improve parsing of emphasis

Prior to PHP 5.4.0 the default encoding for `htmlentities()`
and `htmlspecialchars` is "ISO-8859-1". For PHP 5.4+ is "UTF-8".

This ensures always the right encoding is used no matter the PHP version
and the locale settings.

.travis.yml

@@ -4,8 +4,4 @@ php:
   - 5.5
   - 5.4
   - 5.3
-  - 5.2
-
-matrix:
-  allow_failures:
-    - php: 5.2
+  - 5.2

Parsedown.php

@@ -110,7 +110,6 @@ class Parsedown
 
 		foreach ($lines as $line)
 		{
-			#
 			# fenced elements
 
 			switch ($element['type'])
@@ -168,7 +167,6 @@ class Parsedown
 				continue;
 			}
 
-			#
 			# composite elements
 
 			switch ($element['type'])
@@ -223,6 +221,8 @@ class Parsedown
 
 							$element['lines'] []= $line;
 
+							unset($element['interrupted']);
+
 							continue 2;
 						}
 					}
@@ -238,20 +238,10 @@ class Parsedown
 					break;
 			}
 
-			# ~
-
-			if ($line[0] >= 'a' and $line[0] !== '~' or $line[0] >= 'A' and $line[0] <= 'Z')
-			{
-				goto paragraph;
-			}
-
-			# ~
+			# indentation sensitive types
 
 			$deindented_line = $line;
 
-			#
-			# indentation sensitive types
-
 			switch ($line[0])
 			{
 				case ' ':
@@ -345,7 +335,6 @@ class Parsedown
 					break;
 			}
 
-			#
 			# indentation insensitive types
 
 			switch ($deindented_line[0])
@@ -500,9 +489,7 @@ class Parsedown
 				continue;
 			}
 
-			# ~
-
-			paragraph:
+			# paragraph
 
 			if ($element['type'] === 'p')
 			{
@@ -579,11 +566,15 @@ class Parsedown
 				case 'code_block':
 				case 'fenced_code_block':
 
-					$text = htmlentities($element['text'], ENT_NOQUOTES);
+					$text = htmlspecialchars($element['text'], ENT_NOQUOTES, 'UTF-8');
 
 					strpos($text, "\x1A\\") !== FALSE and $text = strtr($text, $this->escape_sequence_map);
 
-					$markup .= '<pre><code>'.$text.'</code></pre>'."\n";
+					$markup .= isset($element['language'])
+						? '<pre><code class="language-'.$element['language'].'">'.$text.'</code></pre>'
+						: '<pre><code>'.$text.'</code></pre>';
+
+					$markup .= "\n";
 
 					break;
 
@@ -638,6 +629,81 @@ class Parsedown
 
 		$index = 0;
 
+		# inline link / inline image (recursive)
+
+		if (strpos($text, '](') !== FALSE and preg_match_all('/(!?)(\[((?:[^\[\]]|(?2))*)\])\((.*?)\)/', $text, $matches, PREG_SET_ORDER))
+		{
+			foreach ($matches as $matches)
+			{
+				$url = $matches[4];
+
+				strpos($url, '&') !== FALSE and $url = preg_replace('/&(?!#?\w+;)/', '&amp;', $url);
+
+				if ($matches[1]) # image
+				{
+					$element = '<img alt="'.$matches[3].'" src="'.$url.'">';
+				}
+				else # link
+				{
+					$element_text = $this->parse_span_elements($matches[3]);
+
+					$element = '<a href="'.$url.'">'.$element_text.'</a>';
+				}
+
+				# ~
+
+				$code = "\x1A".'$'.$index;
+
+				$text = str_replace($matches[0], $code, $text);
+
+				$map[$code] = $element;
+
+				$index ++;
+			}
+		}
+
+		# reference link / reference image (recursive)
+
+		if ($this->reference_map and strpos($text, '[') !== FALSE and preg_match_all('/(!?)\[(.+?)\](?:\n?[ ]?\[(.*?)\])?/ms', $text, $matches, PREG_SET_ORDER))
+		{
+			foreach ($matches as $matches)
+			{
+				$link_definition = isset($matches[3]) && $matches[3]
+					? $matches[3]
+					: $matches[2]; # implicit
+
+				$link_definition = strtolower($link_definition);
+
+				if (isset($this->reference_map[$link_definition]))
+				{
+					$url = $this->reference_map[$link_definition];
+
+					strpos($url, '&') !== FALSE and $url = preg_replace('/&(?!#?\w+;)/', '&amp;', $url);
+
+					if ($matches[1]) # image
+					{
+						$element = '<img alt="'.$matches[2].'" src="'.$url.'">';
+					}
+					else # link
+					{
+						$element_text = $this->parse_span_elements($matches[2]);
+
+						$element = '<a href="'.$url.'">'.$element_text.'</a>';
+					}
+
+					# ~
+
+					$code = "\x1A".'$'.$index;
+
+					$text = str_replace($matches[0], $code, $text);
+
+					$map[$code] = $element;
+
+					$index ++;
+				}
+			}
+		}
+
 		# code span
 
 		if (strpos($text, '`') !== FALSE and preg_match_all('/`(.+?)`/', $text, $matches, PREG_SET_ORDER))
@@ -645,7 +711,7 @@ class Parsedown
 			foreach ($matches as $matches)
 			{
 				$element_text = $matches[1];
-				$element_text = htmlentities($element_text, ENT_NOQUOTES);
+				$element_text = htmlspecialchars($element_text, ENT_NOQUOTES, 'UTF-8');
 
 				# decodes escape sequences
 
@@ -669,104 +735,37 @@ class Parsedown
 			}
 		}
 
-		# inline link or image
-
-		if (strpos($text, '](') !== FALSE and preg_match_all('/(!?)(\[((?:[^\[\]]|(?2))*)\])\((.*?)\)/', $text, $matches, PREG_SET_ORDER)) # inline
-		{
-			foreach ($matches as $matches)
-			{
-				$url = $matches[4];
-
-				strpos($url, '&') !== FALSE and $url = preg_replace('/&(?!#?\w+;)/', '&amp;', $url);
-
-				if ($matches[1]) # image
-				{
-					$element = '<img alt="'.$matches[3].'" src="'.$url.'">';
-				}
-				else
-				{
-					$element_text = $this->parse_span_elements($matches[3]);
-
-					$element = '<a href="'.$url.'">'.$element_text.'</a>';
-				}
-
-				# ~
-
-				$code = "\x1A".'$'.$index;
-
-				$text = str_replace($matches[0], $code, $text);
-
-				$map[$code] = $element;
-
-				$index ++;
-			}
-		}
-
-		# reference link or image
-
-		if ($this->reference_map and strpos($text, '[') !== FALSE and preg_match_all('/(!?)\[(.+?)\](?:\n?[ ]?\[(.*?)\])?/ms', $text, $matches, PREG_SET_ORDER))
-		{
-			foreach ($matches as $matches)
-			{
-				$link_definition = isset($matches[3]) && $matches[3]
-					? $matches[3]
-					: $matches[2]; # implicit
-
-				$link_definition = strtolower($link_definition);
-
-				if (isset($this->reference_map[$link_definition]))
-				{
-					$url = $this->reference_map[$link_definition];
-
-					strpos($url, '&') !== FALSE and $url = preg_replace('/&(?!#?\w+;)/', '&amp;', $url);
-
-					if ($matches[1]) # image
-					{
-						$element = '<img alt="'.$matches[2].'" src="'.$url.'">';
-					}
-					else # anchor
-					{
-						$element_text = $this->parse_span_elements($matches[2]);
-
-						$element = '<a href="'.$url.'">'.$element_text.'</a>';
-					}
-
-					# ~
-
-					$code = "\x1A".'$'.$index;
-
-					$text = str_replace($matches[0], $code, $text);
-
-					$map[$code] = $element;
-
-					$index ++;
-				}
-			}
-		}
-
 		# automatic link
 
-		if (strpos($text, '<') !== FALSE and preg_match_all('/<((https?|ftp|dict):[^\^\s]+?)>/i', $text, $matches, PREG_SET_ORDER))
+		if (strpos($text, '://') !== FALSE)
 		{
-			foreach ($matches as $matches)
+			switch (TRUE)
 			{
-				$url = $matches[1];
+				case preg_match_all('{<(https?:[/]{2}[^\s]+)>}i', $text, $matches, PREG_SET_ORDER):
+				case preg_match_all('{\b(https?:[/]{2}[^\s]+)\b}i', $text, $matches, PREG_SET_ORDER):
 
-				strpos($url, '&') !== FALSE and $url = preg_replace('/&(?!#?\w+;)/', '&amp;', $url);
+					foreach ($matches as $matches)
+					{
+						$url = $matches[1];
 
-				$element = '<a href=":href">:text</a>';
-				$element = str_replace(':text', $url, $element);
-				$element = str_replace(':href', $url, $element);
+						strpos($url, '&') !== FALSE and $url = preg_replace('/&(?!#?\w+;)/', '&amp;', $url);
 
-				# ~
+						$element = '<a href=":href">:text</a>';
+						$element = str_replace(':text', $url, $element);
+						$element = str_replace(':href', $url, $element);
 
-				$code = "\x1A".'$'.$index;
+						# ~
 
-				$text = str_replace($matches[0], $code, $text);
+						$code = "\x1A".'$'.$index;
 
-				$map[$code] = $element;
+						$text = str_replace($matches[0], $code, $text);
 
-				$index ++;
+						$map[$code] = $element;
+
+						$index ++;
+					}
+
+					break;
 			}
 		}
 
