brokle-testing

Unit tests for Go services use table-driven patterns with mocked repository dependencies. Mock the data layer to isolate business logic, verify calculations and orchestrations without database calls, and assert results against expected outcomes. This approach tests what matters while keeping tests fast and focused.

Test complex business logic, multi-step orchestrations, and domain calculations. Skip trivial CRUD operations, simple framework behavior, and pass-through methods. Pragmatic testing prioritizes high-value assertions that catch real bugs, reducing test maintenance overhead while meeting coverage targets.

Integration tests run against actual database dependencies to verify end-to-end behavior. Set up a test database, seed required data, execute the full workflow, and assert results. Use this for critical paths after unit tests confirm individual layers work in isolation.

Table-driven tests are the standard pattern for Go service layer testing. Define test cases as rows with inputs, mocked outputs, and expected results. This structure scales test coverage efficiently, makes failures clear, and follows Go testing conventions across the Brokle platform.

Mocking repositories isolates business logic from data layer concerns, making unit tests faster, more reliable, and easier to debug. Mocks let you simulate error conditions and edge cases without database setup. Integration tests verify real database interactions separately, creating a balanced test pyramid.

Target coverage by testing business logic, domain rules, and orchestration paths rather than line count. Use service layer tests with mocked dependencies to reach metrics efficiently. Focus assertions on value-producing decisions and error handling paths that impact system behavior.