php
+++ b/library/HTMLPurifier/Strategy/MakeWellFormed.php
@ -7,26 +7,6 @@ require_once 'HTMLPurifier/Generator.php';
 require_once 'HTMLPurifier/Injector/AutoParagraph.php';
 require_once 'HTMLPurifier/Injector/Linkify.php';

-HTMLPurifier_ConfigSchema::define(
-    'Core', 'AutoParagraph', false, 'bool', '
-<p>
-  This directive will cause HTML Purifier to automatically paragraph text
-  in the document fragment root based on two newlines and block tags.
-  This directive has been available since 2.0.1.
-</p>
-'
-);
-
-HTMLPurifier_ConfigSchema::define(
-    'Core', 'AutoLinkify', false, 'bool', '
-<p>
-  This directive will cause HTML Purifier to automatically linkify
-  text that looks like URLs. This directive has been available since
-  2.0.1.
-</p>
-'
-);
-
 /**
 * Takes tokens makes them well-formed (balance end tags, etc.)
 */
@ -70,11 +50,11 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
        
        // we need a generic way of adding injectors, and also its own
        // configuration namespace
-        if ($config->get('Core', 'AutoParagraph')) {
+        if ($config->get('AutoFormat', 'AutoParagraph')) {
            $this->injectors[] = new HTMLPurifier_Injector_AutoParagraph();
        }
        
-        if ($config->get('Core', 'AutoLinkify')) {
+        if ($config->get('AutoFormat', 'Linkify')) {
            $this->injectors[] = new HTMLPurifier_Injector_Linkify();
        }
        
@ -163,7 +143,7 @@ class HTMLPurifier_Strategy_MakeWellFormed extends HTMLPurifier_Strategy
                
                // injector handler code; duplicated for performance reasons
                foreach ($this->injectors as $i => $x) {
-                    if (!$x->skip[$i]) $x->handleStart($token, $config, $context);
+                    if (!$x->skip) $x->handleStart($token, $config, $context);
                    if (is_array($token)) {
                        $this->currentInjector = $i;
                        break;
--- a/tests/HTMLPurifier/ConfigTest.php
+++ b/tests/HTMLPurifier/ConfigTest.php
@ -278,8 +278,8 @@ class HTMLPurifier_ConfigTest extends UnitTestCase
    }
    
    function test_getDefinition() {
-        CS::defineNamespace('Core', 'Core stuff');
-        CS::define('Core', 'DefinitionCache', null, 'string/null', 'Cache?');
+        CS::defineNamespace('Cache', 'Cache stuff');
+        CS::define('Cache', 'DefinitionImpl', null, 'string/null', 'Cache?');
        CS::defineNamespace('Crust', 'Krusty Krabs');
        $config = HTMLPurifier_Config::createDefault();
        $this->expectError("Definition of Crust type not supported");
--- a/tests/HTMLPurifier/Strategy/MakeWellFormedTest.php
+++ b/tests/HTMLPurifier/Strategy/MakeWellFormedTest.php
@ -76,7 +76,7 @@ class HTMLPurifier_Strategy_MakeWellFormedTest extends HTMLPurifier_StrategyHarn
    }
    
    function testAutoParagraph() {
-        $this->config = array('Core.AutoParagraph' => true);
+        $this->config = array('AutoFormat.AutoParagraph' => true);
        
        $this->assertResult(
            'Foobar',
@ -175,19 +175,140 @@ Par
            '<p>Par</p>'
        );
        
+        $this->assertResult(
+'<div>Par1
+
+Par2</div>',
+            '<div><p>Par1</p><p>Par2</p></div>'
+        );
+        
+        $this->assertResult(
+'<div><b>Par1</b>
+
+Par2</div>',
+            '<div><p><b>Par1</b></p><p>Par2</p></div>'
+        );
+        
+        $this->assertResult('<div>Par1</div>');
+        
+        $this->assertResult(
+'<div><b>Par1</b>
+
+<i>Par2</i></div>',
+            '<div><p><b>Par1</b></p><p><i>Par2</i></p></div>'
+        );
+        
+        $this->assertResult(
+'<pre><b>Par1</b>
+
+<i>Par2</i></pre>',
+            true
+        );
+        
+        $this->assertResult(
+'<div><p>Foo
+
+Bar</p></div>',
+            '<div><p>Foo</p><p>Bar</p></div>'
+        );
+        
+        $this->assertResult(
+'<div><p><b>Foo</b>
+
+<i>Bar</i></p></div>',
+            '<div><p><b>Foo</b></p><p><i>Bar</i></p></div>'
+        );
+        
+        $this->assertResult(
+'<div><b>Foo</b></div>',
+            '<div><b>Foo</b></div>'
+        );
+        
+        $this->assertResult(
+'<blockquote>Par1
+
+Par2</blockquote>',
+            '<blockquote><p>Par1</p><p>Par2</p></blockquote>'
+        );
+        
+        $this->assertResult(
+'<ul><li>Foo</li>
+
+<li>Bar</li></ul>', true
+        );
+        
+        $this->assertResult(
+'<div>
+
+Bar
+
+</div>', 
+        '<div><p>Bar</p></div>'
+        );
+        
+        $this->assertResult(
+'<b>Par1</b>a
+
+
+
+Par2', 
+        '<p><b>Par1</b>a</p><p>Par2</p>'
+        );
+        
+        $this->assertResult(
+'Par1
+
+Par2</p>', 
+        '<p>Par1</p><p>Par2</p>'
+        );
+        
+        $this->assertResult(
+'Par1
+
+Par2</div>', 
+        '<p>Par1</p><p>Par2</p>'
+        );
+        
+        $this->assertResult(
+'<div>
+Par1
+</div>', true
+        );
+        
+        $this->assertResult(
+'<div>Par1
+
+<div>Par2</div></div>',
+'<div><p>Par1</p><div>Par2</div></div>'
+        );
+        
+        $this->assertResult(
+'<div>Par1
+<div>Par2</div></div>',
+'<div><p>Par1
+</p><div>Par2</div></div>'
+        );
+        
+        $this->assertResult(
+'Par1
+<div>Par2</div>',
+'<p>Par1
+</p><div>Par2</div>'
+        );
+        
        $this->assertResult(
 'Par

 Par2',
            true,
-            array('Core.AutoParagraph' => true, 'HTML.Parent' => 'span')
+            array('AutoFormat.AutoParagraph' => true, 'HTML.Parent' => 'span')
        );
        
    }
    
    function testLinkify() {
        
-        $this->config = array('Core.AutoLinkify' => true);
+        $this->config = array('AutoFormat.Linkify' => true);
        
        $this->assertResult(
            'http://example.com',
@ -212,7 +333,7 @@ Par2',
    
    function testMultipleInjectors() {
        
-        $this->config = array('Core.AutoParagraph' => true, 'Core.AutoLinkify' => true);
+        $this->config = array('AutoFormat.AutoParagraph' => true, 'AutoFormat.Linkify' => true);
        
        $this->assertResult(
            'Foobar',