Merge pull request #13 from askonomm/12-html-entity-usage-results-in-…

…an-exception Convert HTML entities to Unicode to avoid compatibility issues with X…
askonomm · Oct 21, 2024 · 350d81e · 350d81e
2 parents a36107b + 60da58b
commit 350d81e
Show file tree

Hide file tree

Showing 2 changed files with 32 additions and 0 deletions.
diff --git a/Htmt/Parser.cs b/Htmt/Parser.cs
@@ -43,6 +43,7 @@ private void Parse()
 
             RemoveDoctype();
             CloseVoidElements();
+            TransformHtmlEntities();
         }
 
         var templateStr = $"<root xmlns:x=\"{HtmtNamespace}\">{Template}</root>";
@@ -137,6 +138,28 @@ private void CloseVoidElements()
             Template = Template.Replace(element, newElement);
         }
     }
+
+    [GeneratedRegex(@"&(?<entity>\w+);")]
+    private static partial Regex HtmlEntityRegex();
+
+    /**
+     * Transforms HTML entities to their respective characters.
+     *
+     * This is necessary because the XML parser does not handle HTML entities.
+     */
+    private void TransformHtmlEntities()
+    {
+        var entityRegex = HtmlEntityRegex();
+        var matches = entityRegex.Matches(Template);
+
+        foreach (Match match in matches)
+        {
+            var entity = match.Groups["entity"].Value;
+            var replacement = System.Net.WebUtility.HtmlDecode($"&{entity};");
+
+            Template = Template.Replace(match.Value, replacement);
+        }
+    }
 
     /**
      * Parses the template and returns it as HTML.

diff --git a/HtmtTests/ParserTest.cs b/HtmtTests/ParserTest.cs
@@ -56,4 +56,13 @@ public void TestVoidElButOpen()
 
         Assert.AreEqual("<html><head></head><body><img src=\"asd\" /><br /><hr /></body></html>", parser.ToHtml());
     }
+
+    [TestMethod]
+    public void TestHtmlEntities()
+    {
+        const string template = "<html><head></head><body>&lt;div&gt;Hello, World!&lt;/div&gt;</body></html>";
+        var parser = new Htmt.Parser { Template = template };
+
+        Assert.AreEqual("<html><head></head><body><div>Hello, World!</div></body></html>", parser.ToHtml());
+    }
 }