Improve regex parsing to skip empty attribute and handle html encoded…

… separators to avoid an extra call to HtmlDecode

src/Html2OpenXml/Collections/HtmlAttributeCollection.cs

@@ -20,8 +20,7 @@ namespace HtmlToOpenXml;
 /// </summary>
 sealed class HtmlAttributeCollection
 {
-    private static readonly Regex stripStyleAttributesRegex = new(@"(?<name>.+?):\s*(?<val>[^;]+);*\s*");
-
+    private static readonly Regex stripStyleAttributesRegex = new(@"(?<name>[^;\s]+)\s?(&\#58;|:)\s?(?<val>[^;&]+)\s?(;|&\#59;)*");
     private readonly Dictionary<string, string> attributes = [];
 
 
@@ -37,13 +36,7 @@ public static HtmlAttributeCollection ParseStyle(string? htmlTag)
 
         // Encoded ':' and ';' characters are valid for browser but not handled by the regex (bug #13812 reported by robin391)
         // ex= <span style="text-decoration&#58;underline&#59;color:red">
-        MatchCollection matches = stripStyleAttributesRegex.Matches(
-#if NET5_0_OR_GREATER
-            System.Web.HttpUtility.HtmlDecode(htmlTag)
-#else 
-            HttpUtility.HtmlDecode(htmlTag)
-#endif
-        );
+        MatchCollection matches = stripStyleAttributesRegex.Matches(htmlTag);
         foreach (Match m in matches)
             collection.attributes[m.Groups["name"].Value] = m.Groups["val"].Value;
 

test/HtmlToOpenXml.Tests/StyleTests.cs

@@ -165,5 +165,21 @@ public void DuplicateStyle_ReturnsLatter()
             var styleAttributes = HtmlAttributeCollection.ParseStyle("color:red;color:blue");
             Assert.That(styleAttributes["color"], Is.EqualTo("blue"));
         }
+
+        [Test(Description = "Encoded ':' and ';' characters are valid")]
+        public void EncodedStyle_ShouldSucceed()
+        {
+            var styleAttributes = HtmlAttributeCollection.ParseStyle("text-decoration:underline&#59;color:red");
+            Assert.That(styleAttributes["text-decoration"], Is.EqualTo("underline"));
+            Assert.That(styleAttributes["color"], Is.EqualTo("red"));
+        }
+
+        [Test(Description = "Key style with no value should be ignored")]
+        public void EmptyStyle_ShouldBeIgnoredd()
+        {
+            var styleAttributes = HtmlAttributeCollection.ParseStyle("text-decoration;color:red");
+            Assert.That(styleAttributes["text-decoration"], Is.Null);
+            Assert.That(styleAttributes["color"], Is.EqualTo("red"));
+        }
     }
 }