Testing

Reliable software depends on a robust testing strategy. We follow the "Testing Pyramid" approach to ensure high quality and fast feedback loops.

The Testing Pyramid¶

Scope: Test individual functions, classes, or components in isolation.
Speed: Extremely fast (milliseconds).
Volume: Largest number of tests.
Responsibility: Developers must write unit tests for all new code. Target

80% code coverage.
Tools: Jest (JS/TS), Pytest (Python), JUnit (Java), Go Test (Golang).

Scope: Test interactions between different modules or services (e.g., API to Database, Service A to Service B).
Speed: Slower than unit tests but faster than E2E.
Volume: Moderate number of tests.
Responsibility: Validate that components work together correctly. Use mocks/stubs for external dependencies where appropriate, but prefer real dependencies (e.g., test containers) for critical paths.

Scope: Test the entire application flow from the user's perspective.
Speed: Slowest.
Volume: Smallest number of tests. focus on critical user journeys (e.g., "Sign Up", "Checkout").
Responsibility: Ensure the system works as a whole.
Tools: Playwright, Cypress, Selenium.

Test-Driven Development (TDD): Encouraged. Write the test before the code to clarify requirements and design.
CI Integration: All tests must pass in CI before merging. A broken build is a top priority to fix.
Flaky Tests: Flaky tests (tests that pass/fail randomly) destroy trust. Quarantine or fix flaky tests immediately.
Mocking: Use mocking for external systems (3rd party APIs) to ensure tests are deterministic. Avoid excessive mocking of internal code, as it can hide integration bugs.
Test Data: Use factories or fixtures to generate test data. Avoid relying on shared static data that can be mutated by other tests.

Refer to the Style Guides for specific testing conventions in each language.