You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
Hi. There is a problem with simulation time since #1556 was merged.
I can see the problem when I use rmw_fastrtps_cpp as rmw_implementation and rclcpp from galactic branch.
The problem is that two different executors (main and local) are working with the same instance of guard condition simultaneously. But rmw_fastrtps_cpp does not support this. Only one of these executors will be process rcl_trigger_guard_condition() call, and second executor will not.
This leads to that TimeSource::clock_cb() is never been called in some scenarios (hardly reproduced on small examples)
let me confirm this problem. I think this is bug in rmw_fastrtps, if i am not mistaken. i think rmw_fastrtps should check and see if it needs to reattach the condition variable in the implementation.
Bug report
Hi. There is a problem with simulation time since #1556 was merged.
I can see the problem when I use rmw_fastrtps_cpp as rmw_implementation and rclcpp from galactic branch.
The problem is that two different executors (main and local) are working with the same instance of guard condition simultaneously. But rmw_fastrtps_cpp does not support this. Only one of these executors will be process rcl_trigger_guard_condition() call, and second executor will not.
This leads to that TimeSource::clock_cb() is never been called in some scenarios (hardly reproduced on small examples)
Required Info:
Steps to reproduce issue
Expected behavior
Actual behavior
Additional information
The text was updated successfully, but these errors were encountered: