英文:
How do programs know the types at runtimes of primitives statically typed languages
问题
Go和Scala都提供了在运行时检查类型的方法:
Scala:
class A
class B extends A
val a: A = new A // 静态类型: A 动态类型: A
val ab: A = new B // 静态类型: A 动态类型: B
val b: B = new B // 静态类型: B 动态类型: B
def thing(x: Any): String = x match {
    case t: Int => "Int"
    case t: A => "A"
    case t: B => "B"
    case t: String => "String"
    case _ => "unknown type"
}
Go:
package main
import (
    "fmt"
    "reflect"
)
type A struct {
    field1 int
    field2 string
}
func printTypeOf(t interface{}) {
    fmt.Println(reflect.TypeOf(t))
}
func main() {
    i := 234234 // int
    s := "hello world" // string
    p := &A{3234234, "stuff"} // *A
    fmt.Print("The type of i is: ")
    printTypeOf(i)
    fmt.Print("The type of s is: ")
    printTypeOf(s)
    fmt.Print("The type of p is: ") // pass a pointer
    printTypeOf(p)
    fmt.Print("The type of the reference of p is: ") // pass a value
    printTypeOf(*p)
}
这些代码是如何在内部工作的?我猜对于结构体和类,对象的类型存储在一个隐藏字段中(所以在golang中的结构实际上是struct { field1 int field2 string type type })。但是,函数如何得到11010110并知道它是指向内存地址214的指针、整数214还是字符Ö?所有的值是否都以表示其类型的字节进行传递?
英文:
Both Go and Scala provide ways of checking types at runtime:
Scala:
class A
class B extends A
val a: A = new A // static type: A dynamic type: A
val ab: A = new B // static type: A dynamic type: B
val b: B = new B // static type: B dynamic type: B
def thing(x: Any): String = x match {
	case t: Int => "Int"
	case t: A => "A"
	case t: B => "B"
	case t: String => "String"
	case _ => "unknown type"
}
Go:
package main
import (
	"fmt"
	"reflect"
)
struct A {
	field1 int
	field2 string
}
func printTypeOf(t interface{}) {
	fmt.Println(reflect.TypeOf(t))
}
func main() {
	i := 234234 // int
	s := "hello world" // string
	p := &A{3234234, "stuff"} // *A
	fmt.Print("The type of i is: ")
	printTypeOf(i)
	fmt.Print("The type of s is: ")
	printTypeOf(s)
	fmt.Print("The type of p is: ") // pass a pointer
	printTypeOf(p)
	fmt.Print("The type of the reference of p is: ") // pass a value
	printTypeOf(*p)
}
How exactly does this work internally? I presume for structs and classes, the type of the object is stored in a hidden field (so the structure in golang is really struct { field1 int field2 string type type }. But how on earth can function be given 11010110 and know whether it's a pointer to memory address at 214, the integer 214 or the character Ö? Are all values secretly passed with a byte which represents its type?
答案1
得分: 2
在Scala中,每个对象都有一个指向其类的指针。当你使用原始参数(Int,Char等)调用thing时,它会自动装箱为一个对象(java.lang.Integer,java.lang.Character等)。你代码中对Int的匹配实际上被转换为对Integer的匹配。
在Go中,类型不存储在结构体中,而是存储在接口值中:
接口值表示为一个两个字的对,其中一个指向存储在接口中的类型信息的指针,另一个指向关联数据的指针。将b分配给类型为Stringer的接口值会设置接口值的两个字。
因此,当你调用printTypeOf(whatever)时,whatever会被转换为interface{},并且它的类型会与whatever本身一起存储在这个新值中。
英文:
In Scala, each object has a pointer to its class. When you call thing with a primitive argument (Int, Char, etc.), it's automatically boxed to an object (java.lang.Integer, java.lang.Character, etc.). The match against Int in your code is actually translated to a match against Integer.
In Go, the type isn't stored in structs, but in interface values:
So when you call printTypeOf(whatever), whatever is converted to interface {} and its type is stored in this new value along with whatever itself.
通过集体智慧和协作来改善编程学习和解决问题的方式。致力于成为全球开发者共同参与的知识库,让每个人都能够通过互相帮助和分享经验来进步。


评论