Expected behavior: I do a service request and it times out. If I then advertise a new service with the same name as the previous service request, I don't expect a request to be received from the previous service request.
Actual behavior: As of #470, INTEGRATION_triggered_publisher has been failing in gz-sim8. The issue is that service requests from a previous test are incorrectly received in the next test. It appears that no amount of sleep in between the tests helps. If the individual tests are run by themselves, they all pass.
Environment
gz-transport13Description
Expected behavior: I do a service request and it times out. If I then advertise a new service with the same name as the previous service request, I don't expect a request to be received from the previous service request.
Actual behavior: As of #470,
INTEGRATION_triggered_publisherhas been failing ingz-sim8. The issue is that service requests from a previous test are incorrectly received in the next test. It appears that no amount of sleep in between the tests helps. If the individual tests are run by themselves, they all pass.Steps to reproduce
Test in branch:
new_service_after_timeout