Fast, Robust, Accurate Or how to love your tests again

An empiric take to testing.

Transcript

1. @xuapsdev
   Fast, Robust, Accurate (WIP)
   Or how to love your tests again
   

   View Slide

2. This is
   not a
   talk
   about
   opinions
   

   View Slide

3. 1 - We progress
   faster with test
   

   View Slide

4. 2 - Bad tests
   could be worst
   than no tests
   

   View Slide

5. View Slide

6. So, what is a
   bad test from
   an empirical
   point of view?
   

   View Slide

7. View Slide

8. Slow tests
   

   View Slide

9. Slow tests
   False alarms
   

   View Slide

10. Slow tests
    False alarms
    Di
    ffi
    cult to read tests
    

    View Slide

11. Brittle tests
    Slow tests
    False alarms
    Di
    ffi
    cult to read tests
    

    View Slide

12. Brittle tests
    Slow tests
    False alarms
    Di
    ffi
    cult to read tests
    What else?
    

    View Slide

13. 3 properties
    Speed
    Detection of
    regressions
    Resistance
    to refactor
    

    View Slide

14. Speed
    

    View Slide

15. Resistance to refactor
    

    View Slide

16. Detection of regression
    

    View Slide

17. Resistance to refactor
    Detection of
    regressions
    Speed
    End to end test Trivial tests
    Brittle tests
    

    View Slide

18. Oh boy my tests are really bad…
    

    View Slide

19. I am so happy that my tests are
    so good…
    

    View Slide

20. View Slide

21. Lets get some data
    

    View Slide

22. Speed
    

    View Slide

23. DEMO
    

    View Slide

24. Resistance to refactor
    

    View Slide

25. Resistance to refactor
    

    View Slide

26. DEMO
    

    View Slide

27. Detection of regression
    

    View Slide

28. Detection of regression
    

    View Slide

29. DEMO
    

    View Slide

30. I need to start using this right
    now
    

    View Slide

31. How am I using
    this data?
    

    View Slide

32. To decide how to test something
    

    View Slide

33. Tune it!
    

    View Slide

34. Tune it!
    Resistance to refactor
    Speed Detection of regressions
    

    View Slide

35. Tune it!
    Resistance to refactor
    Speed Detection of regressions
    Resistance to refactor
    Detection of
    regressions
    Speed
    End to end test Trivial tests
    Brittle tests
    

    View Slide

36. To decide if write a
    test is worthy or
    not
    

    View Slide

37. No test is better than a bad test
    but, then we need something else
    

    View Slide

38. O
    b
    s
    e
    r
    v
    a
    b
    i
    l
    i
    t
    y
    

    View Slide

39. @xuapsdev
    @xuaps@fosstodon.org
    

    View Slide

40. Do you want an extra ball?
    

    View Slide

41. Maintainability
    Extra ball
    

    View Slide

42. View Slide

43. @xuapsdev
    @xuaps@fosstodon.org
    

    View Slide