test-driven-development（测试驱动开发）

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.

Reference Files

Read only the references needed for the current test cycle:

• references/testing-patterns.md for naming, AAA structure, near-miss negatives, behavioral focus, and anti-pattern checks
• references/test-data-and-mocks.md for factories, mock boundaries, common boundary mocks, and env/time handling
• references/integration-and-live-proof.md when unit tests are not enough, or the plan requires real APIs, seeded data, browser flows, or stress proof

When to Use

Always:

• New features
• Bug fixes
• Refactoring
• Behavior changes

Exceptions (ask your human partner):

• Throwaway prototypes
• Generated code
• 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 tests. Period.

Test Process Discipline (CRITICAL)

Problem: Test runners (Vitest, Jest) default to watch mode, leaving processes hanging indefinitely.

Mandatory Rules:

1. Always use run mode — Never invoke watch mode:

- Vitest: npx vitest run (NOT npx vitest) - Jest: CI=true npx jest or npx jest --watchAll=false - npm scripts: CI=true npm test or npm test -- --run

1. Prefer CI=true prefix for all test commands: CI=true npm test
2. After TDD cycle complete, verify no orphaned processes: pgrep -f "vitest|jest" || echo "Clean"
3. Kill if found: pkill -f "vitest" 2>/dev/null || true

Red-Green-Refactor

    ┌─────────┐       ┌─────────┐       ┌───────────┐
    │   RED   │──────>│  GREEN  │──────>│ REFACTOR  │
    │ (Fail)  │       │ (Pass)  │       │ (Clean)   │
    └─────────┘       └─────────┘       └───────────┘
         ^                                    │
         │                                    │
         └────────────────────────────────────┘
                    Next Feature

Vertical Slicing (CRITICAL)

WRONG (horizontal — all tests then all code):
  RED:   test1, test2, test3, test4, test5
  GREEN: impl1, impl2, impl3, impl4, impl5

RIGHT (vertical — one feature at a time):
  RED->GREEN: test1->impl1
  RED->GREEN: test2->impl2
  RED->GREEN: test3->impl3

DO NOT write all tests first, then all implementation. This produces bad tests:

• Tests written in bulk test *imagined* behavior, not *actual* behavior
• You end up testing the *shape* of things rather than user-facing behavior
• Tests become insensitive to real changes — pass when behavior breaks, fail when behavior is fine
• You outrun your headlights, committing to test structure before understanding the implementation

Correct approach: One test → one implementation → repeat. Each test responds to what you learned from the previous cycle.

RED - Write Failing Test

Write one minimal test showing what should happen.

Good:

test('retries failed operations 3 times', async () => {
  let attempts = 0;
  const operation = () => {
    attempts++;
    if (attempts < 3) throw new Error('fail');
    return 'success';
  };

  const result = await retryOperation(operation);

  expect(result).toBe('success');
  expect(attempts).toBe(3);
});

Clear name, tests real behavior, one thing

Bad:

test('retry works', async () => {
  const mock = jest.fn()
    .mockRejectedValueOnce(new Error())
    .mockRejectedValueOnce(new Error())
    .mockResolvedValueOnce('success');
  await retryOperation(mock);
  expect(mock).toHaveBeenCalledTimes(3);
});

Vague name, tests mock not code

Requirements:

• One behavior
• Clear name
• Real code (no mocks unless unavoidable)

Verify RED - Watch It Fail

MANDATORY. Never skip.

CI=true npm test path/to/test.test.ts

Confirm:

• Test fails (not errors)
• Failure message is expected
• Fails because feature missing (not typos)

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

Test errors? Fix error, re-run until it fails correctly.

GREEN - Minimal Code

Write simplest code to pass the test.

Good:

async function retryOperation<T>(fn: () => Promise<T>): Promise<T> {
  for (let i = 0; i < 3; i++) {
    try {
      return await fn();
    } catch (e) {
      if (i === 2) throw e;
    }
  }
  throw new Error('unreachable');
}

Just enough to pass

Bad:

async function retryOperation<T>(
  fn: () => Promise<T>,
  options?: {
    maxRetries?: number;
    backoff?: 'linear' | 'exponential';
    onRetry?: (attempt: number) => void;
  }
): Promise<T> {
  // YAGNI - You Ain't Gonna Need It
}

Over-engineered

Don't add features, refactor other code, or "improve" beyond the test. Don't hard-code test values - implement general logic that works for ALL inputs.

Verify GREEN - Watch It Pass

MANDATORY.

CI=true npm test path/to/test.test.ts

Confirm:

• Test passes
• Other tests still pass
• Output pristine (no errors, warnings)

Test fails? Fix code, not test.

