Test driven development

Test-Driven Development (TDD)

Overview

Write the test first. Watch it fail. Write minimal code to pass.

Core principle: If you didn't watch the test fail, you don't know if it tests the right thing.

Violating the letter of the rules is violating the spirit of the rules.

This skill is language-agnostic. Wherever it says "test runner", use your stack's tool (Jest, pytest, RSpec, go test, JUnit, cargo test, …). Wherever it shows pseudocode, translate to your language's testing idiom.

When to Use

Always:

• New features
• Bug fixes
• Refactoring
• Behavior changes

Exceptions (ask your human partner):

• Throwaway prototypes
• Generated code (scaffolds you immediately rewrite)
• Configuration files

Thinking "skip TDD just this once"? Stop. That's rationalization.

The Iron Law

NO PRODUCTION CODE WITHOUT A FAILING TEST FIRST

Write code before the test? Delete it. Start over.

No exceptions:

• Don't keep it as "reference"
• Don't "adapt" it while writing tests
• Don't look at it
• Delete means delete

Implement fresh from the tests. Period.

Red-Green-Refactor

digraph tdd_cycle {
    rankdir=LR;
    red [label="RED\nWrite failing test", shape=box, style=filled, fillcolor="#ffcccc"];
    verify_red [label="Verify fails\ncorrectly", shape=diamond];
    green [label="GREEN\nMinimal code", shape=box, style=filled, fillcolor="#ccffcc"];
    verify_green [label="Verify passes\nAll green", shape=diamond];
    refactor [label="REFACTOR\nClean up", shape=box, style=filled, fillcolor="#ccccff"];
    next [label="Next", shape=ellipse];

    red -> verify_red;
    verify_red -> green [label="yes"];
    verify_red -> red [label="wrong\nfailure"];
    green -> verify_green;
    verify_green -> refactor [label="yes"];
    verify_green -> green [label="no"];
    refactor -> verify_green [label="stay\ngreen"];
    verify_green -> next;
    next -> red;
}

RED - Write Failing Test

Write one minimal test showing what should happen.

<Good> ``` test "retries failed operations up to 3 times": attempts = 0 operation = () => { attempts += 1 if attempts < 3: throw Error("fail") return "success" }

result = retry(operation)

assert result == "success"
assert attempts == 3

Clear name, tests real behavior through the public API, one thing
</Good>

<Bad>

test "retry works": operation = mock() operation.stub_to_fail_twice_then_return("success")

retry(operation)

assert operation.was_called(3)   // only proves the mock was called

Vague name, tests the mock not the code
</Bad>

**Requirements:**

- One behavior per test
- Clear name that describes the behavior
- Real code and real collaborators (no mocks/stubs unless unavoidable)

