working on fixing scheme normalization

Aaronontheweb · Jan 3, 2025 · c51aafb · c51aafb
1 parent 4c00ffb
commit c51aafb
Show file tree

Hide file tree

Showing 2 changed files with 73 additions and 2 deletions.
diff --git a/src/LinkValidator.Tests/ParseHelperSpecs.cs b/src/LinkValidator.Tests/ParseHelperSpecs.cs
@@ -1,6 +1,72 @@
-namespace LinkValidator.Tests;
+using FluentAssertions;
+using LinkValidator.Actors;
+using LinkValidator.Util;
+
+namespace LinkValidator.Tests;
 
 public class ParseHelperSpecs
 {
+    // create a string that contains HTML linking to a few different URLs using relative links
+    private const string RelativeHtml = """
+                                        
+                                                <html>
+                                                    <head>
+                                                        <title>Test Page</title>
+                                                    </head>
+                                                    <body>
+                                                        <a href="/about">About</a>
+                                                        <a href="/contact">Contact</a>
+                                                        <a href="/faq">FAQ</a>
+                                                    </body>
+                                                </html>
+                                        """;
+
+    [Fact]
+    public void ParseHelper_should_return_absolute_uris()
+    {
+        // Arrange
+        var uri = new AbsoluteUri(new Uri("http://example.com"));
+
+        // Act
+        var uris = ParseHelpers.ParseLinks(RelativeHtml, uri);
+
+        // Assert
+        uris.Should().HaveCount(3);
+        uris.Should().Contain(new AbsoluteUri(new Uri("http://example.com/about")));
+        uris.Should().Contain(new AbsoluteUri(new Uri("http://example.com/contact")));
+        uris.Should().Contain(new AbsoluteUri(new Uri("http://example.com/faq")));
+    }
+
+    // create a string that contains HTML linking to a few different URLs using absolute links
+    private const string AbsoluteHtml = """
+                                        
+                                                <html>
+                                                    <head>
+                                                        <title>Test Page</title>
+                                                    </head>
+                                                    <body>
+                                                        <a href="http://example.com/about">About</a>
+                                                        <a href="http://example.com/contact">Contact</a>
+                                                        <a href="http://example.com/faq">FAQ</a>
+                                                    </body>
+                                                </html>
+                                        """;
 
+    [Fact]
+    public void ParseHelper_should_return_absolute_uris_when_given_absolute_links()
+    {
+        // Arrange
+        var uri = new AbsoluteUri(new Uri("https://example.com"));
+
+        // Act
+        var uris = ParseHelpers.ParseLinks(AbsoluteHtml, uri);
+
+        // Assert
+        uris.Should().HaveCount(3);
+
+        // notice that we convert the scheme to https
+        uris.Should().Contain(new AbsoluteUri(new Uri("https://example.com/about")));
+        uris.Should().Contain(new AbsoluteUri(new Uri("https://example.com/contact")));
+        uris.Should().Contain(new AbsoluteUri(new Uri("https://example.com/faq")));
+    }
 }
diff --git a/src/LinkValidator.Tests/UriHelperSpecs.cs b/src/LinkValidator.Tests/UriHelperSpecs.cs
@@ -63,7 +63,12 @@ public void AbsoluteUriIsInDomain_should_NOT_return_true_when_host_is_diffrent(s
     {
         { new AbsoluteUri(new Uri("http://example.com")), "http://example.com/some/path", new AbsoluteUri(new Uri("http://example.com/some/path")) },
         { new AbsoluteUri(new Uri("http://example.com")), "/some/path", new AbsoluteUri(new Uri("http://example.com/some/path")) },
-        { new AbsoluteUri(new Uri("http://example.com")), "some/path", new AbsoluteUri(new Uri("http://example.com/some/path")) }
+        { new AbsoluteUri(new Uri("http://example.com")), "some/path", new AbsoluteUri(new Uri("http://example.com/some/path")) },
+        { new AbsoluteUri(new Uri("https://example.com")), "some/path", new AbsoluteUri(new Uri("https://example.com/some/path")) },
+
+        // HTTP vs. HTTPS - should normalize to use whatever the baseUri uses
+        { new AbsoluteUri(new Uri("https://example.com")), "http://example.com/some/path", new AbsoluteUri(new Uri("https://example.com/some/path")) },
+        { new AbsoluteUri(new Uri("http://example.com")), "https://example.com/some/path", new AbsoluteUri(new Uri("http://example.com/some/path")) }
     };
 
     [Theory]