Triggered jobs 324-331. Update: only one failed, new commit to extend that timeout.

Triggered jobs 332-339. Update: 338 failed, but I can't find any "RESULT: FAIL" in the output...

@kmactavish If it's of any help: It looks like the failing test is a nosetest added( using catkin_add_nosetests) and not a rostest.
Nosetests output results as "OK" or FAILED (failures=1) or FAILED (errors=1). The "SUMMARY" is part of rostest I believe, and thus not printed by "pure" nose of gtest tests.

Build 345 failed (not this branch). It was in raceConditionCallback, which is flawed.
Explanation (main is the main thread, add is the spawned thread):

• main gets to line 471 (queue is cleared)
• add sees sync is true, does its thing (setting sync to false, adding the callback to the queue)
• main sets sync back to true! and calls the callback (setting one to false)
• add sees sync is true, does its thing (setting one to true)
• main wakes up, and having called the callback, checks to make sure one is false. It's not!
  So, ironically, there was a secondary, unintentional race condition in the race condition check 😱.

Fixed by only setting the sync flag in one thread. It means adding another check to see if the queue is empty in the second thread.

@mikaelarguedas Thanks! That helped.

Build 355 (not this branch) also failed due to the compression ratio test thresholds, but in the other direction, relaxed them more.

@esteve, I had to change your raceConditionCallback test (see above). You might want to have a look and make sure it still tests for the right things after my modifications.

Thank you for improving the stability of the tests.