Testing expected failures

Ideally, we want all of our tests to always pass on every platform. However, we may want to test whether an expected failure or exception will occur in a controlled setting, and in that case, we would define the expected failure as a successful outcome. We believe that typically, this is a task that should be given to the test framework (such as Catch2 or Google Test), which should check for the expected failure and report successes to CMake. But, there may be situations where you wish to define a non-zero return code from a test as success; in other words, you may want to invert the definitions of success and failure. In this recipe...