TF_ImportGraphDefOptionsAddInputMapping(opts, "scalar", 0, {scalar, 0});
  TF_ImportGraphDefOptionsAddReturnOutput(opts, "feed", 0);
  TF_ImportGraphDefOptionsAddReturnOutput(opts, "scalar", 0);
  EXPECT_EQ(2, TF_ImportGraphDefOptionsNumReturnOutputs(opts));
  TF_ImportGraphDefOptionsAddReturnOperation(opts, "scalar");
  EXPECT_EQ(1, TF_ImportGraphDefOptionsNumReturnOperations(opts));

      TestTensorHandleWithDims<T, datatype>(eager_ctx, data, dims, num_dims);
  *tensor =
      unwrap(TF_CreateAbstractTensorFromEagerTensor(input_eager, status.get()));
  return absl::OkStatus();
}

// Return a scalar tensor handle with given value.
template <class T, TF_DataType datatype>
Status TestScalarTensorHandle(AbstractContext* ctx, const T value,
                              AbstractTensorHandle** tensor) {

    if (dst->dims() != 0) {
      return InvalidArgument(
          "Malformed TF_RESOURCE tensor: expected a scalar, got a tensor with "
          "shape ",
          dst->shape().DebugString());
    }
    *dst = tensorflow::Tensor(tensorflow::DT_RESOURCE, dst->shape());
    if (!dst->scalar<tensorflow::ResourceHandle>()().ParseFromString(
            string(static_cast<const char*>(tensor_interface->Data()),

  }

 public:
  bool UseMlir() const { return strcmp(std::get<0>(GetParam()), "mlir") == 0; }
  bool UseFunction() const { return std::get<2>(GetParam()); }
};

// Checks that inputs[0] is a scalar.
Status TestScalarShape(AbstractContext* ctx,
                       absl::Span<AbstractTensorHandle* const> inputs,
                       absl::Span<AbstractTensorHandle*> outputs) {
  PartialTensorShape shape;

// APIs for sampler buckets
typedef struct TFE_MonitoringBuckets TFE_MonitoringBuckets;
TF_CAPI_EXPORT extern TFE_MonitoringBuckets*
TFE_MonitoringNewExponentialBuckets(double scale, double growth_factor,
                                    int bucket_count);
TF_CAPI_EXPORT extern void TFE_MonitoringDeleteBuckets(
    TFE_MonitoringBuckets* buckets);

// APIs for Sampler without label.

  TFE_OpSetAttrType(fill_op, "T", TF_FLOAT);
  TFE_OpSetAttrType(fill_op, "Tshape", TF_INT32);

  float five = 5.0;
  TFE_TensorHandle* scalar = TestScalarTensorHandle(tfe_context_, five);
  TF_Tensor* scalarTensor = TFE_TensorHandleResolve(scalar, status_);
  CHECK_EQ(TF_OK, TF_GetCode(status_)) << TF_Message(status_);
  CheckOutputShapes(fill_op,

  Run({{func_feed, Int32Tensor(3)}}, func_op, 2 + 3);
  VerifyFDef(
      {"add_0", "scalar"}, M({{"feed1"}, {"feed2"}}), M({{"add"}}),
      {{"feed1", "add_0:0"}, {"feed2", "add_0:1"}, {"add_0:sum:0", "add"}},
      {{"^scalar", "add_0:2"}});
}

TEST_F(CApiFunctionTest, ControlDependencyOutsideOfBody) {
  /*
   *                  |  |    scalar
   *                  |  |    .
   *                  v  v   . <---- control dependency

}

TFE_MonitoringBuckets* TFE_MonitoringNewExponentialBuckets(double scale,
                                                           double growth_factor,
                                                           int bucket_count) {
  return new TFE_MonitoringBuckets([scale, growth_factor, bucket_count]() {
    return tensorflow::monitoring::Buckets::Exponential(scale, growth_factor,

      * If you passed a value to the `loss_scale` argument (the second argument) of `Policy`:

// Abstract interface to a context.
//
// A context is responsible for creating key objects such as Tensors,
// TensorHandles & Operations.
class ImmediateExecutionContext : public AbstractContext {
 public:
  // Optimized scalar creation functions
  virtual AbstractTensorInterface* CreateInt64Scalar(int64_t value) = 0;
  virtual AbstractTensorInterface* CreateUint64Scalar(uint64 value) = 0;