Closed arjo129 closed 3 years ago
@koonpeng would you be able to test this out on the web CI (without the transport_sleep
parameter) to see if this also provides an acceptable solution to the busy waiting problem?
If these changes fix the problem then we could revert the changes done here.
It won't be straight forward as the CI uses a nightly build on master. We can revert the hack job changes as the CI is not using it actually.
I've reverted the changes of #253 .
Eliminates the busy wait previously used in
Transport
by replacing it with a customrclcpp::Executor
.