Return type consistency #17

Currently, this is how the queue functions when it is empty:

assert self.queue.peek(items=0) == []
assert self.queue.peek(items=1) is None
assert self.queue.peek(items=2) == []
assert self.queue.peek(items=100) == []

assert self.queue.get(items=0, block=False) == []
with pytest.raises(queue.Empty):
    self.queue.get(items=1, block=False)

When the queue is has data in it, the following is true:

assert isinstance(self.queue.get(items=0), list)
assert isinstance(self.queue.get(items=1), int)
assert isinstance(self.queue.get(items=2), list)

assert isinstance(self.queue.peek(items=0), list)
assert isinstance(self.queue.peek(items=1), int)
assert isinstance(self.queue.peek(items=2), list)

It seems like the return values are slightly inconsistent. It makes sense that when peeking and items == 1 a value is returned or None if there are no items in the queue, and when items > 1 a list is returned. What should be the behavior for items == 0?

I propose that when items == 0, None is returned for both peek and get operations, rather than an empty list.

This means that peek can return 3 types of values: None, a value, and a list. get can return 4 types of values: None, a value, a list, and an Empty exception. That seems like a lot of different return types.

@Kriechi do you have any thoughts on this?