Fix emphasis with HTML entities (#97)

xoofx · Feb 24, 2017 · b47ae8d · b47ae8d
1 parent a926e1a
commit b47ae8d
Show file tree

Hide file tree

Showing 3 changed files with 41 additions and 1 deletion.
diff --git a/src/Markdig.Tests/Specs/EmphasisExtraSpecs.md b/src/Markdig.Tests/Specs/EmphasisExtraSpecs.md
@@ -52,3 +52,16 @@ Marked text can be used to specify that a text has been marked in a document.  T
 .
 <p><mark>Marked text</mark></p>
 ````````````````````````````````
+## Emphasis on Html Entities
+
+
+```````````````````````````````` example
+This is text MyBrand ^&reg;^ and MyTrademark ^&trade;^
+This is text MyBrand^&reg;^ and MyTrademark^&trade;^
+This is text MyBrand~&reg;~ and MyCopyright^&copy;^
+.
+<p>This is text MyBrand <sup>®</sup> and MyTrademark <sup>TM</sup>
+This is text MyBrand<sup>®</sup> and MyTrademark<sup>TM</sup>
+This is text MyBrand<sub>®</sub> and MyCopyright<sup>©</sup></p>
+````````````````````````````````
+
diff --git a/src/Markdig.Tests/Specs/Specs.cs b/src/Markdig.Tests/Specs/Specs.cs
@@ -17616,6 +17616,30 @@ public void Example005()
             Console.WriteLine("Example {0}" + Environment.NewLine + "Section: {0}" + Environment.NewLine, 5, "Extensions Marked");
 			TestParser.TestSpec("==Marked text==", "<p><mark>Marked text</mark></p>", "emphasisextras|advanced");
         }
+    }
+        // ## Emphasis on Html Entities
+    [TestFixture]
+    public partial class TestExtensionsEmphasisonHtmlEntities
+    {
+        [Test]
+        public void Example006()
+        {
+            // Example 6
+            // Section: Extensions Emphasis on Html Entities
+            //
+            // The following CommonMark:
+            //     This is text MyBrand ^&reg;^ and MyTrademark ^&trade;^
+            //     This is text MyBrand^&reg;^ and MyTrademark^&trade;^
+            //     This is text MyBrand~&reg;~ and MyCopyright^&copy;^
+            //
+            // Should be rendered as:
+            //     <p>This is text MyBrand <sup>®</sup> and MyTrademark <sup>TM</sup>
+            //     This is text MyBrand<sup>®</sup> and MyTrademark<sup>TM</sup>
+            //     This is text MyBrand<sub>®</sub> and MyCopyright<sup>©</sup></p>
+
+            Console.WriteLine("Example {0}" + Environment.NewLine + "Section: {0}" + Environment.NewLine, 6, "Extensions Emphasis on Html Entities");
+			TestParser.TestSpec("This is text MyBrand ^&reg;^ and MyTrademark ^&trade;^\nThis is text MyBrand^&reg;^ and MyTrademark^&trade;^\nThis is text MyBrand~&reg;~ and MyCopyright^&copy;^", "<p>This is text MyBrand <sup>®</sup> and MyTrademark <sup>TM</sup>\nThis is text MyBrand<sup>®</sup> and MyTrademark<sup>TM</sup>\nThis is text MyBrand<sub>®</sub> and MyCopyright<sup>©</sup></p>", "emphasisextras|advanced");
+        }
     }
         // # Extensions
         //

diff --git a/src/Markdig/Helpers/CharHelper.cs b/src/Markdig/Helpers/CharHelper.cs
@@ -162,6 +162,9 @@ public static bool IsWhiteSpaceOrZero(this char c)
             return IsWhitespace(c) || IsZero(c);
         }
 
+        // Note that we are not considering the character & as a punctuation in HTML
+        // as it is used for HTML entities, print unicode, so we assume that when we have a `&` 
+        // it is more likely followed by a valid HTML Entity that represents a non punctuation
         public static void CheckUnicodeCategory(this char c, out bool space, out bool punctuation)
         {
             // Credits: code from CommonMark.NET
@@ -170,7 +173,7 @@ public static void CheckUnicodeCategory(this char c, out bool space, out bool pu
             if (c <= 'ÿ')
             {
                 space = c == '\0' || c == ' ' || (c >= '\t' && c <= '\r') || c == '\u00a0' || c == '\u0085';
-                punctuation = c == '\0' || (c >= 33 && c <= 47) || (c >= 58 && c <= 64) || (c >= 91 && c <= 96) || (c >= 123 && c <= 126);
+                punctuation = c == '\0' || (c >= 33 && c <= 47 && c != 38) || (c >= 58 && c <= 64) || (c >= 91 && c <= 96) || (c >= 123 && c <= 126);
             }
             else
             {