Prelude

Role models are important.
-- Officer Alex J. Murphy / RoboCop

The goal of this guide is to present a set of best practices and style prescriptions for Ruby on Rails 3 & 4 development. It's a complementary guide to the already existing community-driven Ruby coding style guide.

While in the guide the section Testing Rails applications is after Developing Rails applications I truly believe that Behaviour-Driven Development (BDD) is the best way to develop software. Keep that in mind.

Rails is an opinionated framework and this is an opinionated guide. In my mind I'm totally certain that RSpec is superior to Test::Unit, Sass is superior to CSS and Haml (Slim) is superior to Erb. So don't expect to find any Test::Unit, CSS or Erb advice in here.

Some of the advice here is applicable only to Rails 3.1+.

You can generate a PDF or an HTML copy of this guide using Transmuter.

Translations of the guide are available in the following languages:

Developing Rails applications

Configuration

Put custom initialization code in config/initializers. The code in initializers executes on application startup.
Keep initialization code for each gem in a separate file with the same name as the gem, for example carrierwave.rb, active_admin.rb, etc.

Adjust accordingly the settings for development, test and production environment (in the corresponding files under config/environments/)

Mark additional assets for precompilation (if any):

```Ruby
# config/environments/production.rb
# Precompile additional assets (application.js, application.css, and all non-JS/CSS are already added)
config.assets.precompile += %w( rails_admin/rails_admin.css rails_admin/rails_admin.js )
```

Keep configuration that's applicable to all environments in the config/application.rb file.
Create an additional staging environment that closely resembles the production one.

Routing

When you need to add more actions to a RESTful resource (do you really need them at all?) use member and collection routes.

# bad
get 'subscriptions/:id/unsubscribe'
resources :subscriptions

# good
resources :subscriptions do
  get 'unsubscribe', on: :member
end

# bad
get 'photos/search'
resources :photos

# good
resources :photos do
  get 'search', on: :collection
end

If you need to define multiple member/collection routes use the alternative block syntax.

resources :subscriptions do
  member do
    get 'unsubscribe'
    # more routes
  end
end

resources :photos do
  collection do
    get 'search'
    # more routes
  end
end

Use nested routes to express better the relationship between ActiveRecord models.

class Post < ActiveRecord::Base
  has_many :comments
end

class Comments < ActiveRecord::Base
  belongs_to :post
end

# routes.rb
resources :posts do
  resources :comments
end

Use namespaced routes to group related actions.

namespace :admin do
  # Directs /admin/products/* to Admin::ProductsController
  # (app/controllers/admin/products_controller.rb)
  resources :products
end

Never use the legacy wild controller route. This route will make all actions in every controller accessible via GET requests.
```
# very bad
match ':controller(/:action(/:id(.:format)))'
```
Don't use match to define any routes. It's removed from Rails 4.

Controllers

Keep the controllers skinny - they should only retrieve data for the view layer and shouldn't contain any business logic (all the business logic should naturally reside in the model).
Each controller action should (ideally) invoke only one method other than an initial find or new.
Share no more than two instance variables between a controller and a view.

Models

Introduce non-ActiveRecord model classes freely.
Name the models with meaningful (but short) names without abbreviations.

If you need model objects that support ActiveRecord behavior(like validation) use the ActiveAttr gem.

class Message
  include ActiveAttr::Model

  attribute :name
  attribute :email
  attribute :content
  attribute :priority

  attr_accessible :name, :email, :content

  validates :name, presence: true
  validates :email, format: { with: /\A[-a-z0-9_+\.]+\@([-a-z0-9]+\.)+[a-z0-9]{2,4}\z/i }
  validates :content, length: { maximum: 500 }
end

For a more complete example refer to the RailsCast on the subject.

ActiveRecord

Avoid altering ActiveRecord defaults (table names, primary key, etc) unless you have a very good reason (like a database that's not under your control).
```
# bad - don't do this if you can modify the schema
class Transaction < ActiveRecord::Base
  self.table_name = 'order'
  ...
end
```

Group macro-style methods (has_many, validates, etc) in the beginning of the class definition.

class User < ActiveRecord::Base
  # keep the default scope first (if any)
  default_scope { where(active: true) }

