[romanovcode]

> too many people would be affected by such regression.

Yes, too many people do stupid "hello world" tests indeed.

Maybe this is a problem with running "hello world" tests and not that much of a real-world problem. Let's see.