I've just always felt that if I have 100% coverage of each sub, I must be testing the return value. To boot, I always overlap tests so that it's often not just one test file that uses the return of a sub, so that if I break that return, I may end up with broken tests in several places, if not just several tests within a single file.
In reply to Re^4: Testing my tests
by stevieb
in thread Testing my tests (mutation testing)
by szabgab
| For: | Use: | ||
| & | & | ||
| < | < | ||
| > | > | ||
| [ | [ | ||
| ] | ] |