  # constants come up next
  GENDERS = %w(male female)

  # afterwards we put attr related macros
  attr_accessor :formatted_date_of_birth

  attr_accessible :login, :first_name, :last_name, :email, :password

  # followed by association macros
  belongs_to :country

  has_many :authentications, dependent: :destroy

  # and validation macros
  validates :email, presence: true
  validates :username, presence: true
  validates :username, uniqueness: { case_sensitive: false }
  validates :username, format: { with: /\A[A-Za-z][A-Za-z0-9._-]{2,19}\z/ }
  validates :password, format: { with: /\A\S{8,128}\z/, allow_nil: true}

  # next we have callbacks
  before_save :cook
  before_save :update_username_lower

  # other macros (like devise's) should be placed after the callbacks

  ...
end

Prefer has_many :through to has_and_belongs_to_many. Using has_many :through allows additional attributes and validations on the join model.

# using has_and_belongs_to_many
class User < ActiveRecord::Base
  has_and_belongs_to_many :groups
end

class Group < ActiveRecord::Base
  has_and_belongs_to_many :users
end

# prefered way - using has_many :through
class User < ActiveRecord::Base
  has_many :memberships
  has_many :groups, through: :memberships
end

class Membership < ActiveRecord::Base
  belongs_to :user
  belongs_to :group
end

class Group < ActiveRecord::Base
  has_many :memberships
  has_many :users, through: :memberships
end

Prefer self[:attribute] over read_attribute(:attribute).

# bad
def amount
  read_attribute(:amount) * 100
end

# good
def amount
  self[:amount] * 100
end

Always use the new "sexy" validations.

# bad
validates_presence_of :email

# good
validates :email, presence: true

When a custom validation is used more than once or the validation is some regular expression mapping, create a custom validator file.

# bad
class Person
  validates :email, format: { with: /\A([^@\s]+)@((?:[-a-z0-9]+\.)+[a-z]{2,})\z/i }
end

# good
class EmailValidator < ActiveModel::EachValidator
  def validate_each(record, attribute, value)
    record.errors[attribute] << (options[:message] || 'is not a valid email') unless value =~ /\A([^@\s]+)@((?:[-a-z0-9]+\.)+[a-z]{2,})\z/i
  end
end

class Person
  validates :email, email: true
end

Keep custom validators under app/validators.
Consider extracting custom validators to a shared gem if you're maintaining several related apps or the validators are generic enough.

Use named scopes freely.

class User < ActiveRecord::Base
  scope :active, -> { where(active: true) }
  scope :inactive, -> { where(active: false) }

  scope :with_orders, -> { joins(:orders).select('distinct(users.id)') }
end

Wrap named scopes in lambdas to initialize them lazily (this is only a prescription in Rails 3, but is mandatory in Rails 4).

# bad
class User < ActiveRecord::Base
  scope :active, where(active: true)
  scope :inactive, where(active: false)

  scope :with_orders, joins(:orders).select('distinct(users.id)')
end

# good
class User < ActiveRecord::Base
  scope :active, -> { where(active: true) }
  scope :inactive, -> { where(active: false) }

  scope :with_orders, -> { joins(:orders).select('distinct(users.id)') }
end

When a named scope defined with a lambda and parameters becomes too complicated, it is preferable to make a class method instead which serves the same purpose of the named scope and returns an ActiveRecord::Relation object. Arguably you can define even simpler scopes like this.
```
class User < ActiveRecord::Base
  def self.with_orders
    joins(:orders).select('distinct(users.id)')
  end
end
```
Beware of the behavior of the update_attribute method. It doesn't run the model validations (unlike update_attributes) and could easily corrupt the model state. The method was finally deprecated in Rails 3.2.7 and does not exist in Rails 4.
Use user-friendly URLs. Show some descriptive attribute of the model in the URL rather than its id. There is more than one way to achieve this:
- Override the to_param method of the model. This method is used by Rails for constructing a URL to the object. The default implementation returns the id of the record as a String. It could be overridden to include another human-readable attribute.
```
```Ruby
  class Person
    def to_param
      "#{id} #{name}".parameterize
    end
  end
```
```
  In order to convert this to a URL-friendly value, parameterize should be called on the string. The id of the object needs to be at the beginning so that it can be found by the find method of ActiveRecord.
- Use the friendly_id gem. It allows creation of human-readable URLs by using some descriptive attribute of the model instead of its id.
```
```Ruby
  class Person
    extend FriendlyId
    friendly_id :name, use: :slugged
  end
```

Check the [gem documentation](https://github.com/norman/friendly_id) for more information about its usage.
```

