[{"_1":2,"_3":-5,"_4":-5},"loaderData",{"_5":6,"_7":8},"actionData","errors","root",{"_11":-5},"routes/user.$name.replies",{"_9":10,"_11":12,"_13":14,"_15":16,"_17":18,"_19":16},"replies",[34],"user",{"_20":21,"_13":14,"_22":23,"_24":14,"_25":26,"_27":28,"_29":30,"_31":16,"_32":33},"loginname","chenjin3","page",1,"limit",50,"total","id","20658","avatar_url","https://avatars.githubusercontent.com/u/8558291?v=3","githubUsername","create_at","2016-01-24 11:30:37.318","score",5,"topic_count",0,"reply_count","is_star",false,{"_20":35,"_36":37,"_38":-5,"_39":40,"_41":42,"_43":44,"_45":33,"_46":33,"_31":47,"_48":49,"_25":50,"_51":52},"110","author_id","27","tab","content","身在天朝，难免会用到代理的时候。比如在学校内网用代理免费上外网，在墙内用代理上404网站等。\r\n
\r\n
现在使用的代理大部分为HTTP和Socket代理。 Socket代理更底层，需要本地解析域名，而HTTP代理则是基于HTTP协议之上的，不需要本地解析域名。下面我讲讲HTTP(S)代理的设计思路以及NodeJS代码实现。\r\n
\r\n
HTTP协议\r\n
\r\n
HTTP协议简单说来就是浏览器把一串字符串发送到目标服务器，然后把目标服务器返回回来的一串字符串显示给用户。\r\n
\r\n
浏览器发送的这串字符主要分为两个部分，一部分是头，里面包含目标服务器域名，当前请求的文件路径等信息。另一部分是正文，一般的GET请求没有正文。\r\n
\r\n
服务器返回来的字符串也分为头和正文。\r\n
\r\n
HTTP代理原理\r\n
\r\n
HTTP代理需要做的事情就是接收浏览器发来的请求字符串，再从请求字符串的头部分找出浏览器请求的目标主机，然后直接把这串请求字符串发给目标主机，再把目标主机返回的数据发给浏览器。 “什么？就这么简单？” “呃。。是啊，但这还没完。。”\r\n
\r\n
现代浏览器一般都是默认采用HTTP/1.1版本，并且默认会发送Connection: keep-alive请求。这些信息是写在请求的头部的，意思是通知目标服务器采用keep-alive技术继续处理后续的请求。但是我们做的代理程序要想支持keep-alive是比较麻烦的。所以干脆就把这个篡改成Connection: close。这样就可以保证浏览器请求的每个文件都会单独发送一个HTTP请求。\r\n
\r\n
下面是NodeJS代码实现\r\n

var net = require('net');\r\n
var local_port = 8893;\r\n
\r\n
//在本地创建一个server监听本地local_port端口\r\n
net.createServer(function (client)\r\n
{\r\n
\r\n
\t//首先监听浏览器的数据发送事件，直到收到的数据包含完整的http请求头\r\n
\tvar buffer = new Buffer(0);\r\n
\tclient.on('data',function(data)\r\n
\t{\r\n
\t\tbuffer = buffer_add(buffer,data);\r\n
\t\tif (buffer_find_body(buffer) == -1) return;\r\n
\t\tvar req = parse_request(buffer);\r\n
\t\tif (req === false) return;\r\n
\t\tclient.removeAllListeners('data');\r\n
\t\trelay_connection(req);\r\n
\t});\r\n
\r\n
\t//从http请求头部取得请求信息后，继续监听浏览器发送数据，同时连接目标服务器，并把目标服务器的数据传给浏览器\r\n
\tfunction relay_connection(req)\r\n
\t{\r\n
\t\tconsole.log(req.method+' '+req.host+':'+req.port);\r\n
\r\n
\t\t//如果请求不是CONNECT方法（GET, POST），那么替换掉头部的一些东西\r\n
\t\tif (req.method != 'CONNECT')\r\n
\t\t{\r\n
\t\t\t//先从buffer中取出头部\r\n
\t\t\tvar _body_pos = buffer_find_body(buffer);\r\n
\t\t\tif (_body_pos < 0) _body_pos = buffer.length;\r\n
\t\t\tvar header = buffer.slice(0,_body_pos).toString('utf8');\r\n
\t\t\t//替换connection头\r\n
\t\t\theader = header.replace(/(proxy\\-)?connection\\:.+\\r\\n/ig,'')\r\n
\t\t\t\t\t.replace(/Keep\\-Alive\\:.+\\r\\n/i,'')\r\n
\t\t\t\t\t.replace(\"\\r\\n\",'\\r\\nConnection: close\\r\\n');\r\n
\t\t\t//替换网址格式(去掉域名部分)\r\n
\t\t\tif (req.httpVersion == '1.1')\r\n
\t\t\t{\r\n
\t\t\t\tvar url = req.path.replace(/http\\:\\/\\/[^\\/]+/,'');\r\n
