const char* attr_name, const void* proto,
                                size_t proto_len, TF_Status* status) {
  // shape.ParseFromArray takes an int as length, this function takes size_t,
  // make sure there is no information loss.
  if (proto_len > std::numeric_limits<int>::max()) {
    status->status = InvalidArgument(
        "proto_len (", proto_len,

// On success, returns a set of TF_Function instances from `text_proto` of
// GraphDef type. These functions must be deleted by calling TF_DeleteFunction.
//
// If `mutate_proto_func` is non-NULL, run it over each FunctionDef proto,
// before creating a TF_Function out of the possibly mutated proto.
static std::vector<UniqueFuncPtr> CreateFunctionsFromTextProto(
    const char* text_proto,

                                  const void* proto, size_t proto_len,
                                  TF_Status* status) {
  tensorflow::AttrValue attr_value;
  if (!attr_value.ParseFromArray(proto, proto_len)) {
    status->status = InvalidArgument(
        "Unparseable AttrValue proto passed to "
        "TF_FunctionSetAttrValueProto");
    return;
  }

  tensorflow::Status Run(const std::string& function_name,
                         const tensorflow::DeviceSet& device_set,
                         const tensorflow::ConfigProto& config_proto,
                         const FunctionOptions& function_options,
                         std::unique_ptr<tensorflow::Graph>* graph,
                         tensorflow::FunctionLibraryDefinition* flib_def,

    }
    if (num_dims_i < 0) {
      proto[i].set_unknown_rank(true);
    } else {
      const int64_t* dims_i = dims[i];
      auto proto_i = &proto[i];
      for (int d = 0; d < num_dims_i; ++d) {
        proto_i->add_dim()->set_size(dims_i[d]);
      }
    }
  }
  forward_op_.attrs.Set(
      attr_name, gtl::ArraySlice<TensorShapeProto>(proto.get(), num_values));

        << entry.InitializationErrorString();
    for (int i = 0; i < entry.slices_size(); ++i) {
      const auto& slice_proto = entry.slices(i);
      CHECK(filtered_keys
                .insert(EncodeTensorNameSlice(
                    string(v2_reader_->key()) /* full var's name */,
                    TensorSlice(slice_proto)))
                .second);
    }
  }

  // Second pass: adds the entries, ignoring the filtered keys.

  string bytes1, bytes2;
  tensorflow::TensorShapeProto proto;

  const int64_t pts1[] = {2, 4, -1, 8};
  tensorflow::PartialTensorShape(pts1).AsProto(&proto);
  proto.SerializeToString(&bytes1);

  const int64_t pts2[] = {1, 3, 5, 7};
  tensorflow::PartialTensorShape(pts2).AsProto(&proto);
  proto.SerializeToString(&bytes2);

  std::unique_ptr<const void*[]> list_ptrs;

  TFE_TensorHandle* retval[1] = {nullptr};
  int num_retvals = 1;
  TFE_Op* op = TFE_NewOp(ctx, "AddFunction", status);
  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);

  // Add a config_proto attr, to trigger grappler graph rewrites in the current
  // eager runtime.
  if (enable_grappler) {
    tensorflow::ConfigProto config;
    // Do not skip grappler optimization even for small graphs.

    }
    if (num_dims_i < 0) {
      proto[i].set_unknown_rank(true);
    } else {
      const int64_t* dims_i = dims[i];
      auto proto_i = &proto[i];
      for (int d = 0; d < num_dims_i; ++d) {
        proto_i->add_dim()->set_size(dims_i[d]);
      }
    }
  }
  forward_op_->attrs.Set(
      attr_name, gtl::ArraySlice<TensorShapeProto>(proto.get(), num_values));

                             const void* proto, size_t proto_len,
                             TF_Status* status) {
  tensorflow::AttrValue attr_value;
  if (!attr_value.ParseFromArray(proto, proto_len)) {
    status->status =
        tensorflow::errors::InvalidArgument("Unparseable AttrValue proto");
    return;
  }
  if (op == nullptr) {