### Verify RED - Watch It Fail

**MANDATORY. Never skip.**

Run the single new test with your test runner:

```bash
<test-runner> path/to/the_new_test    # e.g. pytest tests/test_retry.py, npx jest retry.test.ts

Confirm:

• Test fails (not errors)
• Failure message is the one you expected
• It fails because the feature is missing (not a typo, import error, or broken setup)

Test passes? You're testing existing behavior. Fix the test.

Test errors? Distinguish the two kinds: "the function/class doesn't exist yet" IS the expected first failure — that's fine. A syntax error or broken test setup is not. Fix the error, re-run until it fails correctly on the assertion.

GREEN - Minimal Code

Write the simplest code to pass the test.

<Good> ``` function retry(operation): for attempt in 1..3: try: return operation() catch error: if attempt == 3: rethrow ``` Just enough to pass </Good> <Bad> ``` function retry(operation, max_retries = 3, backoff = "exponential", on_retry = null, jitter = false): // YAGNI — options nobody asked for ``` Over-engineered </Bad>

Don't add features, refactor other code, or "improve" beyond the test.

Verify GREEN - Watch It Pass

MANDATORY.

Run the new test, then the whole suite.

Confirm:

• The new test passes
• All other tests still pass (run the full suite)
• Output pristine (no errors, no warnings, no leftover print/debug statements)

Test fails? Fix the code, not the test.

Other tests fail? Fix now.

REFACTOR - Clean Up

After green only:

• Remove duplication
• Improve names
• Extract helpers / functions / modules

Keep tests green. Don't add behavior.

Repeat

Next failing test for the next behavior.

Good Tests

Quality	Good	Bad
Minimal	One thing. "and" in the name? Split it.	test "validates email and domain and whitespace"
Clear	Name describes behavior	test "works" / test "test1"
Shows intent	Demonstrates the desired public API	Obscures what the code should do

Structure tests as arrange → act → assert. Use your framework's grouping and shared-setup facilities (describe/context blocks, fixtures, setup hooks) — but keep each test asserting one behavior.

Why Order Matters

"I'll write tests after to verify it works"

Tests written after code pass immediately. Passing immediately proves nothing:

• Might test the wrong thing
• Might test implementation, not behavior
• Might miss edge cases you forgot
• You never saw it catch the bug

Test-first forces you to see the test fail, proving it actually tests something.

"I already manually tested all the edge cases"

Manual testing is ad-hoc. You think you tested everything but:

• No record of what you tested
• Can't re-run when code changes
• Easy to forget cases under pressure
• "It worked in the console / on the page" ≠ comprehensive

Automated tests are systematic. The suite runs the same way every time.

"Deleting X hours of work is wasteful"

Sunk cost fallacy. The time is already gone. Your choice now:

• Delete and rewrite with TDD (X more hours, high confidence)
• Keep it and add tests after (30 min, low confidence, likely bugs)

The "waste" is keeping code you can't trust. Working code without real tests is technical debt.

"TDD is dogmatic, being pragmatic means adapting"

TDD IS pragmatic:

• Finds bugs before commit (faster than debugging after)
• Prevents regressions (tests catch breaks immediately)
• Documents behavior (tests show how to use the code)
• Enables refactoring (change freely, tests catch breaks)

"Pragmatic" shortcuts = debugging in production = slower.

"Tests after achieve the same goals - it's spirit not ritual"

No. Tests-after answer "What does this do?" Tests-first answer "What should this do?"

Tests-after are biased by your implementation. You test what you built, not what's required. You verify remembered edge cases, not discovered ones.

Tests-first force edge case discovery before implementing. Tests-after verify you remembered everything (you didn't).

30 minutes of tests after ≠ TDD. You get coverage, lose proof the tests work.

Common Rationalizations

Excuse	Reality
"Too simple to test"	Simple code breaks. The test takes 30 seconds.
"I'll test after"	Tests passing immediately prove nothing.
"Tests after achieve same goals"	Tests-after = "what does this do?" Tests-first = "what should this do?"
"Already manually tested"	Ad-hoc ≠ systematic. No record, can't re-run.
"Deleting X hours is wasteful"	Sunk cost fallacy. Keeping unverified code is technical debt.
"Keep as reference, write tests first"	You'll adapt it. That's testing after. Delete means delete.
"Need to explore first"	Fine. Throw away the spike, start with TDD.
"Test hard to write = skip it"	Listen to the test. Hard to test = hard to use. Fix the design.
"TDD will slow me down"	TDD is faster than debugging. Pragmatic = test-first.
"Manual test faster"	Manual doesn't prove edge cases. You'll re-test every change.
"Existing code has no tests"	You're improving it. Add tests for existing code.

Red Flags - STOP and Start Over

• Code before test
• Test after implementation
• Test passes immediately
• Can't explain why the test failed
• Tests added "later"
• Rationalizing "just this once"
• "I already manually tested it"
• "Tests after achieve the same purpose"
• "It's about spirit not ritual"
• "Keep as reference" or "adapt existing code"
• "Already spent X hours, deleting is wasteful"
• "TDD is dogmatic, I'm being pragmatic"
• "This is different because..."

All of these mean: Delete code. Start over with TDD.

Example: Bug Fix

Bug: Empty email accepted at registration

RED

test "rejects an empty email":
    response = POST /registrations, { email: "" }

    assert response.status == 422
    assert response.errors mention "email"

Verify RED

$ <test-runner> registrations_test
FAIL: expected status 422, got 200

GREEN

Add the validation — and only the validation:

function register(input):
    if input.email is blank:
        return validation_error("email is required")
    ...

Verify GREEN

$ <test-runner> registrations_test
PASS (1 test, 0 failures)

REFACTOR Extract shared validation logic if multiple fields need it. Stay green.

Verification Checklist

Before marking work complete:

• Every new code path has a test
• Watched each test fail before implementing
• Each test failed for the expected reason (behavior missing, not a typo)
• Wrote minimal code to pass each test
• Full test suite is GREEN
• Output pristine (no errors, no warnings, no stray print/debug statements)
• Tests use real code (real objects and data; mocks/stubs only at external boundaries)
• Edge cases and errors covered

Can't check all boxes? You skipped TDD. Start over.

Choosing the Test Layer

Match the test to the behavior — do not prove everything through end-to-end tests, and do not prove integration behavior with unit tests.

• Unit tests — pure logic, validations, invariants, data transformations. Fast, isolated, no I/O. The default home for most TDD cycles.
• Integration / API tests — behavior that crosses component boundaries: HTTP endpoints (status codes, response contract, authorization), database interactions, module wiring.
• End-to-end / UI tests — full user journeys through the running system in a real client/browser. Few, high-value flows only; everything provable at a lower layer belongs at that lower layer. Select UI elements by stable test identifiers (e.g. data-test attributes), never by visible text or DOM position.

Build test data with your stack's factory/fixture/builder mechanism inside the test — never assume pre-existing state.

When Stuck

Problem	Solution
Don't know how to test	Write the wished-for public call. Write the assertion first. Ask your human partner.
Test too complicated	Design too complicated. Simplify the interface / extract a smaller unit.
Must mock everything	Code too coupled. Inject the dependency (pass the collaborator in) instead.
Setup huge	Extract factories / builders / setup helpers. Still complex? Simplify the design.

Debugging Integration

Bug found? Write a failing test reproducing it. Follow the TDD cycle. The test proves the fix and prevents regression.

Never fix bugs without a test.

Testing Anti-Patterns

When adding mocks, stubs, or test utilities, read testing-anti-patterns.md to avoid common pitfalls:

• Testing mock behavior instead of real behavior (asserting "was called" where a state assertion suffices)
• Adding test-only methods to production code
• Over-stubbing without understanding dependencies
• Leaky shared state between tests

Final Rule

Production code → a test exists and failed first
Otherwise → not TDD

No exceptions without your human partner's permission.