constructor base call without empty curly brackets

[marcosikkens]

When writing C# constructors with a base call the following code is often created:

public class TestB : TestA
{
 
      public TestB(TestA a): base(a)
      {
          //nothing happens here
      }
 
}

public class TestA
{
   public TestA(TestA a)
   {
       //do something with the members ie:
       this.Something = a.Something;
   }
} 

The main thing about this is that sometimes the curly brackets in TestB do not contain any code. it would be nice to be able to just remove those. Like in this example:

public class TestB : TestA
{
      public TestB(TestA a): base(a)
 }

An alternative I currently see used a lot in this case is the following but it just adds a bit of clutter to the end of the line in my opinion:

public class TestB : TestA
{
 
      public TestB(TestA a): base(a)   {  }
 
}

If you have any questions (or if it isn't clear) please do not hesitate to ask.

[Phantonia]

To help the parser, this would only be done by completing the constructor declaration with a semicolon, like so:

public TestB(TestA a) : base(a);

And does this really save much over

public TestB(TestA a) : base(a) { }

Additionally, ending a method declaration in a semicolon has a clear meaning in C#. It is done for abstract, partial or extern methods, all of which do have a body, it is just not specified at the point where the method is declared.
Using a semicolon to end an empty "method" is not really clear, because it implies a body coming from somewhere else, whereas actually it has a body, it is just empty. The curly brackets clearly communicate that the body is empty.

[KieranDevvs]

Using a semicolon to end an empty "method" is not really clear, because it implies a body coming from somewhere else, whereas actually it has a body, it is just empty. The curly brackets clearly communicate that the body is empty.

I think its perfectly clear. Where else would the body be coming from?

Take away inheritance for a second, if you don't have a constructor, is that unclear? Because it has a constructor, its just empty... Its also not like people can some how screw up their code if they don't understand this feature, as the body is optional. People who are used to creating a blank body can still do so.

[CyrusNajmabadi]

I think its perfectly clear.

I disagree.

if you don't have a constructor, is that unclear?

Yes :)

People who are used to creating a blank body can still do so.

It literally saves one character. That's it. You can just write {} today. It's been supported since C#1.0. It's trivially understood as 'has an empty body'. Why do we need a 1-char version of that?

[KieranDevvs]

if you don't have a constructor, is that unclear?

Yes :)

You've just contradicted your own point...

[CyrusNajmabadi]

I thought you wrote constructor body.

[marcosikkens]

My goal never was to write less characters. I think the expression without the braces is just clearer and causes less clutter.

We can do a similar thing with automatic usings and namespaces now. So it can be seen as a extension to that.

[colejohnson66]

While one could argue that records and primary constructors give precedence to empty method bodies (from their ability to have empty type definitions), I, personally, do not like the idea of empty methods being denoted with a single semicolon. It could also conflict with default interface methods (DIMs) where a ; denotes an abstract method but {} denotes a default implementation that does nothing.

[marcosikkens]

I see your point however, you also do not use {} in primary constructors to indicate an empty body. This way you can also have multiple constructors in a class which you cannot do with a primary constructor.

[nogic1008]

This proposal is useful for classes that use a Primary Constructor on C#12.

public class Sample(int id, string name)
{
    public Sample(): this(0, "");
}

The difference between ; and {} may seem like only one character, but there is a big difference in lint and format.

    // Editor might be written:
    public Sample(): this(0, "") {}

    // With whitespace...?
    public Sample(): this(0, "") { }

    // Or line break...?
    public Sample(): this(0, "")
    {
    }

[KieranDevvs]

Its definitely not inconsistent or confusing in my opinion.

It's confusing to me. There are many cases where we do not require the constructor initializer. But now we would, just to be able to use a semicolon for the constructor body. This is very different from all the other places we allow a semicolon body for other constructs.

I'm not sure I understand your example, why would this feature result in constructor initializers being required in some cases?
To my understanding of what is described by the original poster, this is purely syntactic sugar and there would be zero behavioural changes.

[CyrusNajmabadi]

The request above said:

public Sample(); // This causes compile error because "no body"

They want semicolon here so it can actually error if they leave off the initializer. Something they don't get with a normal block body.

[CyrusNajmabadi]

what is described by the original poster, this is purely syntactic sugar

But it is inconsistently applied. If it represents an empty body, why is it just for constructors?

[marcosikkens]

Thanks for the explanation, so let's just say. We remove the semicolon and allow it for other methods.

What would a good example be if allowing this for other methods? The best thing I can think about are constructors.

[CyrusNajmabadi]

Any method that ends up having an empty body because it doesn't need to do anything