Fix for #empty? responding correctly to empty responses

[pye2k]

Pull request for #35.

[coveralls]

Coverage remained the same at 97.638% when pulling e669bdb on pye2k:master into 7868f8b on aws:master.

[awood45]

I'm checking this against some integration test cases I have now. Pagination in the query/scan operations could produce new edge cases to test for.

[awood45]

Actually, I'm pretty sure for a couple reasons that the test case I have in mind is both impossible (service won't do it) and already covered, but let's add this to make sure we handle it gracefully, including going forward. Test case is like so:

it "is not empty" do
  stub_client.stub_responses(:scan, truncated_resp, resp_empty)
  c = ItemCollection.new(
    :scan,
    { table_name: "TestTable" },
      model,
      stub_client
  )
  expect(c.empty?).to be_falsy
end

You'll need to expand the scope of where truncated_resp is available to test this. This ensures that if we aren't only checking the last page. With the current implementation, I don't think we even get to the second client call.

[awood45]

Actually, I'll just add that before pushing a release. Thanks!

[pye2k]

👍

[awood45]

I added additional support for an edge case related to empty pages followed by pages with content, which is possible in scan operations.

I also added documentation warnings, given that this can easily do full table scans.