Currently it returns false if it fails, which is error prone as it means the test author needs to remember to check the return value or re-check the condition after calling it. Making it throw on error makes the behavior more obvious and harder to mis-use, at the cost of possibly getting less useful error messages in some cases.
Currently it returns false if it fails, which is error prone as it means the test author needs to remember to check the return value or re-check the condition after calling it. Making it throw on error makes the behavior more obvious and harder to mis-use, at the cost of possibly getting less useful error messages in some cases.