Handle cases where sanitiseHtml empty

digirati-co-uk · Nov 16, 2023 · f113a03 · f113a03
1 parent a20ba40
commit f113a03
Show file tree

Hide file tree

Showing 2 changed files with 16 additions and 1 deletion.
diff --git a/src/IIIF/IIIF.Tests/Presentation/HtmlSanitiserTests.cs b/src/IIIF/IIIF.Tests/Presentation/HtmlSanitiserTests.cs
@@ -11,6 +11,19 @@ public class HtmlSanitiserTests
     public void SanitiseHtml_ReturnsGivenString_IfNullOrEmpty(string val)
         => val.SanitiseHtml().Should().Be(val);
 
+    [Theory]
+    [InlineData(true)]
+    [InlineData(false)]
+    public void SanitiseHtml_ReturnsEmptyString_IfGivenInvalidHtml(bool ignoreNonHtml)
+    {
+        const string input = "<div>invalid html</div>";
+        const string expected = "";
+
+        var actual = input.SanitiseHtml(ignoreNonHtml);
+
+        actual.Should().Be(expected);
+    }
+
     [Fact]
     public void SanitiseHtml_Trims_Whitespace_FromBeginningAndEnd_IfIgnoreNonHtmlFalse()
     {

diff --git a/src/IIIF/IIIF/Presentation/HtmlSanitiser.cs b/src/IIIF/IIIF/Presentation/HtmlSanitiser.cs
@@ -72,6 +72,7 @@ public static string SanitiseHtml(this string propertyValue, bool ignoreNonHtml
 
         var workingString = Sanitizer.Sanitize(propertyValue.Trim());
 
+        if (string.IsNullOrEmpty(workingString)) return workingString;
         if (IsHtmlString(workingString)) return workingString;
 
         if (!HtmlSanitizerOptions.AllowedTags.Contains(nonHtmlWrappingTag))
@@ -85,5 +86,6 @@ public static string SanitiseHtml(this string propertyValue, bool ignoreNonHtml
         return Sanitizer.Sanitize(workingString);
     }
 
-    private static bool IsHtmlString(string candidate) => candidate[0] == '<' && candidate[^1] == '>';
+    private static bool IsHtmlString(string candidate) 
+        => candidate[0] == '<' && candidate[^1] == '>';
 }