# pretend the path is a directory.
TFCI_ARTIFACT_FINAL_GCS_ENABLE=1
TFCI_ARTIFACT_FINAL_GCS_SA_PATH="${KOKORO_KEYSTORE_DIR}/73361_tensorflow_release_binary_uploader_service_account"
TFCI_ARTIFACT_FINAL_GCS_URI="gs://tensorflow/versions/"
TFCI_ARTIFACT_FINAL_PYPI_ARGS="--config-file=$KOKORO_KEYSTORE_DIR/73361_tensorflow_pypirc_using_global_api_token --repository pypi-warehouse"
TFCI_ARTIFACT_FINAL_PYPI_ENABLE=1

  pip install twine==3.6.0
fi

# Scheduled nightly and release builds upload build artifacts (Pip packages,
# Libtensorflow archives) to GCS buckets. TFCI Mac VMs need to authenticate as
# a service account that has the right permissions to be able to do so.
set +x
if [[ -n "${GOOGLE_APPLICATION_CREDENTIALS:-}" ]]; then
  # Python 3.12 removed the module `imp` which is needed by gcloud CLI so we set

        shell: bash

      )
   ```

4) Add the version to `SUPPORTED_VERSIONS` in `updater.sh`, and
   `release_updater.sh`

5) Run the `updater.sh` shell script. \
   If the base requirements file hasn't yet been updated to account for the new
   Python version, which will require different versions for at least some
   dependencies, it will need to be updated now, for the script to run
   successfully.

# pretend the path is a directory.
# 1. Upload nightlies
TFCI_ARTIFACT_FINAL_GCS_ENABLE=1
TFCI_ARTIFACT_FINAL_GCS_SA_PATH="${KOKORO_KEYSTORE_DIR}/73361_tensorflow_release_binary_uploader_service_account"
TFCI_ARTIFACT_FINAL_GCS_URI="gs://tensorflow/nightly/"
TFCI_ARTIFACT_FINAL_PYPI_ARGS="--config-file=$KOKORO_KEYSTORE_DIR/73361_tensorflow_pypirc_using_global_api_token --repository pypi-warehouse"
TFCI_ARTIFACT_FINAL_PYPI_ENABLE=1

# gs://tensorflow/nightly/2.16.0-dev20240105 (nightly), overwriting previous values.
if [[ "$TFCI_ARTIFACT_FINAL_GCS_ENABLE" == 1 ]]; then
  gcloud auth activate-service-account --key-file="$TFCI_ARTIFACT_FINAL_GCS_SA_PATH"

  # $TF_VER_FULL will resolve to e.g. "2.15.0-rc2". Since $TF_VER_FULL comes
  # from get_versions.sh, which must be run *after* update_version.py, FINAL_URI

    memset(buffer, 'x', n);
    TF_SetStatus(status, TF_OK, "");
    return n;
  };
  const uint32 block_count = 256;
  tf_gcs_filesystem::RamFileBlockCache cache(
      block_size, block_count * block_size, 0, fetcher);
  std::vector<char> out;
  out.resize(block_count, 0);
  // The cache has space for `block_count` blocks. The loop with i = 0 should
  // fill the cache, and the loop with i = 1 should be all cache hits. The

      new TensorSliceReader::VarToDataTypeMap);
  v2_reader_->Seek(kHeaderEntryKey);
  for (v2_reader_->Next(); v2_reader_->Valid(); v2_reader_->Next()) {
    if (filtered_keys.count(string(v2_reader_->key())) > 0) continue;
    CHECK(entry.ParseFromArray(v2_reader_->value().data(),
                               v2_reader_->value().size()))
        << entry.InitializationErrorString();

      EXPECT_EQ(orig_ptr, TF_TensorData(t));
    } else {
      // In sync mode, forwarding can't really happen since the client code will
      // have a reference count on the input tensor while the kernel is being
      // executed and thus it cannot be re-used for the return tensor.
      EXPECT_NE(orig_ptr, TF_TensorData(t));
    }
  }

    if (edge->src_output() == oper_out.index) {
      if (count < max_consumers) {
        consumers[count] = {ToOperation(edge->dst()), edge->dst_input()};
      }
      ++count;
    }
  }
  return count;
}

int TF_OperationNumControlInputs(TF_Operation* oper) {
  int count = 0;
  for (const auto* edge : oper->node.in_edges()) {
    if (edge->IsControlEdge() && !edge->src()->IsSource()) {