The RunCMake infrastructure is more well-suited to check the results of each case. It can also report more precisely what went wrong when a case fails.