CSC321.01 2016S, Class 11: Behavior-Driven Development ====================================================== _Overview_ * Preliminaries. * Admin. * Upcoming Work. * Extra Credit. * Questions. * Modeling Lessons. * BDD Basics. * Cucumber Basics. * A Cucumber Example. * Cucumber in Practice. Preliminaries ------------- ### Admin ### Upcoming Work * SaaSBook 8 for Friday. * Reading Journal for Friday (due Thursday at 8pm) * Subject: Reading Journal for Class 12 (Your Name) * How does Fox and Patterson's presentation of test-driven development compare to your prior experiences with unit testing? * Choose the most confusing idea from 8.2 - 8.6. Make a serious attempt to explain this idea in your own words. What questions remain? * Choose the most confusing idea from 8.7 - 8.9. Make a serious attempt to explain this idea in your own words. What questions remain? * What do you see as the relationship(s) between BDD and TDD? Which do you prefer? * What do you see as the most important or enduring concept in this chapter overall? * Citation: Questions are based on journal questions from JLND. ### Good Things to Do #### Academic * Misc. department stuff * Thursday convo: Hilary Mason '00 * Thursday extra: Hilary Mason '00 #### Peer ### Questions _Why does Dr. Racket make some words green and some words black?_ > Words in the dictionary are green. Other words are black. Lessons from Modeling Exercise ------------------------------ What do you take as the key lessons from the modeling exercise? * As you add more attributes, you often find that you want to add more tables. * Try to be DRY. Put the value in the database once, use ids, and let the DBMS handle the relationships. * Movie: ID, Title, Ave Rating * Reviewer: ID, Name * Review: ReviewerID, MovieID, Rating, Text * What do we do if we know that the movies might be searched regularly by MovieID * Sort them by MovieID * What if we're also searching them by ReviewerID? * Overall solution: Build another table. * Fortunately, DBMS's do that for us. * DBMS = Database Management Systems * Because we try not to duplicate data, we can create organizations that seem inefficient. * We need to consider the relationships between tables, A and B * 1:1 for each element in A there is exactly one corresponding element in B * 1:many for each element in A there may be 0 or more corresponding elements in B * many:many for each element in A there may be 0 or more corresponding elements in B, and vice versa. * How do we handle each of these if we are doing things in SQL (rather than in Rails, which tries to hide some of the obvious things) * 1:1, same table * 1:many, two tables, in table B, you have a pointer (id) to the corresponding value in table 1. * many:many, make a separate table that pairs ids from A with ids from B. * Don't forget that fields usually need types! Cucumber Basics --------------- * A toolkit to facilitate behavior driven design The stack Features Scenarios Steps Step Definitions - regular expressions that turn "natural" language into calls to functions /I go to the page for user number ([0-9]+)/ gotopage "/users/\1" Support Code Where we write things like gotopage Automation Library Things your support code relies on Capybara When you write Cucumber, you need to write the "human stuff" (features, scenarios, steps), all reside in the feature file; and the programmer stuff (step definitions, support code) Makes sense for code reuse Makes it harder to do updates And working in what feels like three different languages is not always all that fun A Cucumber Example ------------------ Let's walk through some scenarios and think about what work we might have to do to implement these Cucumber scenarios. Do they affect model, view, or controller (or all three)? (Suppose we had the repaired version of the original database, with no director.) 
Feature: search for movies by director

  As a movie buff
  So that I can find movies with my favorite director
  I want to include and search on director information in movies I enter

Background: movies in database

  Given the following movies exist:
  | title        | rating | director     | release_date |
  | Star Wars    | PG     | George Lucas |   1977-05-25 |
  | Blade Runner | PG     | Ridley Scott |   1982-06-25 |
  | Alien        | R      |              |   1979-05-25 |
  | THX-1138     | R      | George Lucas |   1971-03-11 |

Scenario: add director to existing movie
  When I go to the edit page for "Alien"
  And  I fill in "Director" with "Ridley Scott"
  And  I press "Update Movie Info"
  Then the director of "Alien" should be "Ridley Scott"

Scenario: find movie with same director
  Given I am on the details page for "Star Wars"
  When  I follow "Find Movies With Same Director"
  Then  I should be on the Similar Movies page for "Star Wars"
  And   I should see "THX-1138"
  But   I should not see "Blade Runner"

Scenario: can't find similar movies if we don't know director (sad path)
  Given I am on the details page for "Alien"
  Then  I should not see "Ridley Scott"
  When  I follow "Find Movies With Same Director"
  Then  I should be on the home page
  And   I should see "'Alien' has no director info"
vs. movies reset movies add new Movie("Star Wars", "PG", "George Lucas", ...); IY says that the poorly named FactoryGirl Gem helps do this kind of code. Cucumber in Practice --------------------