::tensorflow::Status s =
      ::tensorflow::Env::Default()->NewWritableFile(filename, &f);
  ::tensorflow::Set_TF_Status_from_Status(status, s);

  if (s.ok()) {
    *handle = reinterpret_cast<TF_WritableFileHandle*>(f.release());
  }
}

void TF_CloseWritableFile(TF_WritableFileHandle* handle, TF_Status* status) {
  auto* cc_file = reinterpret_cast<::tensorflow::WritableFile*>(handle);
  TF_SetStatus(status, TF_OK, "");

# Download specific version of libstdc++ shared library based on the value of
# the `VERSION` parameter
  # Download binary libstdc++ 4.8 shared library release
wget --retry-connrefused --waitretry=1 --read-timeout=20 --timeout=15 --tries=5 "http://old-releases.ubuntu.com/ubuntu/pool/main/g/gcc-4.8/libstdc++6_4.8.1-10ubuntu8_arm64.deb" && \
    unar "libstdc++6_4.8.1-10ubuntu8_arm64.deb" && \

  _logging.warning("""

  TensorFlow's `tf-nightly` package will soon be updated to TensorFlow 2.0.

  Please upgrade your code to TensorFlow 2.0:
    * https://www.tensorflow.org/guide/migrate

  Or install the latest stable TensorFlow 1.X release:
    * `pip install -U "tensorflow==1.*"`

  Otherwise your code may be broken by the change.

  """)

#      Ex. no_docker  -- Disable docker on enabled platforms
#    See full examples below for more details on these. Some other modifiers are:
#      Ex. versions_upload -- for TF official release versions
#      Ex. nightly_upload -- for TF nightly official builds; changes version numbers
#      Ex. no_upload      -- Disable all uploads, usually for temporary CI issues

# Recommended: use a local+remote cache.
#

  // Create a tensor instance from the given data buffer and description.
  // `memory_releaser` will be called on destruction, and it's responsible for
  // cleaning up the underlying buffer.
  virtual AbstractTensorInterface* CreateTensor(
      DataType dtype, const int64_t* dims, int num_dims, void* data, size_t len,
      MemoryReleaser memory_releaser, void* memory_releaser_arg) = 0;

  // Create a handle to wrap and manage a Tensor

  status->status = new_op->Reset(op_or_function_name, nullptr);
  if (!status->status.ok()) {
    new_op->Release();
    new_op = nullptr;
  }
  return tensorflow::wrap(new_op);
}

void TFE_DeleteOp(TFE_Op* op) {
  if (op == nullptr) {
    return;
  }

  tensorflow::unwrap(op)->Release();
}

const char* TFE_OpGetName(const TFE_Op* op, TF_Status* status) {

  TF_RETURN_IF_ERROR(ops::Exp(ctx, inputs[0], &exp_output, "Exp"));
  std::unique_ptr<GradientFunction> gradient_function(
      new PassThroughGradientFunction);
  tape.RecordOperation(inputs, {exp_output}, gradient_function.release());
  TF_RETURN_IF_ERROR(tape.ComputeGradient(ctx,
                                          /*targets*/ {exp_output},
                                          /*sources=*/inputs,

  TFE_DeleteContext(ctx);

  TF_DeleteStatus(status);

  // TODO(b/136478427): Figure out how to correctly shut the server down.
  worker_server1.release();
  worker_server2.release();
}

TEST(CAPI, TestLocalFunctionWithPackedInput) {
  TestFunctionWithPackedInput(/*remote=*/false);
}

TEST(CAPI, TestRemoteFunctionWithPackedInput) {

    the bare minimum necessary to generate the problem. If possible, please
    share a link to Colab/Jupyter/any notebook. placeholder: Tell us what you
    see! value: render: shell validations: required: true

-   type: textarea id: logs attributes: label: Relevant log output description:
    Please copy and paste any relevant log output. This will be automatically

  TFE_DeleteExecutor(executor);
  TFE_DeleteContext(ctx);

  TF_DeleteStatus(status);

  // TODO(b/136478427): Figure out how to correctly shut the server down.
  worker_server.release();
}

TEST(CAPI, RemoteExecute) { TestRemoteExecute(false); }
TEST(CAPI, RemoteExecuteAsync) { TestRemoteExecute(true); }

void TestRemoteExecuteSilentCopiesOp(bool async, bool remote,