关于“Effective Java Item#29 Favor Generic Types”中解决方案#1“堆污染”的问题

Question

I am reading Effective Java Item#29 Favor Generic Types.

I'm confused about heap pollution that this item mentioned when it talked about the reason to choose solution#1 or solution#2.

Both techniques for eliminating the generic array creation have their adherents. The first is more readable: the array is declared to be of type E[], clearly indicating that it contains only E instances. It is also more concise: in a typical generic class, you read from the array at many points in the code; the first technique requires only a single cast (where the array is created), while the second requires a separate cast each time an array element is read. Thus, the first technique is preferable and more commonly used in practice. It does, however, cause heap pollution (Item 32): the runtime type of the array does not match its compile-time type (unless E happens to be Object). This makes some programmers sufficiently queasy that they opt for the second technique, though the heap pollution is harmless in this situation.

In my understanding, that means some programmers might concern about "heap pullution" at solution#1, so they might choose solution#2. Is my understanind correcyt? Could Java expert help me to figure out why heap pollution might happen at solution#1 or solution#2?

Original Code without Generic

// Object-based collection - a prime candidate for generics
public class Stack {
    private Object[] elements;
    private int size = 0;
    private static final int DEFAULT_INITIAL_CAPACITY = 16;

    public Stack() {
        elements = new Object[DEFAULT_INITIAL_CAPACITY];
    }

    public void push(Object e) {
        ensureCapacity();
        elements[size++] = e;
    }

    public Object pop() {
        if (size == 0)
            throw new EmptyStackException();
        Object result = elements[--size];
        elements[size] = null; // Eliminate obsolete reference
        return result;
    }

    public boolean isEmpty() {
        return size == 0;
    }

    private void ensureCapacity() {
        if (elements.length == size)
            elements = Arrays.copyOf(elements, 2 * size + 1);
    }
}

solution#1 for Generic

public class Stack<E> {
    private E[] elements;
    private int size = 0;
    private static final int DEFAULT_INITIAL_CAPACITY = 16;

    public Stack() {
        elements = (E[]) new Object[DEFAULT_INITIAL_CAPACITY];
    }

    public void push(E e) {
        ensureCapacity();
        elements[size++] = e;
    }

    public E pop() {
        if (size == 0)
            throw new EmptyStackException();
        E result = elements[--size];
        elements[size] = null; // Eliminate obsolete reference
        return result;
    }
    ... // no changes in isEmpty or ensureCapacity
}

solution#2 for Generic

public class Stack<E> {
    private Object[] elements;
    private int size = 0;
    private static final int DEFAULT_INITIAL_CAPACITY = 16;

    public Stack() {
        elements = new Object[DEFAULT_INITIAL_CAPACITY]; 
    }

    public void push(E e) {
        ensureCapacity();
        elements[size++] = e;
    }

    public E pop() {
        if (size == 0)
            throw new EmptyStackException();
        E result = (E) elements[--size];
        elements[size] = null; // Eliminate obsolete reference
        return result;
    }
    ... // no changes in isEmpty or ensureCapacity
}

Answer 1

In solution #2 you have to cast in pop method what is not recommended. Your code is not type-safety. If methods are more complicated it can generate bugs. You don't have any benefits because of it.

Another similar example is method get. You have to cast also:

public E get(int n) {
    return (E) elements[n];
}