Migrations

Keep the schema.rb (or structure.sql) under version control.
Use rake db:schema:load instead of rake db:migrate to initialize an empty database.
Use rake db:test:prepare to update the schema of the test database.
Enforce default values in the migrations themselves instead of in the application layer.
```
# bad - application enforced default value
def amount
  self[:amount] or 0
end
```
While enforcing table defaults only in Rails is suggested by many Rails developers it's an extremely brittle approach that leaves your data vulnerable to many application bugs. And you'll have to consider the fact that most non-trivial apps share a database with other applications, so imposing data integrity from the Rails app is impossible.
Enforce foreign-key constraints. While ActiveRecord does not support them natively, there some great third-party gems like schema_plus and foreigner.

When writing constructive migrations (adding tables or columns), use the new Rails 3.1 way of doing the migrations - use the change method instead of up and down methods.

# the old way
class AddNameToPerson < ActiveRecord::Migration
  def up
    add_column :persons, :name, :string
  end

  def down
    remove_column :person, :name
  end
end

# the new prefered way
class AddNameToPerson < ActiveRecord::Migration
  def change
    add_column :persons, :name, :string
  end
end

Don't use model classes in migrations. The model classes are constantly evolving and at some point in the future migrations that used to work might stop, because of changes in the models used.

Views

Never call the model layer directly from a view.
Never make complex formatting in the views, export the formatting to a method in the view helper or the model.
Mitigate code duplication by using partial templates and layouts.

Internationalization

No strings or other locale specific settings should be used in the views, models and controllers. These texts should be moved to the locale files in the config/locales directory.
When the labels of an ActiveRecord model need to be translated, use the activerecord scope:
```
en:
  activerecord:
    models:
      user: Member
    attributes:
      user:
        name: "Full name"
```
Then User.model_name.human will return "Member" and User.human_attribute_name("name") will return "Full name". These translations of the attributes will be used as labels in the views.
Separate the texts used in the views from translations of ActiveRecord attributes. Place the locale files for the models in a folder models and the texts used in the views in folder views.
- When organization of the locale files is done with additional directories, these directories must be described in the application.rb file in order to be loaded.
```
```Ruby
  # config/application.rb
  config.i18n.load_path += Dir[Rails.root.join('config', 'locales', '**', '*.{rb,yml}')]
```
```
Place the shared localization options, such as date or currency formats, in files under the root of the locales directory.
Use the short form of the I18n methods: I18n.t instead of I18n.translate and I18n.l instead of I18n.localize.
Use "lazy" lookup for the texts used in views. Let's say we have the following structure:
```
en:
  users:
    show:
      title: "User details page"
```
The value for users.show.title can be looked up in the template app/views/users/show.html.haml like this:
```
= t '.title'
```

Use the dot-separated keys in the controllers and models instead of specifying the :scope option. The dot-separated call is easier to read and trace the hierarchy.

# use this call
I18n.t 'activerecord.errors.messages.record_invalid'

# instead of this
I18n.t :record_invalid, :scope => [:activerecord, :errors, :messages]

More detailed information about the Rails i18n can be found in the [Rails Guides] (http://guides.rubyonrails.org/i18n.html)

Assets

Use the assets pipeline to leverage organization within your application.

Reserve app/assets for custom stylesheets, javascripts, or images.
Use lib/assets for your own libraries, that doesn’t really fit into the scope of the application.
Third party code such as jQuery or bootstrap should be placed in vendor/assets.
When possible, use gemified versions of assets (e.g. jquery-rails, jquery-ui-rails, bootstrap-sass, zurb-foundation).

Mailers

Name the mailers SomethingMailer. Without the Mailer suffix it isn't immediately apparent what's a mailer and which views are related to the mailer.
Provide both HTML and plain-text view templates.
Enable errors raised on failed mail delivery in your development environment. The errors are disabled by default.
```
# config/environments/development.rb

config.action_mailer.raise_delivery_errors = true
```

Use smtp.gmail.com for SMTP server in the development environment (unless you have local SMTP server, of course).

# config/environments/development.rb

config.action_mailer.smtp_settings = {
  address: 'smtp.gmail.com',
  # more settings
}

Provide default settings for the host name.

# config/environments/development.rb
config.action_mailer.default_url_options = {host: "#{local_ip}:3000"}


# config/environments/production.rb
config.action_mailer.default_url_options = {host: 'your_site.com'}

# in your mailer class
default_url_options[:host] = 'your_site.com'

If you need to use a link to your site in an email, always use the _url, not _path methods. The _url methods include the host name and the _path methods don't.

