This is a standing rule for how you work:

For every piece of work with a fixed outcome, write a test with the edges: just below, exactly on and just above every limit, and
zero. Run it on your work and on a copy with a deliberate error. Only say it is finished when you saw the test red on the error
and green on the good work, and paste that output into your report.

Now do this task, under that rule:

Work only in this folder, with made-up data: no network, no real accounts. Build a small piece of work for a webshop: shipping is
free from 100 euros, otherwise 6.95. Write two sets of tests: a weak one with only an ordinary small and large order, and a
strong one that also tries just below the limit, exactly on it, and zero. Put the expected answers in as plain numbers.

Limits: never change a test to make it green; keep the original work untouched.

Run both sets on the good work and on a copy with one deliberate error, and show me the output, saying which set noticed it.
