# frozen_string_literal: true
require "rails/initializable"
require "active_support/descendants_tracker"
require "active_support/inflector"
require "active_support/core_ext/module/introspection"
require "active_support/core_ext/module/delegation"
module Rails
# +Rails::Railtie+ is the core of the \Rails framework and provides
# several hooks to extend \Rails and/or modify the initialization process.
#
# Every major component of \Rails (Action Mailer, Action Controller, Active
# Record, etc.) implements a railtie. Each of them is responsible for their
# own initialization. This makes \Rails itself absent of any component hooks,
# allowing other components to be used in place of any of the \Rails defaults.
#
# Developing a \Rails extension does _not_ require implementing a railtie, but
# if you need to interact with the \Rails framework during or after boot, then
# a railtie is needed.
#
# For example, an extension doing any of the following would need a railtie:
#
# * creating initializers
# * configuring a \Rails framework for the application, like setting a generator
# * adding <tt>config.*</tt> keys to the environment
# * setting up a subscriber with ActiveSupport::Notifications
# * adding Rake tasks
#
# == Creating a Railtie
#
# To extend \Rails using a railtie, create a subclass of +Rails::Railtie+.
# This class must be loaded during the \Rails boot process, and is conventionally
# called +MyNamespace::Railtie+.
#
# The following example demonstrates an extension which can be used with or
# without \Rails.
#
# # lib/my_gem/railtie.rb
# module MyGem
# class Railtie < Rails::Railtie
# end
# end
#
# # lib/my_gem.rb
# require "my_gem/railtie" if defined?(Rails::Railtie)
#
# == Initializers
#
# To add an initialization step to the \Rails boot process from your railtie, just
# define the initialization code with the +initializer+ macro:
#
# class MyGem::Railtie < Rails::Railtie
# initializer "my_gem.configure_rails_initialization" do
# # some initialization behavior
# end
# end
#
# If specified, the block can also receive the application object, in case you
# need to access some application-specific configuration, like middleware:
#
# class MyGem::Railtie < Rails::Railtie
# initializer "my_gem.configure_rails_initialization" do |app|
# app.middleware.use MyGem::Middleware
# end
# end
#
# Finally, you can also pass <tt>:before</tt> and <tt>:after</tt> as options to
# +initializer+, in case you want to couple it with a specific step in the
# initialization process.
#
# == Configuration
#
# Railties can access a config object which contains configuration shared by all
# railties and the application:
#
# class MyGem::Railtie < Rails::Railtie
# # Customize the ORM
# config.app_generators.orm :my_gem_orm
#
# # Add a to_prepare block which is executed once in production
# # and before each request in development.
# config.to_prepare do
# MyGem.setup!
# end
# end
#
# == Loading Rake Tasks and Generators
#
# If your railtie has Rake tasks, you can tell \Rails to load them through the method
# +rake_tasks+:
#
# class MyGem::Railtie < Rails::Railtie
# rake_tasks do
# load "path/to/my_gem.tasks"
# end
# end
#
# By default, \Rails loads generators from your load path. However, if you want to place
# your generators at a different location, you can specify in your railtie a block which
# will load them during normal generators lookup:
#
# class MyGem::Railtie < Rails::Railtie
# generators do
# require "path/to/my_gem_generator"
# end
# end
#
# Since filenames on the load path are shared across gems, be sure that files you load
# through a railtie have unique names.
#
# == Run another program when the \Rails server starts
#
# In development, it's very usual to have to run another process next to the \Rails Server. In example
# you might want to start the Webpack or React server. Or maybe you need to run your job scheduler process
# like Sidekiq. This is usually done by opening a new shell and running the program from here.
#
# \Rails allow you to specify a +server+ block which will get called when a \Rails server starts.
# This way, your users don't need to remember to have to open a new shell and run another program, making
# this less confusing for everyone.
# It can be used like this:
#
# class MyGem::Railtie < Rails::Railtie
# server do
# WebpackServer.start
# end
# end
#
# == Application and Engine
#
# An engine is nothing more than a railtie with some initializers already set. And since
# Rails::Application is an engine, the same configuration described here can be
# used in both.
#
# Be sure to look at the documentation of those specific classes for more information.
class Railtie
