The problem is really that the failing tests are not *unit* tests. They ought to be classified better.
The problem is really that the failing tests are not *unit* tests. They ought to be classified better.