if (!enabled_) {
      return absl::OkStatus();
    }
    if (first_call_) {
      first_call_ = false;
      return absl::OkStatus();
    }
    return tensorflow::errors::Internal("Graph pass runs for more than once!");
  }

 private:
  bool first_call_ = true;
};

// After the graph pass is registered, it takes effect globally and can affect

                "Got a non-parallel tensor ",
                tensorflow::unwrap(absl::get<TFE_TensorHandle*>(input))
                    ->DebugString(),
                " as input to a parallel operation. First pack non-parallel "
                "tensors for each device into a parallel tensor explicitly.")
                .c_str());
        return absl::nullopt;
      }
    } else {

  Define(-1, {}, {feed1, feed2}, {feed2, feed1}, {"first", "second"});

  // Use, run, and verify
  TF_Operation* two = ScalarConst(2, host_graph_, s_);
  TF_Operation* func_feed = Placeholder(host_graph_, s_);
  TF_Operation* func_op = Use({two, func_feed});
  Run({{func_feed, Int32Tensor(3)}}, {{func_op, 0}, {func_op, 1}}, {3, 2});
  VerifyFDef(empty_, M({{"feed1"}, {"feed2"}}), M({{"first"}, {"second"}}),

        tensorflow::unwrap(handles[i]);
    if (tensorflow::CustomDeviceTensorHandle::classof(unwrapped_handle)) {
      // One of the inputs we're trying to pack is on a custom device. We'll let
      // the first custom device we see handle all of the packing.
      auto* custom_device_handle =
          tensorflow::down_cast<tensorflow::CustomDeviceTensorHandle*>(
              unwrapped_handle);

  TFE_Op* func_op = TFE_NewOp(ctx, func.name().data(), status);
  for (const auto& attr : func.attr()) {
    if (!status->status.ok()) return nullptr;
    SetOpAttrValueScalar(ctx, func_op, attr.second, attr.first.data(), status);
    if (!status->status.ok()) return nullptr;
  }
  return func_op;
}
}  // namespace

void TFE_ContextStartStep(TFE_Context* ctx) {
  tensorflow::unwrap(ctx)->StartStep();
}

TEST(RamFileBlockCacheTest, OutOfRange) {
  // Tests reads of a 24-byte file with block size 16.
  const size_t block_size = 16;
  const size_t file_size = 24;
  bool first_block = false;
  bool second_block = false;
  auto fetcher = [block_size, file_size, &first_block, &second_block](
                     const string& filename, size_t offset, size_t n,
                     char* buffer, TF_Status* status) -> int64_t {

    if (!v2_reader_->status().ok()) {
      tsl::Set_TF_Status_from_Status(status, v2_reader_->status());
      return;
    }
    auto result = BuildV2VarMaps();
    var_to_shape_map_.swap(result.first);
    var_to_data_type_map_.swap(result.second);
  } else {
    reader_.reset(new TensorSliceReader(filename));
    if (!reader_->status().ok()) {
      tsl::Set_TF_Status_from_Status(status, reader_->status());

  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);
  auto* arg1 = TF_AddFunctionParameter(graph_ctx, TF_FLOAT, {-1, nullptr}, s);
  ASSERT_EQ(TF_OK, TF_GetCode(s)) << TF_Message(s);

  // Create a first "Add" computing `arg0 + arg1`.
  TF_AbstractTensor* add_output1;
  {
    // Build an abstract operation, inputs and output.
    auto* add_op = TF_NewAbstractOp(graph_ctx);
    TF_AbstractOpSetOpType(add_op, "Add", s);

    TFE_ContextSetExecutorForThread(ctx, executor);

    TFE_DeleteExecutor(executor);
    TFE_DeleteContext(ctx);
  }
  TF_DeleteStatus(status);
}

TEST(CAPI, Function_ident_CPU) {
  // First create a simple identity function.
  TF_Graph* function_graph = TF_NewGraph();
  TF_OperationDescription* arg_descr =
      TF_NewOperation(function_graph, "Placeholder", "arg");
  TF_SetAttrType(arg_descr, "dtype", TF_INT32);

  // Validate NodeDef is what we expect.
  EXPECT_TRUE(IsNeg(node_def, "add"));

  // Serialize to GraphDef.
  GraphDef graph_def2;
  ASSERT_TRUE(GetGraphDef(graph, &graph_def2));

  // Compare with first GraphDef + added NodeDef.
  NodeDef* added_node = graph_def.add_node();
  *added_node = node_def;
  EXPECT_EQ(graph_def.DebugString(), graph_def2.DebugString());

  // Look up some nodes by name.