@@ -777,20 +776,27 @@ class Parsedown
 
 		# ~
 
+		if (strpos($text, '~~') !== FALSE)
+		{
+			$text = preg_replace('/~~(?=\S)(.+?)(?<=\S)~~/s', '<del>$1</del>', $text);
+		}
+
 		if (strpos($text, '_') !== FALSE)
 		{
-			$text = preg_replace('/__(?=\S)(.+?)(?<=\S)__(?!_)/s', '<strong>$1</strong>', $text);
-			$text = preg_replace('/_(?=\S)(.+?)(?<=\S)_/s', '<em>$1</em>', $text);
+			$text = preg_replace('/__(?=\S)([^_]+?)(?<=\S)__/s', '<strong>$1</strong>', $text, -1, $count);
+			$text = preg_replace('/(\b|_)_(?=\S)([^_]+?)(?<=\S)_(\b|_)/s', '$1<em>$2</em>$3', $text);
+			$text = preg_replace('/__(?=\S)([^_]+?)(?<=\S)__/s', '<strong>$1</strong>', $text, -1, $count);
 		}
 
 		if (strpos($text, '*') !== FALSE)
 		{
-			$text = preg_replace('/\*\*(?=\S)(.+?)(?<=\S)\*\*(?!\*)/s', '<strong>$1</strong>', $text);
-			$text = preg_replace('/\*(?=\S)(.+?)(?<=\S)\*/s', '<em>$1</em>', $text);
+			$text = preg_replace('/\*\*(?=\S)([^*]+?)(?<=\S)\*\*/s', '<strong>$1</strong>', $text);
+			$text = preg_replace('/\*(?=\S)([^*]+?)(?<=\S)\*/s', '<em>$1</em>', $text);
+			$text = preg_replace('/\*\*(?=\S)([^*]+?)(?<=\S)\*\*/s', '<strong>$1</strong>', $text);
 		}
 
 		$text = strtr($text, $map);
 
 		return $text;
 	}
-}
+}

README.md

@@ -1,8 +1,20 @@
 ## Parsedown
 
-Parsedown is a Markdown parser for PHP. It is fast, consistent and easy to use.
+Better [Markdown](http://daringfireball.net/projects/markdown/) parser for PHP.
 
-[Home](http://parsedown.org)  · [Demo](http://parsedown.org/explorer/) · [Tests](http://parsedown.org/tests/)
+***
+
+[demo](http://parsedown.org/demo) · [tests](http://parsedown.org/tests/)
+
+***
+
+### Features
+
+* [fast](http://parsedown.org/speed)
+* [consistent](http://parsedown.org/consistency)
+* [GitHub Flavored](https://help.github.com/articles/github-flavored-markdown)
+* [tested](https://travis-ci.org/erusev/parsedown) in PHP 5.2, 5.3, 5.4 and 5.5
+* friendly to international input
 
 ### Installation
 

tests/Test.php

@@ -5,7 +5,7 @@ include 'Parsedown.php';
 class Test extends PHPUnit_Framework_TestCase
 {
 	const provider_dir = 'data/';
-	
+
 	/**
 	 * @dataProvider provider
 	 */
@@ -15,33 +15,41 @@ class Test extends PHPUnit_Framework_TestCase
 
 		$this->assertEquals($expected_markup, $actual_markup);
 	}
-	
+
 	function provider()
 	{
 		$provider = array();
-		
-		$DirectoryIterator = new DirectoryIterator(__DIR__ . '/' . self::provider_dir);
-		
+
+		$path = dirname(__FILE__).'/';
+
+		$DirectoryIterator = new DirectoryIterator($path . '/' . self::provider_dir);
+
 		foreach ($DirectoryIterator as $Item)
 		{
-			if ($Item->isFile() and $Item->getExtension() === 'md')
+			if ($Item->isFile())
 			{
+				$filename = $Item->getFilename();
+
+				$extension = pathinfo($filename, PATHINFO_EXTENSION);
+
+				if ($extension !== 'md')
+					continue;
+
 				$basename = $Item->getBasename('.md');
-				
-				$markdown = file_get_contents(__DIR__ . '/' . self::provider_dir . $basename . '.md');
-				
+
+				$markdown = file_get_contents($path . '/' . self::provider_dir . $basename . '.md');
+
 				if (!$markdown)
 					continue;
-				
-				$expected_markup = file_get_contents(__DIR__ . '/' . self::provider_dir . $basename . '.html');
+
+				$expected_markup = file_get_contents($path . '/' . self::provider_dir . $basename . '.html');
 				$expected_markup = str_replace("\r\n", "\n", $expected_markup);
 				$expected_markup = str_replace("\r", "\n", $expected_markup);
-				
+
 				$provider [] = array($markdown, $expected_markup);
 			}
 		}
-		
+
 		return $provider;
 	}
-}
-
+}

