I knew what you meant. Perhaps you don't work for a similar industry and with similar software. Different people work on different things, and that's fine.

Namely, running the same code over and over does not reproduce an external stimulus, and there is no mechanism for writing enough tests to cover even 50% of the possible things that may arise.

Anyhow, if you have actually written code that simulates network failures and firmware quirks and so forth, good for you, otherwise realize that while testing is a decent piece of things, real companies test against real hardware. Simply put, you do not know what kind of errors will come at you...

Moral of the story -- The problem of "when have I written enough tests" is about like the halting problem. Simply ensuring each line of code is executed 50 times proves nothing in this case.