为什么不能nullables被声明为const？ [英] Why can't nullables be declared const?

问题描述

[TestClass]
public class MsProjectIntegration {
    const int? projectID = null;
    // The type 'int?' cannot be declared const
    // ...
}

为什么我不能有一个 const int的？？

编辑：我想一个可空INT作为一个const的原因是因为我只是用它从数据库加载一些示例数据。如果是空，我只是要在运行时初始化样本数据。这是一个非常快速的测试项目，很明显，我可以用0或-1，但 INT？只是觉得像我想要做正确的数据结构。只读似乎是要走的路。

The reason I wanted a nullable int as a const is because I'm just using it for loading some sample data from a database. If it's null I was just going to initialize sample data at runtime. It's a really quick test project and obviously I could use 0 or -1 but int? just felt like the right data structure for what I wanted to do. readonly seems like the way to go

推荐答案

这不仅仅是nullables;仅内置于运行时类型可以声明常量（从内存中，它的bool，各类整型，花车/双打和字符串）。

It's not just nullables; only types built into the runtime can be declared const (from memory, it's bools, the various types of int, floats/doubles, and strings).

为什么呢？因为该值被直接嵌入到在编译时组装，而且也没有办法嵌入用户定义类型。

Why? Because the value gets embedded directly into the assembly at compile time, and there's no way to embed user-defined types.

只读关键字应该做你所需要的，但是。通过与常量相反，任何只读字段将会在运行时初始化而不是编译的时候，这样他们就可以用更多的初始化以下任何你想表达

The readonly keyword should do what you need, however. By contrast with const, any readonly fields get initialized at runtime rather than compile time, so they can be initialized with more or less any expression you want.

编辑：作为埃里克利珀指出的，它不是这个简单。例如，常量十进制和

as Eric Lippert points out, it's not this straightforward. For instance, const decimal works.

这：

private const decimal TheAnswer = 42;

...编译（当然，反光镜），以这样的：

...compiles (well, Reflectors) to this:

[DecimalConstant(0, 0, (uint) 0, (uint) 0, (uint) 42)]
private static readonly decimal TheAnswer;

这篇关于为什么不能nullables被声明为const？的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！