tdd-guide

Test-driven development for Backend.AI starts with defining clear test scenarios, writing failing tests first, implementing minimal code to pass them, and refactoring safely. This workflow targets feature development and bug fixes using pytest, pants, and with_tables patterns.

Structuring pytest fixtures for repository-first design involves using explicit test structures, strong typing, and small focused functions. You define test targets clearly, leverage with_tables patterns, and use scaffolding templates to ensure tests remain maintainable and isolated.

Yes, you can use pants with pytest to run TDD workflows incrementally. The workflow encourages observing a red state by running tests first, then implementing minimal code to achieve green, and continuously re-running tests during refactoring to ensure stability.

You need the with_tables pattern when defining test scenarios requiring specific database states or repository interactions. It provides scaffolding templates and guidelines for pytest, ensuring repository-first design tests have the necessary structured data context.

Refactoring breaks green tests when implementations grow too large or lose focus. TDD enforces small, focused functions and strong typing to prevent this. Re-run tests continuously during refactoring to catch regressions immediately and maintain the green state.