improve extensibility

Parsedown.php

@@ -372,7 +372,9 @@ class Parsedown
         if (isset($Line['text'][3]) and $Line['text'][3] === '-' and $Line['text'][2] === '-' and $Line['text'][1] === '!')
         {
             $Block = array(
-                'element' => $Line['body'],
+                'element' => array(
+                    'text' => $Line['body'],
+                ),
             );
 
             if (preg_match('/-->$/', $Line['text']))
@@ -391,7 +393,7 @@ class Parsedown
             return;
         }
 
-        $Block['element'] .= "\n" . $Line['body'];
+        $Block['element']['text'] .= "\n" . $Line['body'];
 
         if (preg_match('/-->$/', $Line['text']))
         {
@@ -643,30 +645,78 @@ class Parsedown
             return;
         }
 
-        if (preg_match('/^<(\w[\w\d]*)(?:[ ][^>]*)?(\/?)[ ]*>/', $Line['text'], $matches))
+        $attrName = '[a-zA-Z_:][\w:.-]*';
-        {
+        $attrValue = '(?:[^"\'=<>`\s]+|".*?"|\'.*?\')';
-            if (in_array($matches[1], $this->textLevelElements))
+
+        preg_match('/^<(\w[\d\w]*)((?:\s'.$attrName.'(?:\s*=\s*'.$attrValue.')?)*)\s*(\/?)>/', $Line['text'], $matches);
+
+        if ( ! $matches or in_array($matches[1], $this->textLevelElements))
         {
             return;
         }
 
         $Block = array(
-                'element' => $Line['body'],
+            'depth' => 0,
+            'element' => array(
+                'name' => $matches[1],
+                'text' => null,
+            ),
         );
 
-            if ($matches[2] or in_array($matches[1], $this->voidElements) or preg_match('/<\/'.$matches[1].'>[ ]*$/', $Line['text']))
+        $remainder = substr($Line['text'], strlen($matches[0]));
+
+        if (trim($remainder) === '')
+        {
+            if ($matches[3] or in_array($matches[1], $this->voidElements))
             {
                 $Block['closed'] = true;
             }
+        }
         else
         {
-                $Block['depth'] = 0;
+            if ($matches[3] or in_array($matches[1], $this->voidElements))
-                $Block['name'] = $matches[1];
+            {
+                return;
+            }
+
+            preg_match('/(.*)<\/'.$matches[1].'>\s*$/i', $remainder, $nestedMatches);
+
+            if ($nestedMatches)
+            {
+                $Block['closed'] = true;
+                $Block['element']['text'] = $nestedMatches[1];
+            }
+            else
+            {
+                $Block['element']['text'] = $remainder;
+            }
+        }
+
+        if ( ! $matches[2])
+        {
+            return $Block;
+        }
+
+        preg_match_all('/\s('.$attrName.')(?:\s*=\s*('.$attrValue.'))?/', $matches[2], $nestedMatches, PREG_SET_ORDER);
+
+        foreach ($nestedMatches as $nestedMatch)
+        {
+            if ( ! isset($nestedMatch[2]))
+            {
+                $Block['element']['attributes'][$nestedMatch[1]] = '';
+            }
+            elseif ($nestedMatch[2][0] === '"' or $nestedMatch[2][0] === '\'')
+            {
+                $Block['element']['attributes'][$nestedMatch[1]] = substr($nestedMatch[2], 1, - 1);
+            }
+            else
+            {
+                $Block['element']['attributes'][$nestedMatch[1]] = $nestedMatch[2];
+            }
         }
 
         return $Block;
     }
-    }
 
     protected function addToMarkup($Line, array $Block)
     {
@@ -675,12 +725,12 @@ class Parsedown
             return;
         }
 
-        if (preg_match('/<'.$Block['name'].'([ ].*[\'"])?[ ]*>/', $Line['text'])) # opening tag
+        if (preg_match('/^<'.$Block['element']['name'].'(?:\s.*[\'"])?\s*>/i', $Line['text'])) # open
         {
             $Block['depth'] ++;
         }
 
