Testing habits, with and without AI

📘What it really is

• Test the edges, where bugs live: empty lists, zero, one item, very large numbers, negative numbers, missing properties, text where a number is expected, Hindi text in a name.

• One bug, one test: when you find a bug, first write a test that fails because of it, then fix the code until it passes.

• Test-first (also called TDD): write the test before the function. It forces you to decide exactly what the function should do, which is useful thinking before asking AI to write it.

• Asking AI for tests: AI writes tests quickly, but check them:

• do they test behaviour you care about, or just repeat the code?
• do they include edge cases and bad input?

• would they actually fail if the code were wrong? (Break the code and see.)

• Never "fix" a failing test by changing what it expects unless you've decided the old expectation was wrong. And never let an AI delete or skip a failing test to make the run go green.

• Keep tests fast, so you'll actually run them after every change.