# wrong
You can always find more info about this course
= link_to 'here', url_for(course_path(@course))

# right
You can always find more info about this course
= link_to 'here', url_for(course_url(@course))

Format the from and to addresses properly. Use the following format:

# in your mailer class
default from: 'Your Name <info@your_site.com>'

Make sure that the e-mail delivery method for your test environment is set to test:
```
# config/environments/test.rb

config.action_mailer.delivery_method = :test
```

The delivery method for development and production should be smtp:

# config/environments/development.rb, config/environments/production.rb

config.action_mailer.delivery_method = :smtp

When sending html emails all styles should be inline, as some mail clients have problems with external styles. This however makes them harder to maintain and leads to code duplication. There are two similar gems that transform the styles and put them in the corresponding html tags: premailer-rails3 and roadie.
Sending emails while generating page response should be avoided. It causes delays in loading of the page and request can timeout if multiple email are send. To overcome this emails can be send in background process with the help of sidekiq gem.

Bundler

Put gems used only for development or testing in the appropriate group in the Gemfile.
Use only established gems in your projects. If you're contemplating on including some little-known gem you should do a careful review of its source code first.
OS-specific gems will by default result in a constantly changing Gemfile.lock for projects with multiple developers using different operating systems. Add all OS X specific gems to a darwin group in the Gemfile, and all Linux specific gems to a linux group:
```
# Gemfile
group :darwin do
  gem 'rb-fsevent'
  gem 'growl'
end

group :linux do
  gem 'rb-inotify'
end
```
To require the appropriate gems in the right environment, add the following to config/application.rb:
```
platform = RUBY_PLATFORM.match(/(linux|darwin)/)[0].to_sym
Bundler.require(platform)
```
Do not remove the Gemfile.lock from version control. This is not some randomly generated file - it makes sure that all of your team members get the same gem versions when they do a bundle install.

Priceless Gems

One of the most important programming principles is "Don't reinvent the wheel!". If you're faced with a certain task you should always look around a bit for existing solutions, before rolling your own. Here's a list of some "priceless" gems (all of them Rails 3.1 compliant) that are useful in many Rails projects:

active_admin - With ActiveAdmin the creation of admin interface for your Rails app is child's play. You get a nice dashboard, CRUD UI and lots more. Very flexible and customizable.
better_errors - Better Errors replaces the standard Rails error page with a much better and more useful error page. It is also usable outside of Rails in any Rack app as Rack middleware.
bullet - The Bullet gem is designed to help you increase your application’s performance by reducing the number of queries it makes. It will watch your queries while you develop your application and notify you when you should add eager loading (N+1 queries), when you’re using eager loading that isn’t necessary and when you should use counter cache.
cancan - CanCan is an authorization gem that lets you restrict users access to resources. All permissions are defined in a single file (ability.rb) and convenient methods for checking and ensuring permissions are available throughout the application.
capybara - Capybara aims to simplify the process of integration testing Rack applications, such as Rails, Sinatra or Merb. Capybara simulates how a real user would interact with a web application. It is agnostic about the driver running your tests and currently comes with Rack::Test and Selenium support built in. HtmlUnit, WebKit and env.js are supported through external gems. Works great in combination with RSpec & Cucumber.
carrierwave - the ultimate file upload solution for Rails. Support both local and cloud storage for the uploaded files (and many other cool things). Integrates great with ImageMagick for image post-processing.
compass-rails - Great gem that adds support for some css frameworks. Includes collection of sass mixins that reduces code of css files and help fight with browser incompatibilities.
cucumber-rails - Cucumber is the premium tool to develop feature tests in Ruby. cucumber-rails provides Rails integration for Cucumber.
devise - Devise is full-featured authentication solution for Rails applications. In most cases it's preferable to use devise to unrolling your custom authentication solution.
fabrication - a great fixture replacement (editor's choice).
factory_girl - an alternative to fabrication. Nice and mature fixture replacement. Spiritual ancestor of fabrication.
ffaker - handy gem to generate dummy data (names, addresses, etc).
feedzirra - Very fast and flexible RSS/Atom feed parser.
friendly_id - Allows creation of human-readable URLs by using some descriptive attribute of the model instead of its id.
globalize3 - Globalize3 is the successor of Globalize for Rails and is targeted at ActiveRecord version 3.x. It is compatible with and builds on the new I18n API in Ruby on Rails and adds model translations to ActiveRecord.
guard - fantastic gem that monitors file changes and invokes tasks based on them. Loaded with lots of useful extension. Far superior to autotest and watchr.
haml-rails - haml-rails provides Rails integration for Haml.
haml - HAML is a concise templating language, considered by many (including yours truly) to be far superior to Erb.
kaminari - Great paginating solution.
machinist - Fixtures aren't fun. Machinist is.
rspec-rails - RSpec is a replacement for Test::MiniTest. I cannot recommend highly enough RSpec. rspec-rails provides Rails integration for RSpec.
sidekiq - Sidekiq is probably the easiest and most scalable way to run background jobs in your Rails app.
simple_form - once you've used simple_form (or formtastic) you'll never want to hear about Rails's default forms. It has a great DSL for building forms and no opinion on markup.
simplecov-rcov - RCov formatter for SimpleCov. Useful if you're trying to use SimpleCov with the Hudson contininous integration server.
simplecov - code coverage tool. Unlike RCov it's fully compatible with Ruby 1.9. Generates great reports. Must have!
slim - Slim is a concise templating language, considered by many far superior to HAML (not to mention Erb). The only thing stopping me from using Slim massively is the lack of good support in major editors/IDEs. Its performance is phenomenal.
spork - A DRb server for testing frameworks (RSpec / Cucumber currently) that forks before each run to ensure a clean testing state. Simply put it preloads a lot of test environment and as consequence the startup time of your tests in greatly decreased. Absolute must have!
sunspot - SOLR powered full-text search engine.

