从Rails的svn资源库下载最新的Rails，我们会发现多了个activeresource包
从此Rails核心模块变为: ActiveRecord、ActionPack、ActionWebService、AcionMailer、ActiveResource、ActiveSupport和Railties

这里有几个问题:

一，ActionController里的resources.rb文件与ActiveResource是什么关系?

当初我看到Edge Rails里多了activeresource这个包时，我第一反应就是ActionController::Resources被ActiveResource替代了吧?
但是我去ActionPack里一看，resources.rb文件依然存在，原来我的判断错了: ActionController::Resources != ActiveResource

ActionController::Resources的工作就是创建了map.resource(s)这个DSL
这个DSL的底层含义是为一组或单个resource(其实是一组或单个对象)做两样事情:
a，生成一组named routes命名
b，为一组http动词(:get/:post/:put/:delete)做一个到一组action(index/new/create/show/edit/update/destroy)的映射
ActionController::Resources是REST风格的Rails实现，它的目的就是让我们做RESTful风格的Rails开发
按ActionController::Resources的指导，我们依照HTTP method规范开发系统，可以对html和xml两种response格式render内容
但我们慢慢发现，ActionController::Resources只为我们提供了REST web service的发布功能，而没有提供消费功能!

ActiveResource的工作看看官方说法则不言而喻了:

ActiveResource不是ActionController::Resources的替代者，而是REST web service的消费者
Active Resource是一个proxy的角色，它连接model与service，加上一点胶水，让REST web service的调用简化无比
Active Resource的架构类似于ActiveRecord，等会我们看看代码来加深认识

二，Rails有ActionWebService，还要ActiveResource干什么?

这个问题是刚才在JavaEye上看到的，其实应该说ActionWebService和ActionController::Resources+ActiveResource才有可比性
最大的区别就是ActionWebService是基于SOAP的，而ActionController::Resources和ActiveResource则是基于轻量的REST、XML的
ActionWebService是Rails对SOAP web service的封装，并提供Client端调用API
不过ActionWebService也支持XML-RPC协议

三，ActiveResource尝鲜

看代码吧!

class Person < ActiveResource::Base
  self.site = "http://api.people.com:3000/"
end

Person类继承ActiveResource::Base，然后调用site=这个类方法，参数为一个URL
就这么简单，现在Person已经映射到如下RESTful resources位置: http://api.people.com:3000/people/
==== Create:

   p = Person.create(:name => 'Robbin')

   # POST http://api.people.com:3000/people.xml
   # Submit:
   # <person><name>Robbin</name></person>

==== Find:

   p = Person.find(1)

   # GET http://api.people.com:3000/people/1.xml
   # Response:
   # <person><id>1</id><name>Robbin</name></person>

   people = Person.find(:all)

   # GET http://api.people.com:3000/people.xml
   # Response:
   # <people>
   #  <person><id>1</id><first>Robbin</first></person>
   #  <person><id>2</id><first>Gigix</first></person>
   # </people>

==== Update:

   p = Person.find(1)
   p.name = 'Hideto'
   p.save

   # PUT http://api.people.com:3000/people/1.xml
   # Submit:
   # <person><name>Hideto</name></person>

==== Delete:

   p = Person.find(2)
   p.destroy

   # DELETE http://api.people.com:3000/people/2.xml

ActiveResource基于HTTP协议并充分使用HTTP规范里的动词:
* GET requests are used for finding and retrieving resources.
* POST requests are used to create new resources.
* PUT requests are used to update existing resources.
* DELETE requests are used to delete resources.

而Person类就是对http://api.people.com:3000/people/链接上的REST web service的代理，我们可以像使用ActiveRecord对象一样使用ActiveResource对象

四、ActiveResource源码分析一二
active_resource/base.rb:

module ActiveResource
  class Base

    class << self
      def site
        if defined?(@site)
          @site
        elsif superclass != Object and superclass.site
          superclass.site.dup.freeze
        end
      end
      def site=(site)
        @connection = nil
        @site = site.nil? ? nil : create_site_uri_from(site)
      end
      def create(attributes = {})
        returning(self.new(attributes)) { |res| res.save }        
      end
      def find(*arguments)
        scope   = arguments.slice!(0)
        options = arguments.slice!(0) || {}
        case scope
          when :all   then find_every(options)
          when :first then find_every(options).first
          when :one   then find_one(options)
          else             find_single(scope, options)
        end
      end
      private
        def find_every(options)
          case from = options[:from]
          when Symbol
            instantiate_collection(get(from, options[:params]))
          when String
            path = "#{from}#{query_string(options[:params])}"
            instantiate_collection(connection.get(path, headers) || [])
          else
            prefix_options, query_options = split_options(options[:params])
            path = collection_path(prefix_options, query_options)
            instantiate_collection( (connection.get(path, headers) || []), prefix_options )
          end
        end
        def create_site_uri_from(site)
          site.is_a?(URI) ? site.dup : URI.parse(site)
        end
    end

    attr_accessor :attributes

    def destroy
      connection.delete(element_path, self.class.headers)
    end

  end
end

我们看到find、create、destroy都是使用connection来做具体的操作，让我们再看connection.rb:

module ActiveResource
  class Connection

    def get(path, headers = {})
      xml_from_response(request(:get, path, build_request_headers(headers)))
    end
    def delete(path, headers = {})
      request(:delete, path, build_request_headers(headers))
    end
    def put(path, body = '', headers = {})
      request(:put, path, body.to_s, build_request_headers(headers))
    end
    def post(path, body = '', headers = {})
      request(:post, path, body.to_s, build_request_headers(headers))
    end
    private
      def http
        unless @http
          @http             = Net::HTTP.new(@site.host, @site.port)
          @http.use_ssl     = @site.is_a?(URI::HTTPS)
          @http.verify_mode = OpenSSL::SSL::VERIFY_NONE if @http.use_ssl
        end
        @http
      end
      def request(method, path, *arguments)
        logger.info "#{method.to_s.upcase} #{site.scheme}://#{site.host}:#{site.port}#{path}" if logger
        result = nil
        time = Benchmark.realtime { result = http.send(method, path, *arguments) }
        logger.info "--> #{result.code} #{result.message} (#{result.body.length}b %.2fs)" % time if logger
        handle_response(result)
      end
      def handle_response(response)
        case response.code.to_i
          when 200...400
            response
          when 404
            raise(ResourceNotFound.new(response))
          when 405
            raise(MethodNotAllowed.new(response))
          when 409
            raise(ResourceConflict.new(response))
          when 422
            raise(ResourceInvalid.new(response))
          when 401...500
            raise(ClientError.new(response))
          when 500...600
            raise(ServerError.new(response))
          else
            raise(ConnectionError.new(response, "Unknown response code: #{response.code}"))
        end
      end
  end
end

核心代码就是这些，最后只是调用Net::HTTP对象的get/post/put/delete方法，socket连接而已
主要是我们的REST web service host(如http://api.people.com:3000/)已经成功发布web服务，我们访问相应的url即可(如people.xml、people/1.xml)
ActiveResource以我们熟悉的ActiveRecord操作方式来封装和架构我们的REST web service客户端API以简化开发，我们只能说两个字:佩服!

五、ActionController::Resources的一些改动

Changeset 6485
源码:

def map_collection_actions(map, resource)
 resource.collection_methods.each do |method, actions|
  actions.each do |action|
   action_options = action_options_for(action, resource, method)

   map.deprecated_named_route("#{action}_#{resource.name_prefix}#{resource.plural}", "#{resource.name_prefix}#{action}_#{resource.plural}", "#{resource.path}/#{action}", action_options)
   map.deprecated_named_route("formatted_#{action}_#{resource.name_prefix}#{resource.plural}", "formatted_#{resource.name_prefix}#{action}_#{resource.plural}", "#{resource.path}/#{action}.:format", action_options)
  end
 end
end

“/messages/1;edit”这种url已经替换成“/messages/1/edit”，早该这样了不是吗?

Ticket #8251
源码例子和上面一样，name_prefix放到path的最前面去了，action名字挪到它后面