if (BlockNotStale(entry->second)) {
      return entry->second;
    } else {
      // Remove the stale block and continue.
      RemoveFile_Locked(key.first);
    }
  }

  // Insert a new empty block, setting the bookkeeping to sentinel values
  // in order to update them as appropriate.
  auto new_entry = std::make_shared<Block>();
  lru_list_.push_front(key);
  lra_list_.push_front(key);

  ASSERT_EQ(TF_INVALID_ARGUMENT, TF_GetCode(status.get()));
  ASSERT_EQ(device_type, nullptr);
  ASSERT_EQ("Invalid handle", string(TF_Message(status.get())));

  TF_SetStatus(status.get(), TF_OK, "");

  int device_id = TFE_TensorHandleDeviceID(h, status.get());
  ASSERT_EQ(TF_INVALID_ARGUMENT, TF_GetCode(status.get()));
  ASSERT_EQ(device_id, -1);

                                      TF_Status* status) -> int64_t {
    EXPECT_EQ(n, block_size);
    EXPECT_EQ(offset % block_size, 0);
    EXPECT_EQ(calls.find(offset), calls.end()) << "at offset " << offset;
    calls.insert(offset);
    memset(buffer, 'x', n);
    TF_SetStatus(status, TF_OK, "");
    return n;
  };
  const uint32 block_count = 256;
  tf_gcs_filesystem::RamFileBlockCache cache(

  auto iter = registry_.find(op_name);
  if (iter != registry_.end()) {
    const string error_msg = "Gradient already exists for op: " + op_name + ".";
    return errors::AlreadyExists(error_msg);
  }
  registry_.insert({op_name, gradient_function_factory});
  return absl::OkStatus();
}
Status GradientRegistry::Lookup(
    const ForwardOperation& op,
    std::unique_ptr<GradientFunction>* gradient_function) const {

                           TF_DataType value) {
  auto iter = builder->attr_names.insert(attr_name).first;
  builder->Set(*iter, static_cast<tensorflow::DataType>(value));
}

void TF_AttrBuilderSetTypeList(TF_AttrBuilder* builder, const char* attr_name,
                               const TF_DataType* values, int num_values) {
  auto iter = builder->attr_names.insert(attr_name).first;

        missing_it->second--;
        VLOG(1) << "Op " << op_id << " missing " << missing_it->second
                << " output gradients";
        if (missing_it->second == 0) {
          op_stack.insert(op_stack.begin(), op_id);
        }
      }
    }
  }
  if (!state.op_tape.empty()) {
    return tensorflow::errors::Internal("Invalid tape state.");
  }
  if (result.size() != source_tensor_ids.size()) {

  tensorflow::JobDef* job_def = cluster_def->add_job();
  job_def->set_name(job_name);
  for (int i = 0; i < num_tasks; i++) {
    int port = tensorflow::testing::PickUnusedPortOrDie();
    job_def->mutable_tasks()->insert(
        {i, tensorflow::strings::StrCat("localhost:", port)});
    LOG(INFO) << "Picked test port: " << port << " for job: " << job_name
              << ", task: " << i;
  }
  return server_def;
}

  TFE_Op* op = TFE_NewOp(ctx, "ReadVariableOp", status);
  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);
  TFE_OpSetAttrType(op, "dtype", TF_FLOAT);
  TFE_OpAddInput(op, var, status);
  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);
  int num_retvals = 1;
  TFE_Execute(op, out_value, &num_retvals, status);
  ASSERT_EQ(TF_OK, TF_GetCode(status)) << TF_Message(status);
  TFE_DeleteOp(op);

  ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());

  // Execute.
  TF_ExecuteOperation(add_op, 2, inputs, add_outputs, status.get());
  ASSERT_EQ(TF_OK, TF_GetCode(status.get())) << TF_Message(status.get());

  // Test that full type information can be accessed.
  auto outs = unwrap(add_outputs);
  auto h = outs->outputs[0];
  ASSERT_NE(h, nullptr);

                                    " into function '", fn_name, "'");

    input_tensors->emplace_back(node, idx);

    const auto& iter = input_nodes->find(node);
    if (iter == input_nodes->end()) {
      input_nodes->insert({node, {idx}});
    } else {
      auto& indices = iter->second;
      if (std::find(indices.begin(), indices.end(), idx) != indices.end()) {
        return InvalidArgument("TF_Output ", node->name(), ":", idx,