\t\t\t\tif (url.path != url) header = header.replace(req.path,url);\r\n
\t\t\t}\r\n
\t\t\tbuffer = buffer_add(new Buffer(header,'utf8'),buffer.slice(_body_pos));\r\n
\t\t}\r\n
\r\n
\t\t//建立到目标服务器的连接\r\n
\t\tvar server = net.createConnection(req.port,req.host);\r\n
\t\t//交换服务器与浏览器的数据\r\n
\t\tclient.on(\"data\", function(data){ server.write(data); });\r\n
\t\tserver.on(\"data\", function(data){ client.write(data); });\r\n
\r\n
\t\tif (req.method == 'CONNECT')\r\n
\t\t\tclient.write(new Buffer(\"HTTP/1.1 200 Connection established\\r\\nConnection: close\\r\\n\\r\\n\"));\r\n
\t\telse\r\n
\t\t\tserver.write(buffer);\r\n
\t}\r\n
}).listen(local_port);\r\n
\r\n
console.log('Proxy server running at localhost:'+local_port);\r\n
\r\n
//处理各种错误\r\n
process.on('uncaughtException', function(err)\r\n
{\r\n
\tconsole.log(\"\\nError!!!!\");\r\n
\tconsole.log(err);\r\n
});\r\n
\r\n
/**\r\n
* 从请求头部取得请求详细信息\r\n
* 如果是 CONNECT 方法，那么会返回 { method,host,port,httpVersion}\r\n
* 如果是 GET/POST 方法，那么返回 { metod,host,port,path,httpVersion}\r\n
*/\r\n
function parse_request(buffer)\r\n
{\r\n
\tvar s = buffer.toString('utf8');\r\n
\tvar method = s.split('\\n')[0].match(/^([A-Z]+)\\s/)[1];\r\n
\tif (method == 'CONNECT')\r\n
\t{\r\n
\t\tvar arr = s.match(/^([A-Z]+)\\s([^\\:\\s]+)\\:(\\d+)\\sHTTP\\/(\\d\\.\\d)/);\r\n
\t\tif (arr && arr[1] && arr[2] && arr[3] && arr[4])\r\n
\t\t\treturn { method: arr[1], host:arr[2], port:arr[3],httpVersion:arr[4] };\r\n
\t}\r\n
\telse\r\n
\t{\r\n
\t\tvar arr = s.match(/^([A-Z]+)\\s([^\\s]+)\\sHTTP\\/(\\d\\.\\d)/);\r\n
\t\tif (arr && arr[1] && arr[2] && arr[3])\r\n
\t\t{\r\n
\t\t\tvar host = s.match(/Host\\:\\s+([^\\n\\s\\r]+)/)[1];\r\n
\t\t\tif (host)\r\n
\t\t\t{\r\n
\t\t\t\tvar _p = host.split(':',2);\r\n
\t\t\t\treturn { method: arr[1], host:_p[0], port:_p[1]?_p[1]:80, path: arr[2],httpVersion:arr[3] };\r\n
\t\t\t}\r\n
\t\t}\r\n
\t}\r\n
\treturn false;\r\n
}\r\n
\r\n
/**\r\n
* 两个buffer对象加起来\r\n
*/\r\n
function buffer_add(buf1,buf2)\r\n
{\r\n
\tvar re = new Buffer(buf1.length + buf2.length);\r\n
\tbuf1.copy(re);\r\n
\tbuf2.copy(re,buf1.length);\r\n
\treturn re;\r\n
}\r\n
\r\n
/**\r\n
* 从缓存中找到头部结束标记(\"\\r\\n\\r\\n\")的位置\r\n
*/\r\n
function buffer_find_body(b)\r\n
{\r\n
\tfor(var i=0,len=b.length-3;i < len;i++)\r\n
\t{\r\n
\t\tif (b[i] == 0x0d && b[i+1] == 0x0a && b[i+2] == 0x0d && b[i+3] == 0x0a)\r\n
\t\t{\r\n
\t\t\treturn i+4;\r\n
\t\t}\r\n
\t}\r\n
\treturn -1;\r\n
}

\r\n
另外，可以用 \"nohup node some.js > /dev/null &\" 命令让nodejs程序在后台运行。","title","用NodeJS实现HTTP/HTTPS代理","last_reply_at","2016-01-24 11:32:10.159","good","top",6,"visit_count",44063,"2011-05-08 15:24:42","author",{"_13":53,"_22":54},"longbill","http://www.gravatar.com/avatar/4c564c9b7fad37a790b560904df49c13?s=48"]