On: An overly aggressive mock can work fine, but break much later

[tayo42]

Tests help make sure that the intent is to just bubble up the error.

What if someone in the future is comes in and modifies that to add some complexity somehow or changes it to log and continue. Tests will catch that behavior

[bccdee → tayo42]

It's rarely possible to future-proof a function by testing its entire range of possible inputs, so I'm not particularly concerned about trying to anticipate trivial cases that might become non-trivial in future—that just feels like a special case of future-proofing.

I think it's more important to ensure that, after each new commit, every non-trivial code path in that function is exercised. If someone adds new logic, they're responsible for adding new tests to cover that logic; similarly, if they turn a trivial code path into a non-trivial one, they're responsible for adding tests to cover that too.