char*** entries, TF_Status* status) {
  struct dirent** dir_entries = nullptr;
  /* we don't promise entries would be sorted */
  int num_entries =
      scandir(path, &dir_entries, RemoveSpecialDirectoryEntries, nullptr);
  if (num_entries < 0) {
    TF_SetStatusFromIOError(status, errno, path);
  } else {
    *entries = static_cast<char**>(

 private:
  TensorTape tensor_tape_;
  OpTape<BackwardFunction, TapeTensor> op_tape_;
  int64_t next_op_id_{0};

  // Map from tensor id to number of remaining usages (i.e. how many entries in
  // the tape refer to it); to aid in tape garbage collection.
  std::unordered_map<int64_t, int64_t> tensor_usage_;

  // If false, all activations are deleted in the first call to ComputeGradient.

  virtual Status SetOrUpdateServerDef(const ServerDef& server_def,
                                      bool reset_context, int keep_alive_secs,
                                      int64_t init_timeout_in_ms, int retries,
                                      bool clear_existing_contexts = false) = 0;

  // Initializes context for the local worker and no contexts will be created

    TFE_Context* ctx, int keep_alive_secs, const void* proto, size_t proto_len,
    int64_t init_timeout_in_ms, TF_Status* status,
    bool clear_existing_contexts);

// Set server def with retries and timeout. This is helpful for fault-tolerant
// initial connection in high-preemption environments, such as
// ParameterServerStrategy training.
// This API is for experimental usage and may be subject to change.

      ctx, keep_alive_secs, proto, proto_len, init_timeout_in_ms,
      /*retries=*/0, status, clear_existing_contexts);
}

// Set server_def on the context, possibly updating it.
// TODO(b/291142876) Simplify TFE_ContextSetServerDefWithTimeoutAndRetries and
// TFE_ContextUpdateServerDefWithTimeout to be simple wrappers around the same
// C++ function.
// Retries are used for CreateContext calls, which is used in

        can fix this mismatch by either calling the model with correctly-shaped
        data, or by relaxing `Input` shape assumptions (note that you can pass
        shapes with `None` entries for axes that are meant to be dynamic). You
        can also disable the input checking entirely by setting
        `model.input_spec = None`.
    *   Several changes have been made to

    - name: Update the RBE Configs
      run: |
        function map() {
          # The "digest" that allows us to pull an image is not the digest as
          # returned by the API, but a sha256sum of the entire chunk of image
          # metadata. gcr.io helpfully includes it in the header of the response
          # as docker-content-digest: sha256:[digest]. Note we use egrep to

# TensorFlow Bazel configuration file.
# This file tries to group and simplify build options for TensorFlow
#
# ----CONFIG OPTIONS----
# Android options:
#    android:
#    android_arm:
#    android_arm64:
#    android_x86:
#    android_x86_64:
#
# iOS options:
#     ios:
#     ios_armv7:
#     ios_arm64:
#     ios_x86_64:
#     ios_fat:
#
# Macosx options
#     darwin_arm64:
#
# Compiler options: