Rails源码阅读(二)_script/server
如果明白了script/console,那么理解script/server会轻松些;
区别在于,启动console其实是需要加载irb,而启动server则要处理http请求;
http-server:动手写rails(一)_Rack标准和HttpServer之WEBrick
script/server的代码:
require File.expand_path('../../config/boot', __FILE__)
require 'commands/server'
第一行是准备环境, 第二行才是启动server
commands/server.rb的主要代码:
server = Rack::Handler.get(ARGV.first) rescue nil
unless server
begin
server = Rack::Handler::Mongrel
rescue LoadError => e
server = Rack::Handler::WEBrick
end
end
puts "=> Booting #{ActiveSupport::Inflector.demodulize(server)}"
puts "=> Rails #{Rails.version} application starting on http://#{options[:Host]}:#{options[:Port]}#{options[:path]}"
%w(cache pids sessions sockets).each do |dir_to_make|
FileUtils.mkdir_p(File.join(RAILS_ROOT, 'tmp', dir_to_make))
end
if options[:detach]
Process.daemon
pid = "#{RAILS_ROOT}/tmp/pids/server.pid"
File.open(pid, 'w'){ |f| f.write(Process.pid) }
at_exit { File.delete(pid) if File.exist?(pid) }
end
ENV["RAILS_ENV"] = options[:environment]
RAILS_ENV.replace(options[:environment]) if defined?(RAILS_ENV)
if File.exist?(options[:config])
config = options[:config]
if config =~ /\.ru$/
cfgfile = File.read(config)
if cfgfile[/^#\\(.*)/]
opts.parse!($1.split(/\s+/))
end
inner_app = eval("Rack::Builder.new {( " + cfgfile + "\n )}.to_app", nil, config)
else
require config
inner_app = Object.const_get(File.basename(config, '.rb').capitalize)
end
else
require RAILS_ROOT + "/config/environment"
inner_app = ActionController::Dispatcher.new
end
if options[:path].nil?
map_path = "/"
else
ActionController::Base.relative_url_root = options[:path]
map_path = options[:path]
end
app = Rack::Builder.new {
use Rails::Rack::LogTailer unless options[:detach]
use Rails::Rack::Debugger if options[:debugger]
map map_path do
use Rails::Rack::Static
run inner_app
end
}.to_app
puts "=> Call with -d to detach"
trap(:INT) { exit }
puts "=> Ctrl-C to shutdown server"
begin
server.run(app, options.merge(:AccessLog => []))
ensure
puts 'Exiting'
end
#(1) 可以指定maping的path,这样在rack中就map到这个路径:map map_path
if options[:path].nil?
map_path = "/"
else
ActionController::Base.relative_url_root = options[:path]
map_path = options[:path]
end
#(2) 在启动rack之前,先要加载rails环境和组建
require RAILS_ROOT + "/config/environment"
#(3) 使用rack启动,用WEBrick做web服务,见下面分析。
commands/server.rb的启动,除了一些配置如端口等外,更重要的是用到了两个东西:
#1 server:WEBrick|Mongrel,用来处理http等
server = Rack::Handler.get(ARGV.first) rescue nil
unless server
begin
server = Rack::Handler::Mongrel
rescue LoadError => e
server = Rack::Handler::WEBrick
end
end
看看WEBrick的源码:继承于::WEBrick::HTTPServlet::AbstractServlet
module Rack
module Handler
class WEBrick < ::WEBrick::HTTPServlet::AbstractServlet
def self.run(app, options={})
end
def self.shutdown
end
def initialize(server, app)
super server
@app = Rack::ContentLength.new(app)
end
def service(req, res)
#2 rack 和 middleware
rack的应用上来说,比较简单直接:
#1
rackXXX = generate_a_rack_app
#2
builder = Rack::Builder.new{
user middlewareXXX
... ...
run rackXXX
}
#3
Rack::Handler::HandlerXXX.run builder
对应的server启动代码为:
#1 generate a racke proc
# 这里只要实现call方法就行
inner_app = ActionController::Dispatcher.new #Dispatcher实现了call方法
#2 RackBuilder
app = Rack::Builder.new {
use Rails::Rack::LogTailer unless options[:detach]
use Rails::Rack::Debugger if options[:debugger]
map map_path do
use Rails::Rack::Static
run inner_app
end
}.to_app
#3 RackHandler
server = Rack::Handler.get(ARGV.first) rescue nil
unless server
begin
server = Rack::Handler::Mongrel
rescue LoadError => e
server = Rack::Handler::WEBrick
end
end
server.run(app, options.merge(:AccessLog => []))
这里要明白流程:
启动server是要启动什么?
HttpServer(CGIserver),这里就是WEBrick或者Mongrel
server怎么和ruby联系上了?
WEBrick等遵循rack标准,server.run(rack_app, options)
ruby写一个rack的app即可
rails里的rack是什么?
是:ActionController::Dispatcher
出处代码:inner_app = ActionController::Dispatcher.new
还需要哪些知识?
rack的介绍,原理,好处,标准,使用等
WEBrick活Mongrel的介绍和使用
====结束====
=== ===
== ==
= =
| |