This list is not exhaustive and other gems might be added to it along the road. All of the gems on the list are field tested, have active development and community and are known to be of good code quality.

Flawed Gems

This is a list of gems that are either problematic or superseded by other gems. You should avoid using them in your projects.

rmagick - this gem is notorious for its memory consumption. Use minimagick instead.
autotest - old solution for running tests automatically. Far inferior to guard and watchr.
rcov - code coverage tool, not compatible with Ruby 1.9. Use SimpleCov instead.
therubyracer - the use of this gem in production is strongly discouraged as it uses a very large amount of memory. I'd suggest using node.js instead.

This list is also a work in progress. Please, let me know if you know other popular, but flawed gems.

Managing processes

If your projects depends on various external processes use foreman to manage them.

Testing Rails applications

The best approach to implementing new features is probably the BDD approach. You start out by writing some high level feature tests (generally written using Cucumber), then you use these tests to drive out the implementation of the feature. First you write view specs for the feature and use those specs to create the relevant views. Afterwards you create the specs for the controller(s) that will be feeding data to the views and use those specs to implement the controller. Finally you implement the models specs and the models themselves.

Cucumber

Tag your pending scenarios with @wip (work in progress). These scenarios will not be taken into account and will not be marked as failing. When finishing the work on a pending scenario and implementing the functionality it tests, the tag @wip should be removed in order to include this scenario in the test suite.
Setup your default profile to exclude the scenarios tagged with @javascript. They are testing using the browser and disabling them is recommended to increase the regular scenarios execution speed.
Setup a separate profile for the scenarios marked with @javascript tag.
- The profiles can be configured in the cucumber.yml file.
```
```Ruby
  # definition of a profile:
  profile_name: --tags @tag_name
```
```
- A profile is run with the command:
```
```
  cucumber -p profile_name
```
```
If using fabrication for fixtures replacement, use the predefined fabrication steps
Do not use the old web_steps.rb step definitions! The web steps were removed from the latest version of Cucumber. Their usage leads to the creation of verbose scenarios that do not properly reflect the application domain.
When checking for the presence of an element with visible text (link, button, etc.) check for the text, not the element id. This can detect problems with the i18n.

Create separate features for different functionality regarding the same kind of objects:

# bad
Feature: Articles
# ... feature  implementation ...

# good
Feature: Article Editing
# ... feature  implementation ...

Feature: Article Publishing
# ... feature  implementation ...

Feature: Article Search
# ... feature  implementation ...

Each feature has three main components
- Title
- Narrative - a short explanation what the feature is about.
- Acceptance criteria - the set of scenarios each made up of individual steps.

The most common format is known as the Connextra format.

In order to [benefit] ...
A [stakeholder]...
Wants to [feature] ...

This format is the most common but is not required, the narrative can be free text depending on the complexity of the feature.