tests/data/em_strong.html

@@ -1,5 +1,6 @@
-<p><strong><em>em strong</em></strong></p>
-<p><strong><em>one</em> at the start</strong></p>
-<p><strong>one at the <em>end</em></strong></p>
-<p><strong>one <em>in the</em> middle</strong></p>
-<p><strong>one with <em>asterisks</em></strong></p>
+<p><strong><em>em strong</em> strong</strong></p>
+<p><strong>strong <em>em strong</em></strong></p>
+<p><strong>strong <em>em strong</em> strong</strong></p>
+<p><strong><em>em strong</em> strong</strong></p>
+<p><strong>strong <em>em strong</em></strong></p>
+<p><strong>strong <em>em strong</em> strong</strong></p>

tests/data/em_strong.md

@@ -1,9 +1,11 @@
-___em strong___
+___em strong_ strong__
 
-___one_ at the start__
+__strong _em strong___
 
-__one at the _end___
+__strong _em strong_ strong__
 
-__one _in the_ middle__
+***em strong* strong**
 
-**one with *asterisks***
+**strong *em strong***
+
+**strong *em strong* strong**

tests/data/emphasis.html

@@ -2,4 +2,6 @@
 <p><em>multiline
 emphasis</em></p>
 <p>_ this _ is not an emphasis, neither is _ this_, _this _, or _this*</p>
-<p>an empty emphasis __ ** is not an emphasis</p>
+<p>this_is_not_an_emphasis</p>
+<p>an empty emphasis __ ** is not an emphasis</p>
+<p>*mixed *<em>double and</em> single asterisk** spans</p>

tests/data/emphasis.md

@@ -5,4 +5,8 @@ emphasis_
 
 _ this _ is not an emphasis, neither is _ this_, _this _, or _this*
 
-an empty emphasis __ ** is not an emphasis
+this_is_not_an_emphasis
+
+an empty emphasis __ ** is not an emphasis
+
+*mixed **double and* single asterisk** spans

tests/data/inline_link.html

@@ -1,2 +1,3 @@
 <p><a href="http://example.com">link</a></p>
+<p><a href="http://example.com"><code>link</code></a></p>
 <p><a href="http://example.com"><img alt="MD Logo" src="http://parsedown.org/md.png"></a></p>

tests/data/inline_link.md

@@ -1,3 +1,5 @@
 [link](http://example.com)
 
+[`link`](http://example.com)
+
 [![MD Logo](http://parsedown.org/md.png)](http://example.com)

tests/data/multiline_list_paragraph.html

@@ -0,0 +1,7 @@
+<ul>
+<li>
+<p>li</p>
+<p>line
+line</p>
+</li>
+</ul>

tests/data/multiline_list_paragraph.md

@@ -0,0 +1,4 @@
+- li
+
+  line
+  line

tests/data/strikethrough.html

@@ -0,0 +1,3 @@
+<p><del>strikethrough</del></p>
+<p>in the <del>middle</del> of a sentence</p>
+<p>in the middle of a w<del>or</del>d</p>

tests/data/strikethrough.md

@@ -0,0 +1,5 @@
+~~strikethrough~~
+
+in the ~~middle~~ of a sentence
+
+in the middle of a w~~or~~d

tests/data/strong_em.html

@@ -0,0 +1,6 @@
+<p><em><strong>strong em</strong></em> </p>
+<p><em>em <strong>strong em</strong></em></p>
+<p><em><strong>strong em</strong> em</em></p>
+<p><em><strong>strong em</strong></em></p>
+<p><em>em <strong>strong em</strong></em></p>
+<p><em><strong>strong em</strong> em</em></p>

tests/data/strong_em.md

@@ -0,0 +1,11 @@
+***strong em*** 
+
+*em **strong em***
+
+***strong em** em*
+
+___strong em___
+
+_em __strong em___
+
+___strong em__ em_

tests/data/text_reference.html

@@ -3,4 +3,5 @@
 <p>[one][404] with no definition</p>
 <p><a href="http://example.com">multiline
 one</a> defined on 2 lines</p>
-<p><a href="http://example.com">one</a> with an upper case label</p>
+<p><a href="http://example.com">one</a> with an upper case label</p>
+<p><a href="http://example.com"><code>link</code></a></p>

tests/data/text_reference.md

@@ -13,4 +13,6 @@ one][website] defined on 2 lines
 
 [one][label] with an upper case label
 
-[LABEL]: http://example.com
+[LABEL]: http://example.com
+
+[`link`][website]

tests/data/url_autolinking.html

@@ -0,0 +1 @@
+<p>Here's an autolink <a href="http://example.com">http://example.com</a>.</p>

tests/data/url_autolinking.md

@@ -0,0 +1 @@
+Here's an autolink http://example.com.