为什么泛型数组中的setter方法可以工作（JAVA）？

Question

class Demo<T>{
    T d[]=(T[])new Object[3];
    T x;
    int l=0;
    void setD(T x){
        d[l++]=x; //Why does this work?
    }
    void getD(int x){
        System.out.println(d[x]); //Why does this work too? as i cant print in main class directly
    }

}

public class Main<T> {

    public static void main(String[] args) {

        Demo<Integer> de=new Demo<>();
        de.x=7; //This doesnt throw any error

        //de.d[0]=3; //This throws runtime error. Runtime error is given below.. I understand why this doesn't work 
//but i wanted to know why does that setter method work as this line of code seems similar to that of setter method
        de.setD(3);

        de.getD(0);

    }
}

Answer 1

我想那是因为字型擦除。看这条线。

Demo<Integer> de=new Demo<>();
de.d[0]=3; //java knows, that de.d should be Integer[], but it actually is Object[]

这里，由于指定了generic参数，java希望de.d是integer[]。但它实际上是object[]。java试图将对象数组视为整数数组。

为此，java必须将Object[]强制转换为Integer[]。它会引发异常。如果执行以下操作，将引发相同的异常

Integer[] ints = (Integer[])new Object[10];

T d[]=(T[])new Object[3];
void setD(T x){
    d[l++]=x; //Why does this work?
}

Object d[]= (Object[])new Object[3];
void setD(Object x){
    d[l++]=x; // x is Integer here
}