The curriculum

Eight tracks on what a test is actually for. Every lesson runs its examples through the same engine the exercises use, so nothing here can demonstrate something the tool does not really do.

0 of 10 finished

What a test is for

The two jobs a suite has, and why most suites only do the first.

The two jobsA suite proves the code works now, and tells you when you break it later. Most only do the first.6 min

Coverage and its limits

What the number measures, what it cannot, and the suite that proves it.

Assertions

Asserting on values rather than on truthiness, and the traps in between.

Boundaries

Empty, one, two, many — and the off-by-one that lives between them.

Empty, one, two, manyFour cases that between them catch a startling share of real defects.8 min

Test doubles

The four kinds, when each is right, and how a double drifts from the real thing.

The double that driftedStub, mock, fake, spy — and the failure mode all four share.8 min

Properties

Invariants instead of examples, and what a shrunk counterexample tells you.

Invariants over examplesAn example tests one input. A property tests the rule, and the generator finds the input you would not have chosen.9 min

Determinism

Clock, randomness, ordering — and why a retried test is a deleted test.

A retried test is a deleted testWhere non-determinism comes from, and why re-running until green is the worst available response.8 min

Reviewing a generated suite

The thesis track. Score a suite before you read it, then check.

Score it before you read itThe thesis exercise. Guess how good a generated suite is, then run the board and find out.10 min