using-tests（使用测试）

Working with Tests

Testing strategy and workflow. Tests run in parallel with isolated data per suite. Prioritize Playwright for UI, integration tests for APIs, unit tests for logic.

Testing Strategy

Follow this hierarchy when deciding what kind of test to write:

1. Playwright tests (browser) - Preferred for most features
2. Integration tests (API) - When Playwright is not practical
3. Unit tests (pure functions) - Only for complex isolated logic

When to Use Each Test Type

Playwright Tests (Default Choice)

Write Playwright tests when the feature involves:

• User interactions (clicking, typing, navigation)
• Visual feedback (toasts, loading states, error messages)
• Form submissions and validation
• Multi-step UI flows
• Protected routes and redirects
• Accessibility behavior

Example features best tested with Playwright:

• Sign-in flow with error handling
• Chat creation and deletion with confirmation dialogs
• Theme toggle
• Form validation messages
• Navigation between pages

Integration Tests

Write integration tests when:

• Testing API responses directly (status codes, JSON structure)
• Verifying database state after operations
• Testing server-side logic without UI
• Playwright would be too slow or complex for the scenario

Example features best tested with integration tests:

• API route returns correct status codes
• User creation populates database correctly
• Session cookies are set on sign-in
• Protected API routes return 401/403

Unit Tests

Write unit tests only when:

• Testing pure functions with complex logic
• Testing code with many edge cases
• Testing type narrowing and error messages
• The function has no external dependencies

Example features best tested with unit tests:

• Assertion helpers
• Config schema validation
• Data transformation functions
• Utility functions

Running Tests

All tests run against an isolated Neon database branch that auto-deletes after 1 hour.

bun run test              # All tests with isolated Neon branch
bun run test:playwright   # Browser tests only
bun run test:integration  # Integration tests only
bun run test:unit         # Unit tests only

Folder Structure

src/
├── lib/
│   ├── common/
│   │   ├── assert.ts
│   │   └── assert.test.ts      # Unit test (co-located)
│   └── config/
│       ├── schema.ts
│       └── schema.test.ts      # Unit test (co-located)
tests/
├── integration/
│   ├── llms.test.ts            # Integration test
│   ├── r.test.ts
│   ├── mcp/
│   │   └── route.test.ts
│   └── recipes/
│       └── [slug]/
│           └── route.test.ts
└── playwright/
    ├── auth.spec.ts            # Playwright test
    ├── chat.spec.ts
    ├── home.spec.ts
    └── lib/
        └── test-user.ts        # Playwright-specific helpers

Writing Tests for New Features

Step 1: Determine Test Type

Ask: "How would a user verify this feature works?"

• If through the UI → Playwright test
• If through API calls → Integration test
• If by calling a function directly → Unit test

Step 2: Create Test File

Playwright tests: tests/playwright/{feature}.spec.ts

import { test, expect } from "@playwright/test";

test.describe("Feature Name", () => {
  test("should do expected behavior", async ({ page }) => {
    await page.goto("/feature");
    // Test implementation
  });
});

Integration tests: tests/integration/{feature}.test.ts

For API routes, import the handler directly for faster, more reliable tests:

import { describe, it, expect } from "bun:test";
import { GET } from "@/app/api/feature/route";

describe("GET /api/feature", () => {
  it("should return expected response", async () => {
    const response = await GET();

    expect(response.status).toBe(200);
    const data = await response.json();
    expect(data.value).toBeDefined();
  });
});

Unit tests: src/lib/{domain}/{file}.test.ts (co-located)

import { describe, it, expect } from "bun:test";
import { myFunction } from "./my-file";

describe("myFunction", () => {
  it("should do expected behavior", () => {
    expect(myFunction()).toBe("expected");
  });
});

Test Data Management

Database Isolation

Tests run against isolated Neon branches. Each test run:

1. Creates a fresh schema-only branch
2. Runs tests against the branch
3. Branch auto-deletes after 1 hour

This ensures tests don't interfere with production data.

Parallel Test Isolation

Tests run in parallel by default. Each test suite must use its own test data to avoid conflicts:

• Different test users - Each spec file should create unique users with distinct emails
• Different resources - Tests creating chats, sessions, etc. should not depend on shared state
• No cleanup required - The branch TTL handles cleanup automatically

// auth.spec.ts - uses auth-specific test user
const testUser = await createTestUser({
  email: `auth-test-${uuid}@example.com`,
});

// chat.spec.ts - uses chat-specific test user
const testUser = await createTestUser({
  email: `chat-test-${uuid}@example.com`,
});

Avoid patterns that rely on global state or specific database contents existing from other tests.

Common Patterns

Testing Protected Routes (Playwright)

test("should redirect unauthenticated user", async ({ page }) => {
  await page.goto("/protected-page");
  await expect(page).toHaveURL(/sign-in/);
});

Testing Error States (Playwright)

test("should show error for invalid input", async ({ page }) => {
  await page.goto("/form");
  await page.getByRole("button", { name: /submit/i }).click();

  await expect(page.getByText(/error|required/i)).toBeVisible({
    timeout: 5000,
  });
});

Testing API Responses (Integration)

Import route handlers directly for cleaner tests:

import { GET } from "@/app/api/endpoint/route";

it("should return 200 for valid request", async () => {
  const response = await GET();
  expect(response.status).toBe(200);
});

Debugging Failed Tests

Playwright

bunx playwright test --headed              # Watch browser
bunx playwright test --debug               # Step through test
bunx playwright show-report                # View HTML report

Integration/Unit

bun test --only "test name"                # Run single test
bun test --watch                           # Re-run on changes

View test artifacts

Failed Playwright tests save screenshots and traces to test-results/. Check this folder when CI fails.