createEndpointSlice(t, s.KubeClient().Kube(), "slice1", "service", namespace, []v1.EndpointPort{{Name: "http", Port: 80}}, []string{"1.2.3.4"})
	expectEndpoints(t, s, "outbound|80||service.namespace.svc.cluster.local", []string{"1.2.3.4:80"}, nil)

	// create an FQDN endpoint that should be ignored
	createEndpointSliceWithType(t, s.KubeClient().Kube(), "slice1", "service",

// CHECK: %[[SLICE_1:.*]] = "tf.Slice"(%[[SHAPE]], %[[CONST]], %[[CONST_2]]) : (tensor<3xi64>, tensor<1xi32>, tensor<1xi32>) -> tensor<1xi64>
// CHECK: %[[SLICE_2:.*]] = "tf.Slice"(%[[SHAPE]], %[[CONST_2]], %[[CONST_1]]) : (tensor<3xi64>, tensor<1xi32>, tensor<1xi32>) -> tensor<2xi64>
// CHECK: %[[BROADCAST_ARGS:.*]] = "tf.BroadcastArgs"(%[[SLICE_1]], %[[CONST_4]]) : (tensor<1xi64>, tensor<0xi64>) -> tensor<1xi64>

    type_ = self._get_inferred_type(node.value, ty)

    # TODO(fengliuai): Here we hardcode the node.slice here to get the index
    # type. Use the visit method once the type inference is done.
    # slice_val, slice_ty = self.visit(node.slice)
    s = node.slice
    if not isinstance(s, (ast.Tuple, ast.Slice)):
      if isinstance(s, ast.Constant):
        # TODO(fengliuai): promote to an assignment

        vector_one);
    auto slice_op =
        CreateSliceOpForTensorList(loc, /*input_list=*/input,
                                   /*start_index=*/scalar_zero, /*size=*/size,
                                   /*item_rank=*/partial_position_shape,
                                   /*result_type=*/result_type, rewriter);
    rewriter->create<func::ReturnOp>(loc, ArrayRef<Value>({slice_op}));
  }
};

  // CHECK:  %[[SLICE1:.*]] = "tf.Slice"(%arg0, %[[CST1]], %[[CST2]]) : (tensor<3x8x4xi32>, tensor<3xi64>, tensor<3xi64>) -> tensor<3x6x4xi32>
  // CHECK:  %[[CONCAT:.*]] = "tf.ConcatV2"(%[[SLICE]], %[[SLICE1]], %[[CST3]]) : (tensor<3x2x4xi32>, tensor<3x6x4xi32>, tensor<i32>) -> tensor<3x8x4xi32>
  // CHECK:  return %[[CONCAT]] : tensor<3x8x4xi32>