I have to add custom code to detect any failed tests in the results so I can fail the build.
If there was a command line arg to basically use the test failure to also mean the command failed. This can wait until everything is cleaned up, but should still return a non 0 exit code if my tests failed.