testing-patterns
Testing Patterns & Strategies
Comprehensive guide to testing JavaScript/TypeScript applications. Complements the tdd skill (which focuses on workflow) by providing concrete patterns, strategies, and anti-patterns.
When to Use This Skill
Activate when the user:
- Wants to write tests for existing or new code
- Needs to establish a testing strategy for a project
- Wants to improve test coverage or reliability
- Has flaky or brittle tests
- Asks about mocking, stubbing, or test doubles
- Needs guidance on unit vs integration vs E2E tests
Testing Pyramid
/ E2E \ Few, slow, high confidence
/----------\
/ Integration \ Moderate count, medium speed
/----------------\
/ Unit \ Many, fast, focused
/____________________\
Rule of thumb: ~70% unit, ~20% integration, ~10% E2E. Adjust per project — legacy codebases often benefit from more integration tests initially.
1. Unit Testing Patterns
Test Structure: AAA (Arrange, Act, Assert)
describe('calculateDiscount', () => {
it('should apply 10% discount for orders above 100€', () => {
// Arrange
const order = { total: 150, items: ['item1', 'item2'] }
// Act
const result = calculateDiscount(order)
// Assert
expect(result).toBe(135)
})
})
Naming Convention
Use descriptive names that read as specifications:
// GOOD — describes behavior
it('should return 401 when token is expired')
it('should retry 3 times before failing')
it('should trim whitespace from patient name')
// BAD — describes implementation
it('should call validateToken')
it('should set retryCount to 3')
it('should use .trim()')
Parameterized Tests (Table-Driven)
describe('parseHL7Date', () => {
it.each([
['20260315', new Date(2026, 2, 15)],
['20260315120000', new Date(2026, 2, 15, 12, 0, 0)],
['', null],
['invalid', null],
])('should parse "%s" to %s', (input, expected) => {
expect(parseHL7Date(input)).toEqual(expected)
})
})
Testing Error Cases
Always test the unhappy path:
describe('PatientService.findById', () => {
it('should throw NotFoundError for unknown patient ID', async () => {
await expect(service.findById('UNKNOWN'))
.rejects.toThrow(NotFoundError)
})
it('should throw ValidationError for empty ID', async () => {
await expect(service.findById(''))
.rejects.toThrow(ValidationError)
})
})
Testing Async Code
// Promises
it('should resolve with patient data', async () => {
const patient = await fetchPatient('PAT123')
expect(patient.name).toBe('DUPONT')
})
// Event emitters
it('should emit "transfer" event on unit change', (done) => {
emitter.on('transfer', (data) => {
expect(data.unit).toBe('CARDIO')
done()
})
service.transferPatient('PAT123', 'CARDIO')
})
2. Integration Testing Patterns
Integration tests verify that components work together correctly.
Database Integration Tests
describe('PatientRepository', () => {
let db: TestDatabase
beforeAll(async () => {
db = await TestDatabase.create() // Real DB, not mock
await db.migrate()
})
afterEach(async () => {
await db.truncate() // Clean between tests
})
afterAll(async () => {
await db.destroy()
})
it('should persist and retrieve a patient', async () => {
const repo = new PatientRepository(db.connection)
await repo.create({
id: 'PAT123',
name: 'DUPONT',
birthDate: '1975-03-15',
})
const found = await repo.findById('PAT123')
expect(found).toMatchObject({
id: 'PAT123',
name: 'DUPONT',
})
})
})
API Integration Tests
describe('POST /api/patients', () => {
it('should create a patient and return 201', async () => {
const response = await request(app)
.post('/api/patients')
.send({ name: 'DUPONT', birthDate: '1975-03-15' })
.set('Authorization', `Bearer ${validToken}`)
expect(response.status).toBe(201)
expect(response.body.data.id).toBeDefined()
})
it('should return 400 for missing required fields', async () => {
const response = await request(app)
.post('/api/patients')
.send({}) // Missing name
.set('Authorization', `Bearer ${validToken}`)
expect(response.status).toBe(400)
expect(response.body.errors).toContainEqual(
expect.objectContaining({ field: 'name' })
)
})
it('should return 401 without authentication', async () => {
const response = await request(app)
.post('/api/patients')
.send({ name: 'DUPONT' })
expect(response.status).toBe(401)
})
})
Message Processing Integration Tests
Particularly relevant for healthcare message processing (HPK, HL7):
describe('HPK Message Pipeline', () => {
it('should parse, validate, and transform an ID|M1 message', async () => {
const rawMessage = 'ID|M1|C|HEXAGONE|20260122120000|USER001|PAT123|DUPONT|JEAN|19750315|M'
const result = await pipeline.process(rawMessage)
expect(result.type).toBe('ID')
expect(result.code).toBe('M1')
expect(result.patient.name).toBe('DUPONT')
expect(result.valid).toBe(true)
})
})
3. E2E Testing Patterns
E2E tests validate complete user workflows through the real application.
