Is your feature request related to a problem? Please describe.
In Storage, a list objects operation returns a response body with both a list of items and prefixes. The list of items are objects ordered lexicographically by name, whereas the prefixes operate like a directory listing.

The storage objects.list API allows a query paramaeter maxResults that represents the maximum combined number of entries in items[] and prefixes[] to return in a single page of responses.

While the Iterator class supports maxResults, it only accounts for the maximum number of items.

Describe the solution you'd like
Is there a way to have HTTPIterator.max_results and HTTPIterator.num_results account for results other than items from the response body? Currently, only items are kept count for num_results and consequently _has_next_page()

    def _items_iter(self):
        """Iterator for each item returned."""
        for page in self._page_iter(increment=False):
            for item in page:
                self.num_results += 1
                yield item

GCS objects.list API call returns a response body with the following structure:

{
  "kind": "storage#objects",
  "nextPageToken": string,
  "prefixes": [
    string
  ],
  "items": [
    objects Resource
  ]
}

Additional context
Customers are concerned running into memory issue when all prefixes are returned and max_results doesn't take any effect. Internal b/401531977

it = client.list_blobs(
    bucket_name,
    delimiter='/',
    prefix="some_prefix_pattern",
    max_results=100,
)
list(it)
print(len(it.prefixes)) # All 755 prefixes are returned regardless setting max_results=100