demos
demos copied to clipboard
QoS_demo_cpp: Process hung on [INFO] [rclcpp]: signal_handler(signal_value=2)
Bug report
Required Info:
- Operating System:
- Ubuntu 18.04
- Installation type:
- Source
- Version or commit hash:
- Eloquent
- DDS implementation:
- Connext
- Client library (if applicable):
- rclcpp
Steps to reproduce issue
RMW_IMPLEMENTATION=rmw_connext_cpp ros2 run quality_of_service_demo_cpp interactive_publisher --delay 0.5 --deadline 1 --liveliness MANUAL_BY_TOPIC --lease 1
RMW_IMPLEMENTATION=rmw_connext_cpp ros2 run quality_of_service_demo_cpp interactive_subscriber --deadline 1 --liveliness MANUAL_BY_TOPIC --lease 1
Expected behavior
^C[INFO] [rclcpp]: signal_handler(signal_value=2)
then exit the process
Actual behavior
[INFO] [rclcpp]: signal_handler(signal_value=2)
[INFO] [rclcpp]: signal_handler(signal_value=2)
[INFO] [rclcpp]: signal_handler(signal_value=2)
[INFO] [rclcpp]: signal_handler(signal_value=2)
never exit.
Out of curiousity, how long did you wait? We have a known issue somewhere where Connext can take a substantial amount of time to cleanup after itself (like 10 seconds or something). I'm wondering if this is a duplicate.
Out of curiousity, how long did you wait? We have a known issue somewhere where Connext can take a substantial amount of time to cleanup after itself (like 10 seconds or something). I'm wondering if this is a duplicate.
I tried again, waited for at least five minutes and got the same result. I had to close the terminal to kill it.
I tried again, waited for at least five minutes and got the same result. I had to close the terminal to kill it.
Sounds like something new :+1:
I came across the same thing. @claireyywang try pressing x
...
The thing is that these demos spin up a separate thread to support stdin
-based user interaction. Because of the way that bit is coded, you have to either Ctrl-C or press x
(to stop the executor) and then press another random key to get yourself out of a blocking getch*()
call.
I fiddled a bit on how to fix it but being cross-platform is going to be time consuming. I'll come back later.
@claireyywang can you push on this issue. @hidmic says the code may need to be refactored to close when doing ctrl-c. We can either backlog this or fix it. Thanks!
@wjwwood sure I'll look into it
Since we don't support rmw_connext_cpp
anymore, I'm going to go ahead and close this out.