-        if (stripos($Line['text'], '</'.$Block['name'].'>') !== false) # closing tag
+        if (preg_match('/(.*?)<\/'.$Block['element']['name'].'>\s*$/i', $Line['text'], $matches)) # close
         {
             if ($Block['depth'] > 0)
             {
@@ -688,18 +738,25 @@ class Parsedown
             }
             else
             {
+                $Block['element']['text'] .= "\n";
+
                 $Block['closed'] = true;
             }
+
+            $Block['element']['text'] .= $matches[1];
         }
 
         if (isset($Block['interrupted']))
         {
-            $Block['element'] .= "\n";
+            $Block['element']['text'] .= "\n";
 
             unset($Block['interrupted']);
         }
 
-        $Block['element'] .= "\n".$Line['body'];
+        if ( ! isset($Block['closed']))
+        {
+            $Block['element']['text'] .= "\n".$Line['body'];
+        }
 
         return $Block;
     }
@@ -871,6 +928,7 @@ class Parsedown
                 'id' => strtolower($matches[1]),
                 'data' => array(
                     'url' => $matches[2],
+                    'title' => null,
                 ),
             );
 
@@ -906,12 +964,21 @@ class Parsedown
 
     protected function element(array $Element)
     {
-        $markup = '<'.$Element['name'];
+        $markup = '';
+
+        if (isset($Element['name']))
+        {
+            $markup .= '<'.$Element['name'];
 
             if (isset($Element['attributes']))
             {
                 foreach ($Element['attributes'] as $name => $value)
                 {
+                    if ($value === null)
+                    {
+                        continue;
+                    }
+
                     $markup .= ' '.$name.'="'.$value.'"';
                 }
             }
@@ -919,7 +986,17 @@ class Parsedown
             if (isset($Element['text']))
             {
                 $markup .= '>';
+            }
+            else
+            {
+                $markup .= ' />';
 
+                return $markup;
+            }
+        }
+
+        if (isset($Element['text']))
+        {
             if (isset($Element['handler']))
             {
                 $markup .= $this->$Element['handler']($Element['text']);
@@ -928,12 +1005,11 @@ class Parsedown
             {
                 $markup .= $Element['text'];
             }
-
-            $markup .= '</'.$Element['name'].'>';
         }
-        else
+
+        if (isset($Element['name']))
         {
-            $markup .= ' />';
+            $markup .= '</'.$Element['name'].'>';
         }
 
         return $markup;
@@ -950,16 +1026,7 @@ class Parsedown
                 continue;
             }
 
-            $markup .= "\n";
+            $markup .= "\n" . $this->element($Element);
-
-            if (is_string($Element)) # because of Markup
-            {
-                $markup .= $Element;
-
-                continue;
-            }
-
-            $markup .= $this->element($Element);
         }
 
         $markup .= "\n";
@@ -972,8 +1039,8 @@ class Parsedown
     #
 
     protected $SpanTypes = array(
-        '!' => array('Link'), # ?
         '"' => array('QuotationMark'),
+        '!' => array('Image'),
         '&' => array('Ampersand'),
         '*' => array('Emphasis'),
         '/' => array('Url'),
@@ -1234,92 +1301,98 @@ class Parsedown
         }
     }
 
-    protected function identifyLink($Excerpt)
+    protected function identifyImage($Excerpt)
     {
-        $extent = $Excerpt['text'][0] === '!' ? 1 : 0;
+        if ( ! isset($Excerpt['text'][1]) or $Excerpt['text'][1] !== '[')
-
-        if (strpos($Excerpt['text'], ']') and preg_match('/\[((?:[^][]|(?R))*)\]/', $Excerpt['text'], $matches))
-        {
-            $Link = array('text' => $matches[1], 'label' => strtolower($matches[1]));
-
-            $extent += strlen($matches[0]);
-
-            $substring = substr($Excerpt['text'], $extent);
-
-            if (preg_match('/^\s*\[([^][]+)\]/', $substring, $matches))
-            {
-                $Link['label'] = strtolower($matches[1]);
-
-                if (isset($this->Definitions['Reference'][$Link['label']]))
-                {
-                    $Link += $this->Definitions['Reference'][$Link['label']];
-
-                    $extent += strlen($matches[0]);
-                }
-                else
-                {
-                    return;
-                }
-            }
-            elseif (isset($this->Definitions['Reference'][$Link['label']]))
-            {
-                $Link += $this->Definitions['Reference'][$Link['label']];
-
-                if (preg_match('/^[ ]*\[\]/', $substring, $matches))
-                {
-                    $extent += strlen($matches[0]);
-                }
-            }
-            elseif (preg_match('/^\([ ]*(.*?)(?:[ ]+[\'"](.+?)[\'"])?[ ]*\)/', $substring, $matches))
-            {
-                $Link['url'] = $matches[1];
-
-                if (isset($matches[2]))
-                {
-                    $Link['title'] = $matches[2];
-                }
-
-                $extent += strlen($matches[0]);
-            }
-            else
-            {
-                return;
-            }
-        }
-        else
         {
             return;
         }
 