Page Object Pattern
class PatientListPage {
constructor(private page: Page) {}
async goto() {
await this.page.goto('/patients')
}
async search(query: string) {
await this.page.fill('[data-testid="search-input"]', query)
await this.page.click('[data-testid="search-button"]')
}
async getResults() {
return this.page.locator('[data-testid="patient-row"]').allTextContents()
}
async clickPatient(name: string) {
await this.page.click(`text=${name}`)
}
}
// Usage in test
test('should find patient by name', async ({ page }) => {
const patientList = new PatientListPage(page)
await patientList.goto()
await patientList.search('DUPONT')
const results = await patientList.getResults()
expect(results).toContain('DUPONT JEAN')
})
Data-testid Convention
Use data-testid attributes for test selectors, never CSS classes or element structure:
// GOOD — stable selector
await page.click('[data-testid="submit-admission"]')
// BAD — brittle, breaks on style/structure changes
await page.click('.btn.btn-primary.submit')
await page.click('form > div:nth-child(3) > button')
Visual Regression Testing
test('patient dashboard matches snapshot', async ({ page }) => {
await page.goto('/dashboard')
await page.waitForLoadState('networkidle')
await expect(page).toHaveScreenshot('dashboard.png', {
maxDiffPixelRatio: 0.01,
})
})
4. Mocking Strategies
When to Mock
| Mock | Don't Mock |
|---|---|
| External HTTP APIs | Your own database (use real test DB) |
| Time/Date (for determinism) | Internal collaborators between modules |
| File system (when impractical) | Simple utility functions |
| Third-party services (payment, email) | Data transformations |
| Environment-specific features | Business logic |
Mock vs Stub vs Spy
// STUB — returns a fixed value
const getPatient = vi.fn().mockResolvedValue({ id: 'PAT123', name: 'DUPONT' })
// SPY — observes calls without changing behavior
const logSpy = vi.spyOn(logger, 'info')
await service.admitPatient(data)
expect(logSpy).toHaveBeenCalledWith('Patient admitted', { id: 'PAT123' })
// MOCK — replaces the entire module
vi.mock('./emailService', () => ({
sendAdmissionNotification: vi.fn().mockResolvedValue(true),
}))
MSW for HTTP Mocking
Prefer MSW (Mock Service Worker) over manual fetch mocking:
import { http, HttpResponse } from 'msw'
import { setupServer } from 'msw/node'
const server = setupServer(
http.get('/api/patients/:id', ({ params }) => {
return HttpResponse.json({
id: params.id,
name: 'DUPONT',
birthDate: '1975-03-15',
})
})
)
beforeAll(() => server.listen())
afterEach(() => server.resetHandlers())
afterAll(() => server.close())
Time Mocking
describe('token expiration', () => {
beforeEach(() => {
vi.useFakeTimers()
})
afterEach(() => {
vi.useRealTimers()
})
it('should reject expired tokens', () => {
const token = createToken({ expiresIn: '1h' })
vi.advanceTimersByTime(2 * 60 * 60 * 1000) // 2 hours later
expect(() => validateToken(token)).toThrow('Token expired')
})
})
5. Test Organization
File Structure — Colocation
src/
├── patients/
│ ├── PatientService.ts
│ ├── PatientService.test.ts # Unit tests next to source
│ ├── PatientRepository.ts
│ └── PatientRepository.test.ts
├── __tests__/
│ └── integration/
│ ├── patient-admission.test.ts # Integration tests
│ └── hl7-message-flow.test.ts
└── e2e/
├── patient-workflow.spec.ts # E2E tests
└── pages/
└── PatientListPage.ts # Page objects
Test Utilities — Factories
Avoid copy-pasting test data. Use factories:
// test/factories/patient.ts
export function buildPatient(overrides: Partial<Patient> = {}): Patient {
return {
id: `PAT${Math.random().toString(36).slice(2, 8)}`,
name: 'DUPONT',
firstName: 'JEAN',
birthDate: '1975-03-15',
sex: 'M',
...overrides,
}
}
// Usage
it('should flag underage patients', () => {
const minor = buildPatient({ birthDate: '2015-06-01' })
expect(isMinor(minor)).toBe(true)
})
6. Common Anti-Patterns
Testing Implementation Details
// BAD — tests internal state
it('should set isLoading to true', () => {
component.fetchData()
expect(component.state.isLoading).toBe(true)
})
// GOOD — tests observable behavior
it('should show loading spinner while fetching', async () => {
render(<PatientList />)
expect(screen.getByTestId('spinner')).toBeVisible()
await waitForElementToBeRemoved(screen.queryByTestId('spinner'))
})
Excessive Mocking
// BAD — mocking everything, testing nothing real
it('should create patient', async () => {
vi.mock('./db')
vi.mock('./validator')
vi.mock('./logger')
// ... what are you even testing?
