Seems to be same issue as #526 (Indigo) #813 (Kinetic) . On #526 the most recent comment also references Melodic (since closing). However I'm experiencing it on Noetic
Dump from term for get_num_connections() and rostopic.get_info_text with instance specific type/connection info replaced by <x>
Topic is latched
Update:
I have tried to create a minimal working example, and with happy flow I can't repeat it. One thing that is still different to the implementation where I see the issue is the publishing node has already received the shutdown signal (via rosnode.kill_nodes(node_name), and is processing pre shutdown hooks. Subscriber that is unregistering has not received shutdown signal.
Creating a shutdown hook and then calling rospy.signal_shutdown() in the node also doesn't seem show the issue in a minimal working example
Seems to be same issue as #526 (Indigo) #813 (Kinetic) . On #526 the most recent comment also references Melodic (since closing). However I'm experiencing it on Noetic
Dump from term for
get_num_connections()
androstopic.get_info_text
with instance specific type/connection info replaced by<x>
Topic is latched
Update: I have tried to create a minimal working example, and with happy flow I can't repeat it. One thing that is still different to the implementation where I see the issue is the publishing node has already received the shutdown signal (via
rosnode.kill_nodes(node_name)
, and is processing pre shutdown hooks. Subscriber that is unregistering has not received shutdown signal.Creating a shutdown hook and then calling
rospy.signal_shutdown()
in the node also doesn't seem show the issue in a minimal working example