platform_velocity_controller publishes tf as TRANSIENT_LOCAL

[taDachs]

Please provide the following information:

• OS: Ubuntu 22.04
• ROS 2 Distro: Humble
• Built from source or installed: tried both
• Package version: 1.0.0-1jammy.20241126.044506 (apt), 2ac65c7 (source)
• Real hardware or simulation: hardware, tried with mock hardware as well (ros2_control mock hardware, not simulation)

Expected behaviour
Normal tf (not tf static) should be published VOLATILE

Actual behaviour
platform_velocity_controller publishes tf as TRANSIENT_LOCAL

To Reproduce
Start the platform_velocity_controller and check the QoS for tf

Other notes
Output of ros2 topic info -v /husky/tf

Node name: platform_velocity_controller
Node namespace: /husky
Topic type: tf2_msgs/msg/TFMessage
Endpoint type: PUBLISHER
GID: 01.0f.57.0b.0d.89.5f.73.00.00.00.00.00.00.4f.03.00.00.00.00.00.00.00.00
QoS profile:
  Reliability: RELIABLE
  History (Depth): UNKNOWN
  Durability: TRANSIENT_LOCAL
  Lifespan: Infinite
  Deadline: Infinite
  Liveliness: AUTOMATIC
  Liveliness lease duration: Infinite

/tf (or in this case /husky/tf) should always be published as VOLATILE, as TRANSIENT_LOCAL latches the topic and old tf messages have no value. It also can lead to error messages because of dropped messages. It is not a critical issue or anything but should still be mentioned

[Mechazo11]

Hi @taDachs I was wondering does this error lead to TF_OLD_DATA warning message?

[taDachs]

@Mechazo11 hey, yes, this can happen. TF_OLD_DATA gets thrown when data is added to your tf_buffer that is older than the cache limit (see here). By latching the topic, old messages might get added to your tf buffer, resulting in that error.

[tonybaltovski]

@taDachs this looks deeper than this repo. It looks to be from here: https://github.com/ros-controls/ros2_controllers/blob/master/diff_drive_controller/src/diff_drive_controller.cpp#L412 Do you want to open an issue there?