Flaky tests

[bmagyar]

    [ackermann_steering_controller.rosunit-ackermann_steering_controller_wheel_radius_param_test/testTurn][passed]
    [ackermann_steering_controller.rosunit-ackermann_steering_controller_wheel_radius_param_test/testOdomFrame][passed]

    SUMMARY
    [NON-XML-CHAR-0x1B][1;31m * RESULT: FAIL[NON-XML-CHAR-0x1B][0m
     * TESTS: 3
     * ERRORS: 0
    [NON-XML-CHAR-0x1B][1;31m * FAILURES: 1[NON-XML-CHAR-0x1B][0m

    rostest log file is in /root/.ros/log/rostest-570fe3c46d99-23237.log
    -- run_tests.py: execute commands
      /usr/bin/python3 /opt/ros/noetic/share/rostest/cmake/../../../bin/rostest --pkgdir=/root/target_ws/src/ros_controllers/ackermann_steering_controller --package=ackermann_steering_controller --results-filename test_ackermann_steering_controller_radius_param_test_ackermann_steering_controller_radius_param.xml --results-base-dir "/root/target_ws/build/ackermann_steering_controller/test_results" /root/target_ws/src/ros_controllers/ackermann_steering_controller/test/ackermann_steering_controller_radius_param_test/ackermann_steering_controller_radius_param.test 
    -- run_tests.py: verify result "/root/target_ws/build/ackermann_steering_controller/test_results/ackermann_steering_controller/rostest-test_ackermann_steering_controller_radius_param_test_ackermann_steering_controller_radius_param.xml"
  >>>
build/ackermann_steering_controller/test_results/ackermann_steering_controller/rosunit-ackermann_steering_controller_wheel_radius_param_test.xml: 3 tests, 0 errors, 1 failure, 0 skipped
- AckermannSteeringControllerTest testForward
  <<< failure message
    /root/target_ws/src/ros_controllers/ackermann_steering_controller/test/ackermann_steering_controller_test/ackermann_steering_controller_test.cpp:73
    The difference between fabs(new_odom.twist.twist.linear.x) and cmd_vel.linear.x is 0.02165762085097149, which exceeds EPS, where
    fabs(new_odom.twist.twist.linear.x) evaluates to 0.1216576208509715,
    cmd_vel.linear.x evaluates to 0.10000000000000001, and
    EPS evaluates to 0.01.
  >>>

@matthew-reynolds have you worked on a similar fix for the diff drive by any chance?

[matthew-reynolds]

I ran into these flakey tests when testing the new CI, I've been poking around at it but don't have anything solid yet. I'll keep investigating it

[matthew-reynolds]

Since we're adding this repo as a downstream test for many of our other repos, I'll put a bunch of work in this week to try to iron out these flakey tests. It will be a pain to have failures popping up all over the place due to a couple flakey tests in diff drive and ackermann drive controllers. Currently looking into resetting the odom accumulators as a potential source of some failures...