Automated testing - how?
Download as PPTX, PDF0 likes184 views
The document outlines various automated testing methodologies, emphasizing the importance of unit, component, integration, end-to-end, and smoke tests. It highlights best practices for test descriptions, code reviews, and continuous integration and deployment processes. Additionally, it discusses tools like Protractor, Mocha, and Jenkins that aid in testing and CI/CD workflows.
Automated testing - how?
- 2. unit tests component tests integration tests end-to-end and smoke tests acceptance and contract tests code review continuous integration and deployment
- 5. 4 clear blocks: setup, operation, validation, teardown Single claim per test, description should not contain “and” Independently executable Unit tests
- 6. Test descriptions first, then code and tests in parallel Code first, then tests Pure TDD (single failing test – single feature – refactor – repeat) Unit tests
- 11. Cover interface Cover possible behaviours of dependencies Common mock for positive scenarios Ad-hock mocks for error scenarios Possible duplication of some unit tests Mock databases with in-memory databases. Component tests
- 14. Cover interface between units E.g. if plugins can be plugged in Often can be omitted because of component tests cover it Sometimes it may be efficient to create integration tests instead of unit tests Integration tests
- 17. ✓ unit tests ✓ component tests ✓ integration tests end-to-end and smoke tests acceptance and contract tests code review continuous integration and deployment
- 18. Use staging environment Stable test data set in database (alternative would be setup and teardown as part of test) Mock external http-based services using Robohydra Create page models for abstracting interaction with html/css Protractor for running tests in browser End-to-end tests for HTTP API can be done using mocha, much faster End-to-end tests
- 20. Technically same as end-to-end tests Use in production environment right after deployment Very limited number of tests, just to validate the deployment success Smoke tests
- 21. Link and track with business requirements Does not have to be end-to-end test Acceptance tests
- 22. Set of customer-specific requirements If the customer vanishes, the tests can be removed Coverage tools can be used to find code that no customer requires Contract tests
- 23. ✓ unit tests ✓ component tests ✓ integration tests ✓ end-to-end and smoke tests ✓ acceptance and contract tests code review continuous integration and deployment
- 24. No seniority – everybody can and has to review Are the tests descriptions working as a specification? Has the author revised the existing tests for the entire unit or just added new? Readability of both code and tests. Third eye-pair code review for pair programming. Code review
- 25. Continuous integration and deployment Development and tests (feature branch) Code review Tests (feature branch) Merge to master Tests (master) End-to-end tests Deploy to production Deploy to stage Smoke tests Development CI (docker) Stage Production
- 26. Protractor – browser end-to-end test runner Karma – JavaScript unit test runner Mocha + chai – JavaScript unit test framework Istanbul – coverage Nock – http interception for nodejs Sinon – mocking tool for JavaScript objects and libraries Robohydra – webserver for mocking Phabricator + arcanist – code review Jenkins – continuous integration Tools