DEV Community

Cover image for AI generated test cases to improve application stability
Karlis
Karlis

Posted on

AI generated test cases to improve application stability

Summer Bug Smash: Smash Stories 🐛🛹

This is a submission for DEV's Summer Bug Smash: Smash Stories powered by Sentry.

AI suggestion

Now my side projects I develop using AI. Usual prompt for me is to ask how to improve the project. Usually one of the suggestions is to add more tests. I am a developer and not a tester, therefore to do the good necessary test coverage part I leave to AI. As for me, application testing is not trivial and I accept that each one can develop his/her own testing strategy and it is totally fine.

Acacia-Log example and gained experience

For example, Acacia-log project. First time AI generated a lot of tests Added testing and improved performance. The generated tests did not find any bugs (AI has hard job at finding bugs in it's own code), but further development improved by far. After each feature addition, AI is able to do improved testing and early catch changes that broke previous logic.

Screenshot of AI adding new features (last step is doing verification):
feature verification

Example of the prompt for Phase 5 generation:
code generation in phases

Top comments (0)