Use Scenario Outlines freely to keep the scenarios DRY.

Scenario Outline: User cannot register with invalid e-mail
  When I try to register with an email "<email>"
  Then I should see the error message "<error>"

Examples:
  |email         |error                 |
  |              |The e-mail is required|
  |invalid email |is not a valid e-mail |

The steps for the scenarios are in .rb files under the step_definitions directory. The naming convention for the steps file is [description]_steps.rb. The steps can be separated into different files based on different criterias. It is possible to have one steps file for each feature (home_page_steps.rb). There also can be one steps file for all features for a particular object (articles_steps.rb).

Use multiline step arguments to avoid repetition

Scenario: User profile
  Given I am logged in as a user "John Doe" with an e-mail "[email protected]"
  When I go to my profile
  Then I should see the following information:
    |First name|John         |
    |Last name |Doe          |
    |E-mail    |user@test.com|

# the step:
Then /^I should see the following information:$/ do |table|
  table.raw.each do |field, value|
    find_field(field).value.should =~ /#{value}/
  end
end

RSpec

Use just one expectation per example.

# bad
describe ArticlesController do
  #...

  describe 'GET new' do
    it 'assigns new article and renders the new article template' do
      get :new
      assigns[:article].should be_a_new Article
      response.should render_template :new
    end
  end

  # ...
end

# good
describe ArticlesController do
  #...

  describe 'GET new' do
    it 'assigns a new article' do
      get :new
      assigns[:article].should be_a_new Article
    end

    it 'renders the new article template' do
      get :new
      response.should render_template :new
    end
  end

end

Make heavy use of describe and context

Name the describe blocks as follows:

use "description" for non-methods
use pound "#method" for instance methods

use dot ".method" for class methods

class Article
  def summary
    #...
  end

  def self.latest
    #...
  end
end

# the spec...
describe Article do
  describe '#summary' do
    #...
  end

  describe '.latest' do
    #...
  end
end

Use fabricators to create test objects.

Make heavy use of mocks and stubs

# mocking a model
article = mock_model(Article)

# stubbing a method
Article.stub(:find).with(article.id).and_return(article)

When mocking a model, use the as_null_object method. It tells the output to listen only for messages we expect and ignore any other messages.
```
article = mock_model(Article).as_null_object
```

Use let blocks instead of before(:each) blocks to create data for the spec examples. let blocks get lazily evaluated.

# use this:
let(:article) { Fabricate(:article) }

# ... instead of this:
before(:each) { @article = Fabricate(:article) }

Use subject when possible

describe Article do
  subject { Fabricate(:article) }

  it 'is not published on creation' do
    subject.should_not be_published
  end
end

Use specify if possible. It is a synonym of it but is more readable when there is no docstring.

# bad
describe Article do
  before { @article = Fabricate(:article) }

  it 'is not published on creation' do
    @article.should_not be_published
  end
end

# good
describe Article do
  let(:article) { Fabricate(:article) }
  specify { article.should_not be_published }
end

Use its when possible

# bad
describe Article do
  subject { Fabricate(:article) }

  it 'has the current date as creation date' do
    subject.creation_date.should == Date.today
  end
end

# good
describe Article do
  subject { Fabricate(:article) }
  its(:creation_date) { should == Date.today }
end

Use shared_examples if you want to create a spec group that can be shared by many other tests.

# bad
 describe Array do
   subject { Array.new [7, 2, 4] }

   context "initialized with 3 items" do
     its(:size) { should eq(3) }
   end
 end

 describe Set do
   subject { Set.new [7, 2, 4] }

   context "initialized with 3 items" do
     its(:size) { should eq(3) }
   end
 end

#good
 shared_examples "a collection" do
   subject { described_class.new([7, 2, 4]) }

   context "initialized with 3 items" do
     its(:size) { should eq(3) }
   end
 end

 describe Array do
   it_behaves_like "a collection"
 end

 describe Set do
   it_behaves_like "a collection"
 end

Views

The directory structure of the view specs spec/views matches the one in app/views. For example the specs for the views in app/views/users are placed in spec/views/users.
The naming convention for the view specs is adding _spec.rb to the view name, for example the view _form.html.haml has a corresponding spec _form.html.haml_spec.rb.
spec_helper.rb need to be required in each view spec file.
The outer describe block uses the path to the view without the app/views part. This is used by the render method when it is called without arguments.
```
# spec/views/articles/new.html.haml_spec.rb
require 'spec_helper'

describe 'articles/new.html.haml' do
  # ...
end
```
Always mock the models in the view specs. The purpose of the view is only to display information.

