Testing

Testing Skill

Testing Pyramid

The testing pyramid represents the ideal distribution of tests:

        /\
       /  \      E2E Tests (few, slow, expensive)
      /____\
     /      \    Integration Tests (some, medium speed)
    /________\
   /          \  Unit Tests (many, fast, cheap)
  /____________\

Unit Tests

• What: Test individual functions/methods in isolation
• Speed: Fast (milliseconds)
• Quantity: Many (80%+ of tests)
• Scope: Single function/class
• Dependencies: Mocked/stubbed
• Example: Test calculateTotal() with various inputs

Integration Tests

• What: Test interactions between components/modules
• Speed: Medium (seconds)
• Quantity: Some (15% of tests)
• Scope: Multiple components working together
• Dependencies: Real databases, APIs, or services (may use test doubles)
• Example: Test API endpoint with database, test service layer with repository

End-to-End (E2E) Tests

• What: Test complete user workflows
• Speed: Slow (minutes)
• Quantity: Few (5% of tests)
• Scope: Entire system
• Dependencies: Real infrastructure (or close to it)
• Example: Test user registration → login → purchase flow

Test Naming Conventions

Structure: should [expected behavior] when [condition]

Good Examples:

• should return 0 when given empty array
• should throw error when user is not authenticated
• should calculate tax correctly when order total exceeds 100
• should send email when user signs up

Bad Examples:

• test1, testCalculate, testUser (too vague)
• should work (doesn't describe behavior)
• test_calculate_total_with_items (missing "should" structure)

Language-Specific Patterns

JavaScript/TypeScript (Jest):

describe('calculateTotal', () => {
  it('should return 0 when given empty array', () => {
    // ...
  });
  
  it('should sum all item prices correctly', () => {
    // ...
  });
});

Python (pytest):

def test_should_return_zero_when_given_empty_array():
    # ...

def test_should_sum_all_item_prices_correctly():
    # ...