Flaky: multithreaded comms tests hanging with connext on osx

[dhood]

This is the first time I've seen this: https://ci.ros2.org/job/nightly_osx_repeated/1045/testReport/junit/(root)/projectroot/gtest_multithreaded__rmw_connext_cpp/

That test uses a wait_for_subscriber call that underneath is waiting for graph changes so there's a chance this is caused by the same thing as ros2/rmw_connext#280.

[dhood]

Don't think it's caused by the wait_for_subscriber call because (1) on another build a different one of those tests hung after the wait_for_subscriber call: https://ci.ros2.org/job/ci_osx/3436/testReport/junit/(root)/projectroot/gtest_multithreaded__rmw_connext_cpp/ and (2) the implementation of wait_for_subscriber shouldn't hang because there's a low timeout being used.

[clalancette]

We haven't seen this in a while (over a week), so I'm going to close this out for now. If it pops back up we can reopen.