为什么f(double)比f(long, int=0)更适合f(long)

本文关键字：long int 为什么 double | 更新日期: 2023-09-27 17:50:56

我的一个同事发现了这种奇怪的情况。我在这里贴一个简单的例子:

using System;
namespace Test
{
    public class F
    {
        public void f(double d) { Console.WriteLine("public void F.f(double d)"); }
        public virtual void f(long l, int q = 0) { Console.WriteLine("public virtual void F.f(long l, int q = 0)"); }
    }
    public class FDerived : F
    {
        public override void f(long l, int q) { Console.WriteLine("public override FDerived.f(long l, int q)"); }
        public void g() { f(2L); }
    }
    public class G : FDerived
    {
        public void h1() { F fd = new FDerived(); fd.f(2L); }
        public void h2() { FDerived fd = new FDerived(); fd.f(2L); }
    }
    class Program
    {
        static void Main(string[] args)
        {
            new FDerived().g();
            new G().h1();
            new G().h2();
            Console.ReadLine();
        }
    }
}

示例的输出是:

public void F.f(double d)
public override FDerived.f(long l, int q)
public void F.f(double d)

我不明白这有什么意义。

为什么f(double)比f(long, int=0)更适合f(long)?为什么它取决于'fd'的类型?!

为什么f(double)比f(long, int=0)更适合f(long)

我这里没有c#规范，但是可选的参数值不会被重写的方法继承。

尝试更改FDerived 

public override void f(long l, int q)


public override void f(long l, int q = 0)

，它将按预期工作。

请注意，这可以在一个简单的示例中显示:

public class F
{
    public virtual void f(long l, int q = 0) { Console.WriteLine("public virtual void F.f(long l, int q = 0)"); }
}
public class FDerived : F
{
    public override void f(long l, int q) { Console.WriteLine("public override FDerived.f(long l, int q)"); }
}
// Doesn't compile: No overload for method 'f' takes 1 arguments
new FDerived().f(5L);

显然这是编译:

new F().f(5L);

甚至这个

F obj = new FDerived();
obj.f(5L);

(这个将输出public override FDerived.f(long l, int q))

相关文章：