Other tests fail? Fix now.

REFACTOR - Clean Up

After green only:

• Remove duplication
• Improve names
• Extract helpers

Keep tests green. Don't add behavior.

Repeat

Next failing test for next feature.

Good Tests

For deeper test structure, near-miss negative tests, behavior-vs-internals, and smell checks, read references/testing-patterns.md.

For factories, mocks, and env/time handling, read references/test-data-and-mocks.md.

Why Order Matters

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

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

• Might test 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 when I tried it" ≠ comprehensive

Automated tests are systematic. They run 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.

Red Flags - STOP and Start Over

If you catch yourself:

• Code before test
• Test after implementation
• Test passes immediately
• Can't explain why 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.

Rationalization Prevention

Example: Bug Fix

Bug: Empty email accepted

RED

test('rejects empty email', async () => {
  const result = await submitForm({ email: '' });
  expect(result.error).toBe('Email required');
});

Verify RED

$ npm test
FAIL: expected 'Email required', got undefined

GREEN

function submitForm(data: FormData) {
  if (!data.email?.trim()) {
    return { error: 'Email required' };
  }
  // ...
}

Verify GREEN

$ npm test
PASS

REFACTOR Extract validation for multiple fields if needed.

Verification Checklist

Before marking work complete:

• Every new function/method has a test
• Watched each test fail before implementing
• Each test failed for expected reason (feature missing, not typo)
• Wrote minimal code to pass each test
• All tests pass
• Output pristine (no errors, warnings)
• Tests use real code (mocks only if unavoidable)
• Edge cases and errors covered
• No hanging test processes (pgrep -f "vitest|jest" returns empty)

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

Integration And Live Proof

When the accepted plan or risk profile goes beyond local behavior, read references/integration-and-live-proof.md.

Unit tests are not enough when the task depends on:

• real API calls
• seeded or resettable data
• browser or worker orchestration
• cross-service side effects
• load or stress behavior

In those cases, keep TDD for the inner loop and escalate verification depth for the outer proof.

Coverage Threshold (Project Default)

Target: 80%+ code coverage across:

• Branches: 80%
• Functions: 80%
• Lines: 80%
• Statements: 80%

Verify with: npm run test:coverage or equivalent.

Below threshold? Add missing tests before claiming completion.

Test Prioritization

80% coverage means deliberate choices about what to test first. Focus effort on:

• Critical user-facing paths (auth, payments, data integrity)
• Complex logic with multiple branches
• Code that has broken before (regression-prone areas)

Do NOT skip tests because code "looks simple" — simple code breaks too. The 80% target is a floor, not a ceiling.

When Stuck

Design for Testability (When Tests Are Hard)

If tests are hard to write, the interface needs work:

1. Accept dependencies, don't create them

- Testable: function processOrder(order, paymentGateway) {} - Hard to test: function processOrder(order) {const gw = new StripeGateway();}

1. Return results, don't produce side effects

- Testable: function calculateDiscount(cart): Discount {} - Hard to test: function applyDiscount(cart): void {cart.total -= discount;}

1. Small surface area — fewer methods = fewer tests needed, fewer params = simpler setup

Behavioral Focus

Test how objects collaborate, not what they contain. If a test inspects .state, .length, or private fields, it is testing structure — and will break when internals change without behavior changing.

This is already implied by the "Testing implementation" smell in the Test Smells table. Make it the default lens: every assertion should answer "what did the user/caller observe?" not "what happened inside?"

Test Contracts Across Agents

When CC10x routes work across multiple agents (planner writes test specs, builder implements, reviewer verifies), the test file IS the contract:

• Planner defines expected behavior as test names and assertions in the plan
• Builder writes tests first, implements to green — the test file proves the contract is met
• Reviewer re-runs the same tests — pass means contract fulfilled, fail means contract broken

Do not duplicate the contract in prose. If the test file expresses the requirement, the test file is the requirement.

Output Format

## TDD Cycle

### Requirements
[What functionality is being built]

### RED Phase
- Test: [test name]
- Command: `npm test -- --grep "test name"`
- Result: exit 1 (FAIL as expected)
- Failure reason: [function not defined / expected X got Y]

### GREEN Phase
- Implementation: [summary]
- File: [path:line]
- Command: `npm test -- --grep "test name"`
- Result: exit 0 (PASS)

### REFACTOR Phase
- Changes: [what was improved]
- Command: `npm test`
- Result: exit 0 (all tests pass)

Final Rule

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

No exceptions without your human partner's permission.