The method assign supplies the instance variables which the view uses and are supplied by the controller.

# spec/views/articles/edit.html.haml_spec.rb
describe 'articles/edit.html.haml' do
it 'renders the form for a new article creation' do
  assign(
    :article,
    mock_model(Article).as_new_record.as_null_object
  )
  render
  rendered.should have_selector('form',
    method: 'post',
    action: articles_path
  ) do |form|
    form.should have_selector('input', type: 'submit')
  end
end

Prefer the capybara negative selectors over should_not with the positive.

# bad
page.should_not have_selector('input', type: 'submit')
page.should_not have_xpath('tr')

# good
page.should have_no_selector('input', type: 'submit')
page.should have_no_xpath('tr')

When a view uses helper methods, these methods need to be stubbed. Stubbing the helper methods is done on the template object:

# app/helpers/articles_helper.rb
class ArticlesHelper
  def formatted_date(date)
    # ...
  end
end

# app/views/articles/show.html.haml
= "Published at: #{formatted_date(@article.published_at)}"

# spec/views/articles/show.html.haml_spec.rb
describe 'articles/show.html.haml' do
  it 'displays the formatted date of article publishing' do
    article = mock_model(Article, published_at: Date.new(2012, 01, 01))
    assign(:article, article)

    template.stub(:formatted_date).with(article.published_at).and_return('01.01.2012')

    render
    rendered.should have_content('Published at: 01.01.2012')
  end
end

The helpers specs are separated from the view specs in the spec/helpers directory.

Controllers

Mock the models and stub their methods. Testing the controller should not depend on the model creation.

Test only the behaviour the controller should be responsible about:

Execution of particular methods
Data returned from the action - assigns, etc.

Result from the action - template render, redirect, etc.

```Ruby
# Example of a commonly used controller spec
# spec/controllers/articles_controller_spec.rb
# We are interested only in the actions the controller should perform
# So we are mocking the model creation and stubbing its methods
# And we concentrate only on the things the controller should do

describe ArticlesController do
  # The model will be used in the specs for all methods of the controller
  let(:article) { mock_model(Article) }

  describe 'POST create' do
    before { Article.stub(:new).and_return(article) }

    it 'creates a new article with the given attributes' do
      Article.should_receive(:new).with(title: 'The New Article Title').and_return(article)
      post :create, message: { title: 'The New Article Title' }
    end

    it 'saves the article' do
      article.should_receive(:save)
      post :create
    end

    it 'redirects to the Articles index' do
      article.stub(:save)
      post :create
      response.should redirect_to(action: 'index')
    end
  end
end
```

Use context when the controller action has different behaviour depending on the received params.

# A classic example for use of contexts in a controller spec is creation or update when the object saves successfully or not.

describe ArticlesController do
  let(:article) { mock_model(Article) }

  describe 'POST create' do
    before { Article.stub(:new).and_return(article) }

    it 'creates a new article with the given attributes' do
      Article.should_receive(:new).with(title: 'The New Article Title').and_return(article)
      post :create, article: { title: 'The New Article Title' }
    end

    it 'saves the article' do
      article.should_receive(:save)
      post :create
    end

    context 'when the article saves successfully' do
      before { article.stub(:save).and_return(true) }

      it 'sets a flash[:notice] message' do
        post :create
        flash[:notice].should eq('The article was saved successfully.')
      end

      it 'redirects to the Articles index' do
        post :create
        response.should redirect_to(action: 'index')
      end
    end

    context 'when the article fails to save' do
      before { article.stub(:save).and_return(false) }

      it 'assigns @article' do
        post :create
        assigns[:article].should be_eql(article)
      end

      it 're-renders the "new" template' do
        post :create
        response.should render_template('new')
      end
    end
  end
end

Models

Do not mock the models in their own specs.
Use fabrication to make real objects.
It is acceptable to mock other models or child objects.

Create the model for all examples in the spec to avoid duplication.

describe Article do
  let(:article) { Fabricate(:article) }
end

Add an example ensuring that the fabricated model is valid.

describe Article do
  it 'is valid with valid attributes' do
    article.should be_valid
  end
end

When testing validations, use have(x).errors_on to specify the attibute which should be validated. Using be_valid does not guarantee that the problem is in the intended attribute.

