Tasj course lesson 5-1 - End to End tests vs One Feature Tests - notes_

tasj-course-lesson-5-1-e2e-vs-one-feature-tests.txt

***

Did the prototype/proof of concept:
  Smoke Coverage:
  - main: End to End scenario
  - additional: Feature Tests ( = 1 feature per test method, or "unit style" tests)
  - 6s total time execution

What next?
  Full Coverage:
  - E2E?
  - FT? 

Let's think...
  E2E (dependent long tests)
  + pros
    + fast in "prototype" execution
    + fast in implementation 
    ++ do integrated testing 
    
  - cons
    - slow in coverage analysis and support
    - slow in support:
      - change coverage
        - because less readable
      - bugs & test running 
        - because "fail fast"
        
  FT (independent small tests)
  + pros
    + fast in coverage analysis
    + fast in support:
      + bugs & test running
      + coverage change
        + readable
    + may be faster in follwing implementation of new feateres
    
  - cons
    - slower in execution  (? how much)      => + THE SAME if used backend helpers for preconditions
    - slower in "first-prototype" implementation
    -- no integrated testing 

Hm... =>

Decided:
  Full Coverage:
  - more FT!
  - and let's refactor old long e2e

Refactored:
  !!! So slow! 16s vs 6s (previous impl.)
  
Here comes:
  Backend helpers to presetup data before each test (which we can use only if use language the same as backend devs)
    Used:
      Yeah!!! 8s (with even better coverage) vs 6s 
      That's good:)