return OkStatus();
}

Status ModularWritableFile::Append(StringPiece data) {
  if (ops_->append == nullptr)
    return errors::Unimplemented(tensorflow::strings::StrCat(
        "Append() not implemented for ", filename_));

  UniquePtrTo_TF_Status plugin_status(TF_NewStatus(), TF_DeleteStatus);
  ops_->append(file_.get(), data.data(), data.size(), plugin_status.get());
  return StatusFromTF_Status(plugin_status.get());
}

    forward_op_.inputs.push_back(input);
  }
  return OkStatus();
}
Status TapeOperation::SetAttrString(const char* attr_name, const char* data,
                                    size_t length) {
  forward_op_.attrs.Set(attr_name, StringPiece(data, length));
  return parent_op_->SetAttrString(attr_name, data, length);
}
Status TapeOperation::SetAttrInt(const char* attr_name, int64_t value) {

  std::vector<tensorflow::string> tensor_id_data;
};

struct TF_ImportGraphDefResults {
  std::vector<TF_Output> return_tensors;
  std::vector<TF_Operation*> return_nodes;
  std::vector<const char*> missing_unused_key_names;
  std::vector<int> missing_unused_key_indexes;

  // Backing memory for missing_unused_key_names values.
  std::vector<tensorflow::string> missing_unused_key_names_data;
};

struct TF_DeviceList {

                 std::vector<char>* out) {
  out->clear();
  out->resize(n, 0);
  TF_Status status;
  auto bytes_transferred =
      cache->Read(filename, offset, n, out->data(), &status);
  if (bytes_transferred >= 0) {
    EXPECT_LE(bytes_transferred, n);
    out->resize(bytes_transferred, n);
  }
  return status.status;
}

TEST(RamFileBlockCacheTest, IsCacheEnabled) {

                  dltensor_out->shape[i + 1] * dltensor_out->strides[i + 1]);
      }
    }
  }
  const float* data_in = static_cast<const float*>(dltensor_in->data);
  const float* data_out = static_cast<const float*>(dltensor_out->data);
  for (size_t j = 0; j < num_elements; ++j) {
    EXPECT_EQ(data_out[j], data_in[j]);
  }

  TFE_CallDLManagedTensorDeleter(dlm_out);
  TFE_DeleteTensorHandle(handle);
}

  bool did_work = false;
};

void SomeThreadFunc(void* data) {
  auto* real_data = static_cast<SomeThreadData*>(data);
  ::tensorflow::mutex_lock l(real_data->mu);
  real_data->did_work = true;
}

}  // namespace

TEST(TestEnv, TestThreads) {
  TF_ThreadOptions options;
  TF_DefaultThreadOptions(&options);
  SomeThreadData data;
  TF_Thread* thread =

  owner->reference.Unref();
  delete owner;
}

// Converts TF_DATAType to DLPack data type.
DLDataType GetDlDataType(TF_DataType data_type, TF_Status* status) {
  DLDataType dtype;
  dtype.lanes = 1;
  dtype.bits = TF_DataTypeSize(data_type) * 8;
  switch (data_type) {
    case TF_DataType::TF_BOOL:
      dtype.code = DLDataTypeCode::kDLBool;
      break;
    case TF_DataType::TF_HALF:

     */

    AbstractTensorHandle* upstream_grad = grad_outputs[0];
    DCHECK(upstream_grad);

    // Recover data format from forward pass for gradient.
    std::string data_format;
    TF_RETURN_IF_ERROR(forward_attrs_.Get("data_format", &data_format));

    // Grad for A
    grad_inputs[0] = upstream_grad;
    grad_inputs[0]->Ref();

    // Grad for bias

  return SetAttrShape(attr_name, shape.dim_sizes().data(), shape.dims());
}

inline Status AbstractOperation::SetAttrStringList(
    const char* attr_name, absl::Span<string const> values) {
  std::vector<const char*> raw_strs;
  std::vector<size_t> lengths;
  raw_strs.reserve(values.size());
  lengths.reserve(values.size());
  for (const auto& s : values) {
    raw_strs.emplace_back(s.data());
    lengths.emplace_back(s.size());

      auto* out_shape_and_type = handle_data.add_shape_and_type();
      ic->ShapeHandleToProto(p.shape, out_shape_and_type->mutable_shape());
      out_shape_and_type->set_dtype(p.dtype);
      *out_shape_and_type->mutable_type() = p.type;
    }
  }
  string str_data;
  handle_data.SerializeToString(&str_data);

  TF_Buffer* result = TF_NewBufferFromString(str_data.c_str(), str_data.size());
  return result;
}