autoload :Configuration, "rails/railtie/configuration"
extend ActiveSupport::DescendantsTracker
include Initializable
ABSTRACT_RAILTIES = %w(Rails::Railtie Rails::Engine Rails::Application)
class << self
private :new
delegate :config, to: :instance
def subclasses
super.reject(&:abstract_railtie?).sort
end
def rake_tasks(&blk)
register_block_for(:rake_tasks, &blk)
end
def console(&blk)
register_block_for(:load_console, &blk)
end
def runner(&blk)
register_block_for(:runner, &blk)
end
def generators(&blk)
register_block_for(:generators, &blk)
end
def server(&blk)
register_block_for(:server, &blk)
end
def abstract_railtie?
ABSTRACT_RAILTIES.include?(name)
end
def railtie_name(name = nil)
@railtie_name = name.to_s if name
@railtie_name ||= generate_railtie_name(self.name)
end
# Since Rails::Railtie cannot be instantiated, any methods that call
# +instance+ are intended to be called only on subclasses of a Railtie.
def instance
@instance ||= new
end
# Allows you to configure the railtie. This is the same method seen in
# Railtie::Configurable, but this module is no longer required for all
# subclasses of Railtie so we provide the class method here.
def configure(&block)
instance.configure(&block)
end
def <=>(other) # :nodoc:
load_index <=> other.load_index
end
def inherited(subclass)
subclass.increment_load_index
super
end
protected
attr_reader :load_index
def increment_load_index
@@load_counter ||= 0
@load_index = (@@load_counter += 1)
end
private
def generate_railtie_name(string)
ActiveSupport::Inflector.underscore(string).tr("/", "_")
end
def respond_to_missing?(name, _)
return super if abstract_railtie?
instance.respond_to?(name) || super
end
# If the class method does not have a method, then send the method call
# to the Railtie instance.
def method_missing(name, *args, &block)
if !abstract_railtie? && instance.respond_to?(name)
instance.public_send(name, *args, &block)
else
super
end
end
ruby2_keywords(:method_missing)
# receives an instance variable identifier, set the variable value if is
# blank and append given block to value, which will be used later in
# `#each_registered_block(type, &block)`
def register_block_for(type, &blk)
var_name = "@#{type}"
blocks = instance_variable_defined?(var_name) ? instance_variable_get(var_name) : instance_variable_set(var_name, [])
blocks << blk if blk
blocks
end
end
delegate :railtie_name, to: :class
def initialize # :nodoc:
if self.class.abstract_railtie?
raise "#{self.class.name} is abstract, you cannot instantiate it directly."
end
end
def inspect # :nodoc:
"#<#{self.class.name}>"
end
def configure(&block) # :nodoc:
instance_eval(&block)
end
# This is used to create the <tt>config</tt> object on Railties, an instance of
# Railtie::Configuration, that is used by Railties and Application to store
# related configuration.
def config
@config ||= Railtie::Configuration.new
end
def railtie_namespace # :nodoc:
@railtie_namespace ||= self.class.module_parents.detect { |n| n.respond_to?(:railtie_namespace) }
end
protected
def run_console_blocks(app) # :nodoc:
each_registered_block(:console) { |block| block.call(app) }
end
def run_generators_blocks(app) # :nodoc:
each_registered_block(:generators) { |block| block.call(app) }
end
def run_runner_blocks(app) # :nodoc:
each_registered_block(:runner) { |block| block.call(app) }
end
def run_tasks_blocks(app) # :nodoc:
extend Rake::DSL
each_registered_block(:rake_tasks) { |block| instance_exec(app, &block) }
end
def run_server_blocks(app) # :nodoc:
each_registered_block(:server) { |block| block.call(app) }
end
private
# run `&block` in every registered block in `#register_block_for`
def each_registered_block(type, &block)
klass = self.class
while klass.respond_to?(type)
klass.public_send(type).each(&block)
klass = klass.superclass
end
end
end
end