Tests that run too long without output tends to break Travis builds. This is typically the case when running code-coverage, so we need to exclude a few test cases to ensure that the builds are stable.