-        $url = str_replace(array('&', '<'), array('&amp;', '&lt;'), $Link['url']);
+        $Excerpt['text'] = substr($Excerpt['text'], 1);
 
-        if ($Excerpt['text'][0] === '!')
+        $Span = $this->identifyLink($Excerpt);
-        {
+
-            $Element = array(
+        $Span['extent'] ++;
+
+        $Span['element'] = array(
             'name' => 'img',
             'attributes' => array(
-                    'src' => $url,
+                'src' => $Span['element']['attributes']['href'],
-                    'alt' => $Link['text'],
+                'alt' => $Span['element']['text'],
+                'title' => $Span['element']['attributes']['title'],
             ),
         );
+
+        return $Span;
     }
-        else
+
+    protected function identifyLink($Excerpt)
     {
         $Element = array(
             'name' => 'a',
             'handler' => 'line',
-                'text' => $Link['text'],
+            'text' => null,
             'attributes' => array(
-                    'href' => $url,
+                'href' => null,
+                'title' => null,
             ),
         );
+
+        $extent = 0;
+
+        $remainder = $Excerpt['text'];
+
+        if (preg_match('/\[((?:[^][]|(?R))*)\]/', $remainder, $matches))
+        {
+            $Element['text'] = $matches[1];
+
+            $extent += strlen($matches[0]);
+
+            $remainder = substr($remainder, $extent);
+        }
+        else
+        {
+            return;
         }
 
-        if (isset($Link['title']))
+        if (preg_match('/^\([ ]*([^ ]+)(?:[ ]+(".+?"|\'.+?\'))?[ ]*\)/', $remainder, $matches))
         {
-            $Element['attributes']['title'] = $Link['title'];
+            $Element['attributes']['href'] = $matches[1];
+
+            if (isset($matches[2]))
+            {
+                $Element['attributes']['title'] = substr($matches[2], 1, - 1);
             }
 
+            $extent += strlen($matches[0]);
+        }
+        else
+        {
+            if (preg_match('/^\s*\[(.*?)\]/', $remainder, $matches))
+            {
+                $definition = $matches[1] ?: $Element['text'];
+                $definition = strtolower($definition);
+
+                $extent += strlen($matches[0]);
+            }
+            else
+            {
+                $definition = strtolower($Element['text']);
+            }
+
+            if ( ! isset($this->Definitions['Reference'][$definition]))
+            {
+                return;
+            }
+
+            $Definition = $this->Definitions['Reference'][$definition];
+
+            $Element['attributes']['href'] = $Definition['url'];
+            $Element['attributes']['title'] = $Definition['title'];
+        }
+
+        $Element['attributes']['href'] = str_replace(array('&', '<'), array('&amp;', '&lt;'), $Element['attributes']['href']);
+
         return array(
             'extent' => $extent,
             'element' => $Element,

test/data/self-closing_html.html

@@ -1,4 +1,4 @@
-<hr>
+<hr />
 <p>paragraph</p>
 <hr />
 <p>paragraph</p>
@@ -8,5 +8,5 @@
 <p>paragraph</p>
 <hr class="foo" id="bar" />
 <p>paragraph</p>
-<hr class="foo" id="bar" >
+<hr class="foo" id="bar" />
 <p>paragraph</p>