In the previous articles, we discussed C# 8 async streams and pattern matching. In this article, we’ll look at C# 8 default interface methods.
Extending interfaces
Before C# 8, it was not possible to add members to an interface without breaking the classes that implement the interface. Because interface members were abstract, classes needed to provide an implementation. C# 8 allows us to extend an interface and provide a default implementation. The runtime (which also needs to support this feature) uses the default implementation when the class does not provide it:
interface IOutput { void PrintMessage(string message); void PrintException(Exception exception) => PrintMessage($"Exception: {exception}"); } class ConsoleOutput : IOutput { public void PrintMessage(string message) => Console.WriteLine(message); }
In this example, ConsoleOutput
does not provide an implementation for PrintException
. When PrintException
is called against a ConsoleOutput
instance, the default method from the IOutput
interface will be called. ConsoleOutput
might provide its own implementation.
A derived interface can provide a more appropriate default implementation by explicitly implementing the base member:
interface IA { void M() { WriteLine("IA.M"); } } interface IB : IA { void IA.M() { WriteLine("IB.M"); } }
We explicitly implement the base member by including IA.
in the name. Without IA.
, the compiler would warn us to either make it explicit, or use the new
keyword if we want to hide it.
C# 8 allows us to add static
members in the interface that can be used by the default interface members:
interface IOutput { private static string s_exceptionPrefix = "Exception"; public static string ExceptionPrefix { get => s_exceptionPrefix; set => s_exceptionPrefix = value; } void PrintMessage(string message); sealed void PrintException(Exception exception) => PrintMessage($"{s_exceptionPrefix}: {exception}"); }
This example shows a private static
field and a public static
method to implement the ExceptionPrefix
property that's used by the sealed PrintException
method. By adding the sealed
keyword, this method can no longer be overridden.
Code inheritance
C# supports inheriting from a single-base class and implementing multiple interfaces. Until C# 8, only the base class could provide code that is usable by the derived class. With C# 8, interfaces can provide code to their implementing classes.
In addition to that enhancement, we can use access modifiers on members and provide static members:
interface IOutput { sealed void PrintException(Exception exception) => PrintMessageCore($"Exception: {exception}"); protected void PrintMessageCore(string message); protected static void PrintToConsole(string message) => Console.WriteLine(message); } class ConsoleOutput : IOutput { void IOutput.PrintMessageCore(string message) { IOutput.PrintToConsole(message); } }
In this example, we see that the IOutput
interface delegates the PrintMessageCore
implementation to the derived class and provides a sealed implementation of PrintException
that makes use of PrintMessageCore
. The example also shows how the static protected PrintToConsole
method can be called from a derived type.
This setup allows us to include the code that is common with the interface, allowing different classes to share this code without a common base class and thus enabling multiple inheritance and trait patterns.
C# 8 allows sharing code, but interfaces are still not allowed to have instance fields (state). Interface methods that require state either need to be abstract (so they are implemented in the class), or they need to accept the state as an argument (provided by the caller, like the implementing class). This approach avoids the inheritance diamond problem for state. For code, the diamond problem is solved by the compiler requiring additional members in ambiguous cases (which can then call the appropriate base).
Conclusion
In this article, we looked at C# 8 default interface methods. Default interface methods provide a way to extend interfaces with new members without breaking previous implementers. The new features also allow code to be shared between types, which enables multiple inheritance and trait patterns. In the next article, we will look at nullable reference types.
C# 8 can be used with the .NET Core 3.1 SDK, which is available on Red Hat Enterprise Linux, Fedora, Windows, macOS, and other Linux distributions.
Last updated: February 22, 2024