})
// GOOD — use real implementations, mock only boundaries
it('should create patient in database', async () => {
const repo = new PatientRepository(testDb)
const service = new PatientService(repo) // Real repo, real DB
const patient = await service.create({ name: 'DUPONT' })
expect(patient.id).toBeDefined()
})
Flaky Test Patterns
Common causes and fixes:
| Cause | Fix |
|---|---|
| Shared mutable state | Use beforeEach to reset state |
| Time-dependent tests | Use vi.useFakeTimers() |
| Race conditions in async tests | Use waitFor / findBy instead of getBy |
| Order-dependent tests | Each test must be independent |
| Network calls in unit tests | Mock HTTP with MSW |
| Hardcoded ports | Use dynamic port assignment |
Snapshot Overuse
// BAD — meaningless snapshot of entire component
it('should render', () => {
const { container } = render(<PatientCard patient={patient} />)
expect(container).toMatchSnapshot() // 200 lines of HTML nobody reads
})
// GOOD — targeted assertions
it('should display patient name and birthdate', () => {
render(<PatientCard patient={patient} />)
expect(screen.getByText('DUPONT JEAN')).toBeVisible()
expect(screen.getByText('15/03/1975')).toBeVisible()
})
7. Testing Checklist
Before submitting a PR, verify:
[ ] New feature has tests covering happy path and error cases
[ ] Tests are independent — can run in any order
[ ] No flaky tests introduced (run suite 3x to verify)
[ ] Test names describe behavior, not implementation
[ ] Mocks are limited to external boundaries
[ ] No hardcoded test data — using factories
[ ] Test utilities are shared, not duplicated
[ ] Integration tests clean up after themselves
[ ] E2E tests use data-testid selectors
[ ] Coverage hasn't decreased
Framework Quick Reference
| Framework | Best For | Command |
|---|---|---|
| Vitest | Unit + integration (Vite projects) | vitest run |
| Jest | Unit + integration (legacy, CRA) | jest --coverage |
| Playwright | E2E, visual regression | playwright test |
| Cypress | E2E, component testing | cypress run |
| Testing Library | DOM testing (React, Vue) | Used with Vitest/Jest |
| MSW | HTTP mocking | Used with any runner |
More from dedalus-erp-pas/hexagone-foundation-skills
vue-best-practices
Guide des bonnes pratiques Vue.js 3 couvrant la Composition API, la conception de composants, les patrons de réactivité, le styling utility-first avec Tailwind CSS, l'intégration native de la bibliothèque de composants PrimeVue et l'organisation du code. À utiliser lors de l'écriture, la revue ou le refactoring de code Vue.js pour garantir des patrons idiomatiques et un code maintenable.
23ubiquitous-language
Extrait un glossaire de langage ubiquitaire style DDD de la conversation en cours, signale les ambiguïtés et propose des termes canoniques. Sauvegarde dans UBIQUITOUS_LANGUAGE.md. À utiliser quand l'utilisateur veut définir des termes métier, construire un glossaire, durcir la terminologie, créer un langage ubiquitaire ou mentionne « domain model », « DDD », « glossaire » ou « langage ubiquitaire ».
23grill-me
Interroge l'utilisateur sans relâche sur un plan ou un design jusqu'à atteindre une compréhension partagée, en résolvant chaque branche de l'arbre de décision. À utiliser quand l'utilisateur veut stress-tester un plan, se faire challenger sur son design, ou mentionne « grill me » / « interroge-moi » / « challenge-moi » / « questionne-moi ».
22meeting
Lance une réunion simulée avec plusieurs personas experts pour analyser un sujet sous des perspectives diverses, prendre une décision et proposer une solution avant implémentation. Peut optionnellement publier l'analyse de la réunion sur une issue GitLab ou GitHub liée.
22changelog-generator
Crée automatiquement des changelogs orientés utilisateur à partir des commits git en analysant l'historique, catégorisant les changements et transformant les commits techniques en notes de version claires et compréhensibles. Transforme des heures de rédaction manuelle en minutes de génération automatisée.
22github-issues
Crée, récupère, met à jour et gère les issues GitHub avec collecte complète du contexte. À utiliser quand l'utilisateur veut créer une nouvelle issue, voir les détails d'une issue, mettre à jour des issues existantes, lister les issues du projet, ajouter des commentaires ou gérer les workflows d'issues dans GitHub.
22