shield-testing

SHIELD 2.0 Testing Guidelines

Pre-Merge Checklist

ALL checks must pass before merging to main:

1. npm run format - Run Prettier
2. npm run lint - Zero errors/warnings
3. npm test -- --run - All tests pass
4. npm run build - Successful build
5. Manual testing completed
6. Documentation updated

Test Requirements

Unit Tests

• Required for all new features
• Aim for >80% coverage on business logic
• Located next to source files (*.test.ts)

Test Structure

import { describe, it, expect, vi } from "vitest";

describe("ComponentName", () => {
  describe("methodName", () => {
    it("should do expected behavior", () => {
      // Arrange
      // Act
      // Assert
    });
  });
});

Mocking

// Mock Electron APIs
vi.mock("electron", () => ({
  ipcRenderer: {
    invoke: vi.fn(),
    on: vi.fn(),
  },
}));

// Mock window.electronAPI
Object.defineProperty(window, "electronAPI", {
  value: {
    llm: { loadModel: vi.fn() },
  },
});

CI/CD Pipeline

GitHub Actions runs on every PR:

1. Format Check: npx prettier --check "**/*.{ts,tsx,js,json,md,yml}"
2. Lint: npm run lint
3. Type Check: npx tsc --noEmit
4. Tests: npm test -- --run
5. Build: npm run build

Test Files Location

src/
├── App.test.ts
├── components/
│   └── chat/
│       └── chat.test.ts
electron/
└── services/
    ├── MCPService.test.ts
    └── AuditLogTypes.test.ts

Running Tests

# Run all tests
npm test -- --run

# Watch mode
npm test

# Specific file
npm test -- src/App.test.ts

# With coverage
npm test -- --coverage

What to Test

• Business logic functions
• State store actions
• IPC handlers
• Utility functions
• Component rendering (basic)

What NOT to Test

• Third-party library internals
• Electron main process directly
• LLM inference (mock it)