Part1 throws an error while part2 does not, why?
//part 1
Object[] arr = new Object[10000000];
Object[] arr1 = arr;
for(int i = 0; i < 100; i++) {
arr1[0] = new Object[10000000];
arr1 = (Object[]) arr1[0];
}
//part 2
Object[] arr1 = new Object[10000000];
for(int i = 0; i < 100; i++) {
arr1[0] = new Object[10000000];
arr1 = (Object[]) arr1[0];
}
In your first snippet, you're retaining a reference to your original huge array, and therefore to every array that you create subsequently.
In your second snippet, you're throwing away the reference to the original huge array - in fact, you're throwing away references to each array except the one you've created, in every iteration of the loop. So that allows all but one of them to be garbage collected, and you only need enough heap space for two of the arrays at any point in the loop.
because in second case you don't have live references for all Object[]
