Egad! How Do We Start Writing (Better) Tests?
summary
A practical path shows a team building a Python test stack using pytest, requests, and selenium webdriver, then scaling with pytest-xdist and cloud browsers to cut an hour-long suite down to a fraction of the time. The talk cites the rule of ones (unit ≈1ms, service ≈1s, end-to-end ≈1min) and a $655M NASA integration failure to argue for feature tests. Common flakiness causes are race conditions, RPCs, stale tests and test bugs; mitigations include timeouts, retries, contract testing (pact), logging and CI-focused test hygiene.
AI-written summary · about
chapters (18)
- 0:00Intro
- 0:34Who is Andy Knight?
- 2:11Why Talk About Test Automation?
- 4:53Meet Amanda the Panda!
- 5:04She Works at Bamboozle
- 5:55Buzzword Bingo
- 8:02The Automated Testing Pyramid
- 10:41A Spectacular Failure
- 13:02Packages
- 13:57Amanda's Test Code
- 15:01Hypothesis and pytest-bdd
- 16:20Test Data Management
- 18:51What Should Be Automated?
- 21:54Continuous Integration and Delivery
- 22:164 Common Failure Reasons
- 24:03Reliability
- 27:31Resources for Python Test Automation
- 28:31Concluding Remarks
related talks
28mEgad! How Do We Start Writing Better Tests?
35mTurning test writing into a consistently brief and pleasant experience
38mTesting for the Terrified: How to Write Tests, Conquer Guilt, and Level Up
24mDon't write a lot of tests!
21mDon't just test, my friend, test better
48mA Reflection on Software Testing
17mTest! (Some OF) Your! Code!: How To Write...
56mIf You Won't Test Your Code, Who Will?