Linting and header cleanup, part 1

[ayrton04]

Draft PR for now. Just want the build farm to test this. I have more commits a-comin'.

[ayrton04]

The random UKF test failures are odd. It passes every time locally. When it fails on the build farm now, it's always as if it's not getting any of the sensor measurements:

[test_ukf_localization_node_interfaces-2] /tmp/ws/src/robot_localization/test/test_ukf_localization_node_interfaces.cpp:309: Failure
[test_ukf_localization_node_interfaces-2] Expected: (::fabs(filtered_.pose.pose.position.x - pose.pose.pose.position.x)) < (0.1), actual: 20 vs 0.1
[test_ukf_localization_node_interfaces-2] /tmp/ws/src/robot_localization/test/test_ukf_localization_node_interfaces.cpp:315: Failure
[test_ukf_localization_node_interfaces-2] Expected: (::fabs(filtered_.pose.pose.position.z - pose.pose.pose.position.z)) < (0.1), actual: 40 vs 0.1
[ukf_node-1] [INFO] [1652184610.335751570] [test_ukf_localization_node_interfaces_ukf]: Received a request to reset filter.

robot_localization/test/test_ukf_localization_node_interfaces.cpp

Lines 309 to 317 in 215636a

	EXPECT_LT(
	::fabs(filtered_.pose.pose.position.x - pose.pose.pose.position.x),
	0.1);
	EXPECT_LT(
	::fabs(filtered_.pose.pose.position.y),
	0.1); // Configuration for this variable for this sensor is false
	EXPECT_LT(
	::fabs(filtered_.pose.pose.position.z - pose.pose.pose.position.z),
	0.1);

The only way we have that error at the end is if:

• The node wasn't receiving the sensor data
• The node wasn't publishing the filtered output
• The test sub wasn't receiving the filtered output

Does anyone know if ROS 2 has ever shown instability in pub and subs being correctly established? It seems unlikely, but maybe it's come up before.

[ayrton04]

I think I have a handle on why that's happening. Will address in a forthcoming PR.