TDD Red Phase - Write Failing Tests First

Focus on writing clear, specific and failing unit tests from Gherkin (BDD) feature files that describe the desired behaviour before any implementation exists.

Gherkin (BDD) Overview

Simple Gherkin example:

Feature: Authentication

  Scenario: Successful login with valid credentials
    Given a user with username "user" and password "password"
    When the signs in with username "user" and password "password"
    Then the user should be signed in successfully

Feature-to-Test Mapping

Translate above Gherkin scenario into a unit test. Note the newlines between Given, When, Then sections for clarity.

Example in TypeScript:

let userRepository: FakeUserRepository;
let client: Client;

test('signs in with valid credentials', () => {
    const user = new User("user", "password");
    userRepository.add(user);

    const response = await client.post("/login", {
      username: user.username,
      password: user.password
    });

    expect(response.status_code).toBe(HttpStatus.OK);
});

Example in Python:

def test_successful_login(user_repository: FakeUserRepository, client: Client):
    user = User(username="user", password="password")
    user_repository.add(user)

    response = client.post("/login", data={
      "username": user.username,
      "password": user.password
    })

    assert response.status_code == HttpStatus.OK

Feature Context Analysis

• •Fetch feature details read and study the prompted files matching glob pattern feature/*.feature files comprehensively
• •Understand the full context from feature scenario description, preconditions (Given), actions (When), and expected outcomes (Then)
• •Requirements extraction - Parse the scenario acceptance criteria
• •Edge case identification - Review feature scenarios for boundary conditions

Core Principles

Test-First Mindset

• •Write the test before the code - Never write production code without a failing test.
• •One test at a time - Focus on a single behaviour or requirement from the feature, never write more than one test at once.
• •Fail for the right reason - Ensure tests fail due to a missing implementation, not syntax errors.
• •Be specific - Tests should clearly express what behaviour is expected per feature requirements.

Test Quality Standards

• •

  Descriptive test names - Use clear, behaviour-focused test naming:

  ts

  describe('User', () => {
      test('can sign in with valid credentials', () => {
          // Test implementation here
      });
  });
  

• •

  Given-When-Then Pattern - Structure tests with clear Given, When, Then sections without denoting these explicitly with comments. Use newlines to separate sections.

• •

  Assertion focus - Each test should verify the specific outcomes from Then & And keywords.

Test Library Patterns

• •Use readable assertions with strict matching (e.g. expect().toBe() and expect().toStrictEqual()) as much as possible.
• •Use custom fixture functions to provide faked test data
• •Use test.each for generating data-driven tests.

Test Isolation & Independence

• •Self-contained setup and teardown - Each test manages its own state.
• •Avoid shared mutable state - Never share data between tests.

Mock/Stub Strategy

• •Prefer real objects when possible - For side effect free logic, always use actual implementations.
• •Fake external dependencies only - Substitute access to databases, APIs, file systems, and third-party services with faked test doubles.
• •Keep test doubles simple - Only fake what's necessary for the test scenario. The less fake code needed, the better.

Test Data Management

• •Fixture factories - Use factory patterns for complex object creation from Given clauses.
• •Builder patterns - Consider builders for objects with many optional fields.
• •Data proximity - Keep test data close to tests for readability.
• •Shared fixtures - Place common fixtures in own module such as fixtures.ts or fixtures.py.
• •Shared preconditions - Convert Gherkin Background sections to fixture functions.

Scenario Priority & Ordering

1. •Cover happy paths first - Ensure successful scenarios from feature files come first.
2. •Error scenarios next - Handle exception and error conditions.
3. •Edge cases last - Cover boundary behaviors and unusual inputs.

Scenario Outlines and Examples

• •Parametrized tests - Map Scenario Outline with Examples to test.each.
• •Match placeholders - Ensure parameter names match Given/When/Then variables.
• •Clear test IDs - Use meaningful test IDs for each example case.

Example in TypeScript:

test.each([
    ["valid_user", "valid_pass", HttpStatus.OK],
    ["invalid_user", "wrong_pass", HttpStatus.UNAUTHORIZED],
])("login scenarios for %s", (username: string, password: string, expected: HttpStatus) => {
    // Implement test logic here
});

Example in Python:

@pytest.mark.parametrize(("username", "password", "expected"), [
    ("valid_user", "valid_pass", HttpStatus.OK),
    ("invalid_user", "wrong_pass", HttpStatus.UNAUTHORIZED),
])
def test_login_scenarios(username: str, password: str, expected: HttpStatus):
    # Implement test logic here

Error Message Quality

• •Clear failure messages - Assertions should explain what specific behavior failed.
• •Reference requirements - Error messages should cite feature file expectations.
• •Custom assertion messages - Use expect(fn).toThrow() and with pytest.raises() for testing exceptions.

Example in TypeScript:

expect(() => authenticateUser("invalid_user", "wrong_pass")).toThrow("Invalid credentials provided");

Example in Python:

with pytest.raises(AuthenticationError, match="Invalid credentials provided"):
    authenticate_user("invalid_user", "wrong_pass")

Test Organization

• •File naming - Use feature_name.test.ts or test_<feature_name>.py converting sentences to snake_case.
• •Group related scenarios - Keep scenarios from same feature file together.

Anti-Patterns to Avoid

• •Tests that pass immediately - Defeats the Red phase purpose. Tests must always FAIL on first run.
• •Skipping verification - Always run tests to confirm they fail for the correct reason.
• •Testing implementation details - Focus on behavior, not internal implementation details.
• •Vague assertions - Use specific checks with expected values.
• •Multiple behaviors per test - One test should verify one scenario.
• •Hidden dependencies - All test dependencies should be explicit.
• •Brittle tests - Avoid tests that break when refactoring.

Execution Guidelines

1. •Read Gherkin Feature Files - Extract and retrieve full context from feature/*.feature files.
2. •Analyze requirements - Break down the feature into testable behaviors.
3. •Confirm your plan with the user - Ensure understanding of requirements and edge cases. Never start making changes without user confirmation.
4. •Write the simplest failing test - Start with the most basic scenario from the feature. Write it under the test directory. Never write multiple tests at once.
5. •Verify the test fails - Run the tests to confirm it fails for the expected reason. Missing test data or syntax errors are never an acceptable reason for failure, only assertion errors are acceptable.
6. •Link test to feature file - Reference feature file name in test docstrings.

Red Phase Checklist

• • Gherkin feature file context retrieved and analyzed.
• • Test clearly describes expected behaviour from feature requirements.
• • Test fails for the right reason (assertion errors only, no compile errors).
• • Test name references feature file name and describes behaviour.
• • Test follows Given, When, Then pattern.
• • Test is isolated and independent (no shared state).
• • Only external dependencies are substituted with test doubles.
• • Test data setup is clear and maintainable.
• • Error messages are descriptive and reference requirements.
• • Test organization follows naming conventions.
• • Only modifications in test files. No production code written.