Test fix: don't keep reference to the controller in the test when it should be destroyed in the controller manager

[sgmurray]

This fixes #827

[sgmurray]

The following error message occurs when a node is created with the same name as a node that already exists.

`[WARN]`` [1664625270.031583774] [rcl.logging_rosout]: Publisher already registered for provided node name. If this is due to multiple nodes with the same name then all logs for that logger name will go out over the existing publisher. As soon as any node with that name is destructed it will unregister the publisher, preventing any further logs for that name from being published on the rosout topic.

With regards to this testcase,
auto ctrl_1 = cm_->get_loaded_controllers()[0]; creates a new ControllerSpec object which contains a shared pointer to the controller's ControllerInterfaceBase, Consequently, even when the controller manager unloads the controller. if ctrl_1 is still in scope then then the ControllerInterfaceBase is not destroyed. Because the ControllerInterfaceBase is not destroyed and ControllerInterfaceBase contains a shared pointer to the node, the node is also not destroyed.

[destogl]

Thanks for doing a deep dive on this!

With regards to this testcase, auto ctrl_1 = cm_->get_loaded_controllers()[0]; creates a new ControllerSpec object which contains a shared pointer to the controller's ControllerInterfaceBase, Consequently, even when the controller manager unloads the controller. if ctrl_1 is still in scope then then the ControllerInterfaceBase is not destroyed. Because the ControllerInterfaceBase is not destroyed and ControllerInterfaceBase contains a shared pointer to the node, the node is also not destroyed.

I understand this explanation, but don't understand what this has to do with this test. Where are we destroying controller, exactly?

Sorry, I see it now at the end. Cool, this looks good, thanks!

[bmagyar]

Any comment on the failing test (the one modified by this PR) ?

[sgmurray]

Any comment on the failing test (the one modified by this PR) ?

I believe that the failing test is TestLoadController.spawner_test_type_in_arg , the test modified in this PR is TestLoadController.spawner_test_type_in_param.

[bmagyar]

The tests prove your point, thanks! :D