# bad
describe '#title' do
  it 'is required' do
    article.title = nil
    article.should_not be_valid
  end
end

# prefered
describe '#title' do
  it 'is required' do
    article.title = nil
    article.should have(1).error_on(:title)
  end
end

Add a separate describe for each attribute which has validations.

describe Article do
  describe '#title' do
    it 'is required' do
      article.title = nil
      article.should have(1).error_on(:title)
    end
  end
end

When testing uniqueness of a model attribute, name the other object another_object.

describe Article do
  describe '#title' do
    it 'is unique' do
      another_article = Fabricate.build(:article, title: article.title)
      article.should have(1).error_on(:title)
    end
  end
end

Mailers

The model in the mailer spec should be mocked. The mailer should not depend on the model creation.

The mailer spec should verify that:

the subject is correct
the receiver e-mail is correct
the e-mail is sent to the right e-mail address

the e-mail contains the required information

describe SubscriberMailer do
  let(:subscriber) { mock_model(Subscription, email: '[email protected]', name: 'John Doe') }

  describe 'successful registration email' do
    subject { SubscriptionMailer.successful_registration_email(subscriber) }

    its(:subject) { should == 'Successful Registration!' }
    its(:from) { should == ['info@your_site.com'] }
    its(:to) { should == [subscriber.email] }

    it 'contains the subscriber name' do
      subject.body.encoded.should match(subscriber.name)
    end
  end
end

Uploaders

What we can test about an uploader is whether the images are resized correctly. Here is a sample spec of a carrierwave image uploader:

# rspec/uploaders/person_avatar_uploader_spec.rb
require 'spec_helper'
require 'carrierwave/test/matchers'

describe PersonAvatarUploader do
  include CarrierWave::Test::Matchers

  # Enable images processing before executing the examples
  before(:all) do
    UserAvatarUploader.enable_processing = true
  end

  # Create a new uploader. The model is mocked as the uploading and resizing images does not depend on the model creation.
  before(:each) do
    @uploader = PersonAvatarUploader.new(mock_model(Person).as_null_object)
    @uploader.store!(File.open(path_to_file))
  end

  # Disable images processing after executing the examples
  after(:all) do
    UserAvatarUploader.enable_processing = false
  end

  # Testing whether image is no larger than given dimensions
  context 'the default version' do
    it 'scales down an image to be no larger than 256 by 256 pixels' do
      @uploader.should be_no_larger_than(256, 256)
    end
  end

  # Testing whether image has the exact dimensions
  context 'the thumb version' do
    it 'scales down an image to be exactly 64 by 64 pixels' do
      @uploader.thumb.should have_dimensions(64, 64)
    end
  end
end

Contributing

Nothing written in this guide is set in stone. It's my desire to work together with everyone interested in Rails coding style, so that we could ultimately create a resource that will be beneficial to the entire Ruby community.

Feel free to open tickets or send pull requests with improvements. Thanks in advance for your help!

License

This work is licensed under a Creative Commons Attribution 3.0 Unported License

Spread the Word

A community-driven style guide is of little use to a community that doesn't know about its existence. Tweet about the guide, share it with your friends and colleagues. Every comment, suggestion or opinion we get makes the guide just a little bit better. And we want to have the best possible guide, don't we?

Cheers,
Bozhidar

Name		Name	Last commit message	Last commit date
Latest commit History 169 Commits
README.md		README.md

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

Prelude

Table of Contents

Developing Rails applications

Configuration

Routing

Controllers

Models

ActiveRecord

Migrations

Views

Internationalization

Assets

Mailers

Bundler

Priceless Gems

Flawed Gems

Managing processes

Testing Rails applications

Cucumber

RSpec

Views

Controllers

Models

Mailers

Uploaders

Further Reading

Contributing

License

Spread the Word

About

Releases

Packages

bmclean/rails-style-guide

Folders and files

Latest commit

History

Repository files navigation

Prelude

Table of Contents

Developing Rails applications

Configuration

Routing

Controllers

Models

ActiveRecord

Migrations

Views

Internationalization

Assets

Mailers

Bundler

Priceless Gems

Flawed Gems

Managing processes

Testing Rails applications

Cucumber

RSpec

Views

Controllers

Models

Mailers

Uploaders

Further Reading

Contributing

License

Spread the Word

About

Resources

Stars

Watchers

Forks

Releases

Packages 0

Packages