fix: added workaround for call to double calls to take_data

This adds a workaround for a known bug in the executor in iron.

Signed-off-by: Janosch Machowinski <J.Machowinski@cellumation.com>

rclcpp_action/src/client.cpp

@@ -161,6 +161,8 @@ class ClientBaseImpl
 
  // next ready event for taking, will be set by is_ready and will be processed by take_data
  std::atomic<size_t> next_ready_event;
+ // used to indicate that next_ready_event has no ready event for processing
+ static constexpr size_t NO_EVENT_READY = std::numeric_limits<size_t>::max();
 
  rclcpp::Context::SharedPtr context_;
  rclcpp::node_interfaces::NodeGraphInterface::WeakPtr node_graph_;
@@ -351,7 +353,7 @@ ClientBase::is_ready(rcl_wait_set_t * wait_set)
  }
  }
 
- pimpl_->next_ready_event = std::numeric_limits<size_t>::max();
+ pimpl_->next_ready_event = ClientBaseImpl::NO_EVENT_READY;
 
  if (is_feedback_ready) {
  pimpl_->next_ready_event = static_cast<size_t>(EntityType::FeedbackSubscription);
@@ -647,10 +649,12 @@ std::shared_ptr<void>
 ClientBase::take_data()
 {
  // next_ready_event is an atomic, caching localy
- size_t next_ready_event = pimpl_->next_ready_event.exchange(std::numeric_limits<uint32_t>::max());
+ size_t next_ready_event = pimpl_->next_ready_event.exchange(ClientBaseImpl::NO_EVENT_READY);
 
- if (next_ready_event == std::numeric_limits<uint32_t>::max()) {
- throw std::runtime_error("Taking data from action client but nothing is ready");
+ if (next_ready_event == ClientBaseImpl::NO_EVENT_READY) {
+ // there is a known bug in iron, that take_data might be called multiple
+ // times. Therefore instead of throwing, we just return a nullptr as a workaround.
+ return nullptr;
  }
 
  return take_data_by_entity_id(next_ready_event);
@@ -751,7 +755,9 @@ void
 ClientBase::execute(std::shared_ptr<void> & data_in)
 {
  if (!data_in) {
- throw std::invalid_argument("'data_in' is unexpectedly empty");
+ // workaround, if take_data was called multiple timed, it returns a nullptr
+ // normally we should throw here, but as an API stable bug fix, we just ignore this...
+ return;
  }
 
  std::shared_ptr<ClientBaseData> data_ptr = std::static_pointer_cast<ClientBaseData>(data_in);

rclcpp_action/src/server.cpp

@@ -273,8 +273,10 @@ ServerBase::take_data()
 {
  size_t next_ready_event = pimpl_->next_ready_event.exchange(ServerBaseImpl::NO_EVENT_READY);
 
- if (next_ready_event == ServerBaseImpl::NO_EVENT_READY;) {
- throw std::runtime_error("ServerBase::take_data() called but no data is ready");
+ if (next_ready_event == ServerBaseImpl::NO_EVENT_READY) {
+ // there is a known bug in iron, that take_data might be called multiple
+ // times. Therefore instead of throwing, we just return a nullptr as a workaround.
+ return nullptr;
  }
 
  return take_data_by_entity_id(next_ready_event);
@@ -363,7 +365,9 @@ void
 ServerBase::execute(std::shared_ptr<void> & data_in)
 {
  if (!data_in) {
- throw std::runtime_error("ServerBase::execute: give data pointer was null");
+ // workaround, if take_data was called multiple timed, it returns a nullptr
+ // normally we should throw here, but as an API stable bug fix, we just ignore this...
+ return;
  }
 
  std::shared_ptr<ServerBaseData> data_ptr = std::static_pointer_cast<ServerBaseData>(data_in);