Improved parsing of emphasis around punctuation

See commonmark/commonmark-spec@2da7c3f
thephpleague · Jan 2, 2015 · f3566b1 · f3566b1
1 parent 1d9dd60
commit f3566b1
Show file tree

Hide file tree

Showing 2 changed files with 12 additions and 2 deletions.
diff --git a/src/Inline/Parser/EmphasisParser.php b/src/Inline/Parser/EmphasisParser.php
@@ -18,6 +18,7 @@
 use League\CommonMark\Delimiter\Delimiter;
 use League\CommonMark\InlineParserContext;
 use League\CommonMark\Inline\Element\Text;
+use League\CommonMark\Util\RegexHelper;
 
 class EmphasisParser extends AbstractInlineParser
 {
@@ -55,8 +56,16 @@ public function parse(ContextInterface $context, InlineParserContext $inlineCont
 
         $charAfter = $cursor->getCharacter() ?: "\n";
 
-        $canOpen = $numDelims > 0 && !preg_match('/\s/', $charAfter);
-        $canClose = $numDelims > 0 && !preg_match('/\s/', $charBefore);
+        $canOpen = $numDelims > 0 && !preg_match('/\pZ|\s/u', $charAfter) &&
+            !(preg_match(RegexHelper::REGEX_PUNCTUATION, $charAfter) &&
+            !preg_match('/\pZ|\s/u', $charBefore) &&
+            !(preg_match(RegexHelper::REGEX_PUNCTUATION, $charBefore)));
+
+        $canClose = $numDelims > 0 && !preg_match('/\pZ|\s/u', $charBefore) &&
+            !(preg_match(RegexHelper::REGEX_PUNCTUATION, $charBefore) &&
+            !preg_match('/\pZ|\s/u', $charAfter) &&
+            !(preg_match(RegexHelper::REGEX_PUNCTUATION, $charAfter)));
+
         if ($character === '_') {
             $canOpen = $canOpen && !preg_match('/[a-z0-9]/i', $charBefore);
             $canClose = $canClose && !preg_match('/[a-z0-9]/i', $charAfter);

diff --git a/src/Util/RegexHelper.php b/src/Util/RegexHelper.php
@@ -51,6 +51,7 @@ class RegexHelper
 
     const REGEX_ESCAPABLE = '[!"#$%&\'()*+,.\/:;<=>?@[\\\\\]^_`{|}~-]';
     const REGEX_ENTITY = '/&(?:#x[a-f0-9]{1,8}|#[0-9]{1,8}|[a-z][a-z0-9]{1,31});/i';
+    const REGEX_PUNCTUATION = '/^[\x{2000}-\x{206F}\x{2E00}-\x{2E7F}\\\\\'!"#\$%&\(\)\*\+,\-\.\\/:;<=>\?@\[\]\^_`\{\|\}~]/u';
 
     protected $regex = array();