Question

A developer has written a `calculateShippingCost(weightKg)` function that charges different rates across weight tiers: 0–5 kg, 5.01–20 kg, and above 20 kg. They select the function and ask Copilot Chat to generate unit tests. Copilot returns tests that check one typical value in each tier (e.g., 2 kg, 10 kg, 30 kg) but the developer wants stronger coverage. What is the most effective follow-up prompt to improve the test suite's ability to catch tier-boundary defects?

Accepted Answer

Ask Copilot to add tests for the exact boundary values and just-inside/just-outside cases (e.g., 0, 5.00, 5.01, 20.00, 20.01, and a negative or zero weight) with explicit expected results.

Answer

Ask Copilot to increase the number of random typical values tested within each tier to at least 20 per tier.

Answer

Ask Copilot to add a test that runs the function 1000 times in a loop to measure average execution time.

Answer

Ask Copilot to refactor the function into three separate functions so each tier can be tested in isolation.