Dump Hook

A library that helps cache your data received from performing user actions in tests. We use it in our acceptance tests to enhance performance of preparing data for the tests. It contains ways to cache Postgres and MySql DBs and you need to take care about other sources your own.

gem 'dump_hook'

For Rspec/Capybara add to rails_helper.rb:

require "dump_hook"

RSpec.configure do |config|
  config.include DumpHook
end

For Cucumber/Capybara add to env.rb:

require "dump_hook"

World(DumpHook)

There are several parameters to run and manage it.

DumpHook.setup do |config|
  config.database = ActiveRecord::Base.configurations[Rails.env]["database"]
  config.actual = Date.today.monday.to_s(:number)
end

• database - database name
• host - DB server
• port - DB server port
• username - user to connect
• password - password to connect, works just for MySql
• database_type - postgres/mysql, by default it's postgres

It looks like database is a single required parameter. Others may be default for your DB connection. In relation to password for postgres you need to use the url way in database parameter in order to set password. You may use the gem database_url to get this url.

config.database = postgresql://username:password@localhost/database

• actual - an attribute to manage how long or when you need to create a new dump for your actions. For instance, you may set it to Date.current to recreate your dumps every day. By default it's empty and your dumps will not recreate.
• remove_old_dumps - when actual is pointed your old dumps will be removed. By default it's true.
• dumps_location - by default it's tmp/dump_hook. You may pass something more exciting, e.g your current git branch or some path to store your dumps in your repo and generate them using CI.
• recreate - by default it's false. It recreates your dumps in the current session. It helps when you change execute_with_dump's body. You can set it through ENV variable DUMP_HOOK=recreate for some certain tests.

 $ DUMP_HOOK=recreate rails test test/system/your_test.rb

Attention! It works for clear DB. In case when you have some previous data you may come across with conflicts or incorrect data which may influence your tests. Yet one vital moment you cannot use the transactional way in your tests as dump_hook will not be able to store anything. Our benchmarks don't note some difference for our system tests and we prefer to use the same way as in the production environment. We hope you go by this rule too.

There is just single method what wrap your actions and restore dump again

execute_with_dump(name, opts={}, &block)

• name - identity what explain meaning of this background
• opts[:actual] - the same to the global parameter actual what override it
• opts[:created_on] - Date time to use in Time travelling for time dependent tests. In this case actual is unnecessary and ignored. If you want to recreate such dumps you need to clean them manually.

execute_with_dump("user_data") do
  create_superadmin
  login_as_superadmin
  go_to_users
  add_user("John Doe")
  add_user("Adam Smith")
  add_user("Robert Martin")
  add_user("Kent Beck")
  logout
end

You can use the capybara way or create a wrap step to have all actions in your features

Given(/^There is "(.*?)" background with:$/) do |name, steps_order|
  execute_with_dump(name) do
    steps steps_order.to_s
  end
end

and use it in your scenarios

Feature: User can modify project
  Background:
    Given There is "project_modification" background with:
    """
    There is the user data
    I login as superadmin
    I create a project "Time tracking"
    I logout
    """

  Scenario: I can rename project
    Given I login as "John Doe"
      And I go to the projects
      And I click "Edit" for "Time tracking" project
     When I fill out "name" with "Time management"
      And I click "Save"
     Then I see the notification "Project is updated"  

In this example There is the user data is step what wraps the capybara example

Disclaimer: We are working on enhancements

If you need to add yet one source, e.g "we have yet one service and need to dump its data" you may use something similar

module YourServiceHook
  def store_dump(filename)
    super
    data = YourServiceAPI.import
    File.open(your_service_filename(filename), "w") do |f| 
      f.write(data)
    end
  end

  def restore_dump(filename)
    super
    buffer = File.open(your_service_filename(filename)).read
    YourServiceAPI.export(buffer)
  end
  
  def your_service_filename
    filename.gsub(".dump", ".your_service")
  end
end

and add it after DumpHook module

include DumpHook
include YourServiceHook

• Dump other sources like cookies, Elastic Search, etc.
• Enhance ways of adding hooks to extend current sources without overriding(In progress)
• Add some sugar to create common dumps for CI or parallel_tests

The gem is available as open source under the terms of the MIT License.