SINGLE_CASE(kF, TF_ATTR_FLOAT, -1);
    SINGLE_CASE(kB, TF_ATTR_BOOL, -1);
    SINGLE_CASE(kType, TF_ATTR_TYPE, -1);
    SINGLE_CASE(kShape, TF_ATTR_SHAPE,
                attr->shape().unknown_rank() ? -1 : attr->shape().dim_size());
    SINGLE_CASE(kTensor, TF_ATTR_TENSOR, -1);
#undef SINGLE_CASE

    case tensorflow::AttrValue::kList:
      metadata.is_list = 1;
      metadata.list_size = 0;

  int dst_fd = open(dst, O_WRONLY | O_CREAT | O_TRUNC, open_mode);
  if (dst_fd < 0) {
    close(src_fd);
    return -1;
  }

  // Both files have been opened, do the transfer.
  // Since errno would be overridden by `close` below, save it here.
  int error_code = 0;
  if (CopyFileContents(dst_fd, src_fd, size) < 0) error_code = errno;

  close(src_fd);
  close(dst_fd);
  if (error_code != 0) {

}

unsigned char TF_GetXlaAutoJitEnabled() {
  tensorflow::XlaAutoJitFlag flag =
      tensorflow::GetMarkForCompilationPassFlags()->xla_auto_jit_flag;
  return static_cast<unsigned char>(flag.optimization_level_single_gpu > 0 ||
                                    flag.optimization_level_general > 0);
}

unsigned char TF_GetXlaConstantFoldingDisabled() {
  return static_cast<unsigned char>(

    const tensorflow::ServerDef& cluster_server_def, int task_index) {
  tensorflow::ServerDef single_host_server_def;
  single_host_server_def.set_job_name("worker");
  single_host_server_def.set_protocol(cluster_server_def.protocol());
  single_host_server_def.set_task_index(0);
  tensorflow::ClusterDef* cluster_def =
      single_host_server_def.mutable_cluster();
  tensorflow::JobDef* job_def = cluster_def->add_job();

  };
  tf_gcs_filesystem::RamFileBlockCache cache(block_size, 2 * block_size, 0,
                                             fetcher);
  std::vector<char> out;
  // Read the second block; this should yield an OK status and a single byte.
  TF_EXPECT_OK(ReadCache(&cache, "", block_size, block_size, &out));
  EXPECT_EQ(out.size(), 1);
  // Now read the first block; this should yield an INTERNAL error because we

  EXPECT_EQ(X, dxy_dyValue);

  TF_DeleteTensor(fetchValues[0]);
  TF_DeleteTensor(fetchValues[1]);
}

// REGISTER_OP for CApiAttributesTest test cases.
// Registers two ops, each with a single attribute called 'v'.
// The attribute in one op will have a type 'type', the other
// will have list(type).
#define ATTR_TEST_REGISTER_OP(type)                           \

  ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());
  TFE_DeleteContextOptions(opts);

  ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());

  /* Want to test simple MatMul example:
    [[0,0],    *   [[0,0],    =   [[0,0],
     [0,0]]         [0,0]]         [0,0]]
  */

  // Build an abstract input tensor.
  int64_t dims[] = {2, 2};  // Matrices will be 2 x 2

    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);

        // artificial restriction and require that when num_opers=-1, such
        // nodes must have a single output.
        if (node->num_outputs() != 1) {
          return InvalidArgument(
              "When `num_opers` is set to -1, nodes referenced in `inputs` "
              "must have a single output. Node ",
              node->name(), " has ", node->num_outputs(),

  }
};

using ExecutorPtr = std::unique_ptr<TFE_Executor, ExecutorDeleter>;

}  // namespace

// Allows a single op at a time to be launched without blocking.
//
// DeviceThread itself is thread-safe, in that StartExecute will block if there
// is a pending execution. Since StartExecute is equivalent to grabbing a lock,
// multiple DeviceThreads should always be accessed in the same order to avoid
// deadlocks.