Uri wonders...or not

It's not the first time the Uri gives me pain. This time I got a bug reported on one of my libraries where parts of the provided URL got replaced. A URL is defined as a two string parts which are then used to construct a URI. Now, I'm most certainly misusing it, or have been putting to much hope to the smart people behind it. Look at the case below. The desired string is: http://mydomain.com:8081/cloud/test

The problem

var b1 = new Uri("http://mydomain.com:8081/cloud");
var b2 = new Uri("http://mydomain.com:8081/cloud/");
            
//How many will match http://mydomain.com:8081/cloud/test
Console.WriteLine(new Uri(b1, "/test"));
Console.WriteLine(new Uri(b1, "test"));
Console.WriteLine(new Uri(b2, "/test"));
Console.WriteLine(new Uri(b2, "test"));

gives:

http://mydomain.com:8081/test
http://mydomain.com:8081/test
http://mydomain.com:8081/test
http://mydomain.com:8081/cloud/test

A hacky solution

Before looking at the answer, this was the "hacky" solution I went with:

public static class UriMagic
{
  public static Uri Abracadabra(params string[] parts)
      => new Uri(string.Join(
          "/",
          parts.Select(p => p.Trim(' ', '/'))));
}

Another "fun" finding... the "OriginalString"

Documentation says:

Gets the original URI string that was passed to the Uri constructor.

Try this:

[Fact]
public void Foo()
{
    new Uri(new Uri("http://foo.com:80/bar/"), "fun")
        .OriginalString
        .Should().Be("http://foo.com:80/bar/fun");
}

[Fact]
public void Bar()
{
    new Uri(new Uri("http://foo.com:8081/bar/"), "fun")
        .OriginalString
        .Should().Be("http://foo.com:8081/bar/fun");
}

[Fact]
public void FooBar()
{
    new Uri("http://foo.com:8081/bar/fun")
        .OriginalString
        .Should().Be("http://foo.com:8081/bar/fun");
}

Outcome?

Summary

I'm clearly using it wrong. But for me, it feels to easy to misuse.

//Daniel

header image src: https://pixabay.com/en/doors-choices-choose-open-decision-1767559/