added py::ellipsis() method for slicing of multidimensional NumPy arrays

This PR adds a new py::ellipsis() method which can be used in
conjunction with NumPy's generalized slicing support. For instance,
the following is now valid (where "a" is a NumPy array):

py::array b = a[py::make_tuple(0, py::ellipsis(), 0)];
diff --git a/tests/test_numpy_array.cpp b/tests/test_numpy_array.cpp
index 79a157e..5702594 100644
--- a/tests/test_numpy_array.cpp
+++ b/tests/test_numpy_array.cpp
@@ -295,4 +295,10 @@
         std::fill(a.mutable_data(), a.mutable_data() + a.size(), 42.);
         return a;
     });
+
+#if PY_MAJOR_VERSION >= 3
+        sm.def("index_using_ellipsis", [](py::array a) {
+            return a[py::make_tuple(0, py::ellipsis(), 0)];
+        });
+#endif
 }
diff --git a/tests/test_numpy_array.py b/tests/test_numpy_array.py
index 1e83135..8ac0e66 100644
--- a/tests/test_numpy_array.py
+++ b/tests/test_numpy_array.py
@@ -408,3 +408,9 @@
     a = m.create_and_resize(2)
     assert(a.size == 4)
     assert(np.all(a == 42.))
+
+
+@pytest.unsupported_on_py2
+def test_index_using_ellipsis():
+    a = m.index_using_ellipsis(np.zeros((5, 6, 7)))
+    assert a.shape == (6,)