Turning test writing into a consistently brief and pleasant experience
summary
Argues for integration tests not too many as the highest-value approach, cites academic ranges of 17–50% time spent on testing and interviews with ~15 Django devs, and shows concrete tooling: Factory Boy, freezegun, time-machine, httpretty/responses/VCR, Hypothesis, pytest plugins, Selenium, GitHub Copilot/ChatGPT. Reports a real codebase of ~1,000 tests dropping local runs from 2–3 minutes to 15s by running tests in parallel. Demonstrates Colo generating editable tests from recorded requests; alpha generation can produce passing tests.
AI-written summary · about
chapters (10)
- 0:00Introduction
- 2:08Testing context
- 4:28Developer attitudes
- 9:32Why we test
- 11:51Integration tests
- 14:26Practical testing advice
- 17:44Speeding up tests
- 20:43Test scaffolding
- 26:44AI in testing
- 28:29Demo and conclusion
related talks
28mEgad! How Do We Start Writing Better Tests?
29mEgad! How Do We Start Writing (Better) Tests?
32mEasy, Breezy, Beautiful... Django Unit Tests with Colleen Dunlap
49mKeynote - Testing in Django
37mUnderstanding the successes and the pain points of different testing strategies
48mTesting Should Be Fun
01h 05mMicrotesting
32mMore testing with fewer tests