selects.config_setting_group(
    name = "mobile",
    match_any = [
        ":android",
        ":chromiumos",
        # TODO(jakeharmon8): Move all platform configs to TSL with fuchsia (non-blocking)
        ":fuchsia",
        ":emscripten",
        ":ios",
    ],
    visibility = ["//visibility:public"],
)

# This flag disables all google production dependencies, intended for

TFCI_BAZEL_TARGET_SELECTING_CONFIG_PREFIX=linux_arm64
# Note: this is not set to "--cpu", because that changes the package name
# to tensorflow_cpu. These ARM builds are supposed to have the name "tensorflow"
# despite lacking Nvidia CUDA support.
TFCI_BUILD_PIP_PACKAGE_ARGS="--repo_env=WHEEL_NAME=tensorflow"
TFCI_DOCKER_ENABLE=1
TFCI_DOCKER_IMAGE=gcr.io/tensorflow-sigs/build-arm64:tf-2-16-multi-python
TFCI_DOCKER_PULL_ENABLE=1

      TF_DeleteTensor);
  ASSERT_TRUE(TF_GetCode(status.get()) == TF_OK) << TF_Message(status.get());
  ASSERT_EQ(111., *static_cast<float*>(TF_TensorData(resolved_value.get())));

  // Free the backing buffer for the variable.
  op.reset(TFE_NewOp(context.get(), "DestroyResourceOp", status.get()));
  TFE_OpAddInput(op.get(), var_handle, status.get());
  TFE_OpSetDevice(op.get(), name, status.get());

  ///
  /// In order to prevent deadlocks, never grab the block-cache-wide mu_ lock
  /// AFTER grabbing any block's mu lock. It is safe to grab mu without locking
  /// mu_.
  struct Block {
    /// The block data.
    std::vector<char> data;
    /// A list iterator pointing to the block's position in the LRU list.
    std::list<Key>::iterator lru_iterator;

automatically on GitHub.

If you want to contribute, start working through the TensorFlow codebase,
navigate to the
[GitHub "issues" tab](https://github.com/tensorflow/tensorflow/issues) and start
looking through interesting issues. If you are not sure of where to start, then
start by trying one of the smaller/easier issues here i.e.

    ASSERT_TRUE(remote_arg->HasLocalMirror(nullptr));
  }

  if (remote_func_outputs) {
    const string backing_device =
        TFE_TensorHandleBackingDeviceName(retvals[0], status);
    ASSERT_EQ(TF_GetCode(status), TF_OK) << TF_Message(status);
    EXPECT_EQ(backing_device, task2_name);
  }

  auto* retval_task0 = TFE_TensorHandleCopyToDevice(

      tf.data service
      (https://www.tensorflow.org/api_docs/python/tf/data/experimental/service)
      to write distributed dataset snapshots. The call is non-blocking and
      returns without waiting for the snapshot to finish. Setting `wait=True` to
      `tf.data.Dataset.load` allows the snapshots to be read while they are
      being written.

### Bug Fixes and Other Changes

  static Variable* Create(TFE_Context* context, TF_DataType type,
                          const int64_t* dims, const int num_dims,
                          const char* device, TF_Status* status);
  // Dereferences the backing buffer for the variable. Note that since this can
  // fail (it runs operations), it must be called explicitly and the resulting
  // `status` checked.
  void Destroy(TFE_Context* context, TF_Status* status);

          f'Bazel invocations.\n'
          f'ResultStore contains individual representations of each target '
          f'that were run/built during the invocation.\n'
          f'These results are generally easier to read than looking through '
          f'the entire build log:\n')
  i = 1
  for url, invocation_results in result_store_dict.items():
    line_str = f'Invocation #{i} ({invocation_results["status"]}):\n'

  if (TF_GetCode(status) != TF_OK) return nullptr;
  return new Variable(var_handle, type);
}

void Variable::Destroy(TFE_Context* context, TF_Status* status) {
  // Free the backing buffer for the variable.
  std::unique_ptr<TFE_Op, decltype(&TFE_DeleteOp)> op(
      TFE_NewOp(context, "DestroyResourceOp", status), &TFE_DeleteOp);
  if (TF_GetCode(status) != TF_OK) return;