python2.7下urllib2的connection自动close的原因及解决办法
python2.7下urllib2的connection自动close的原因及解决办法
前文介绍了urllib2的常见问题的解决方案，今天来特别讨论下urllib2中短连接问题。
1、urllib2代码
如下文代码所示，自定义 'Connection': 'keep-alive'，通知服务器交互结束后，不断开连接，即所谓长连接。
1 #测试8 使用urllib2 测试Connection=keep-alive
2 import urllib2
3 import cookielib
4
5
6 httpHandler = urllib2.HTTPHandler(debuglevel=1)
7 httpsHandler = urllib2.HTTPSHandler(debuglevel=1)
8 opener = urllib2.build_opener(httpHandler, httpsHandler)
9 urllib2.install_opener(opener)
10
11 loginHeaders={
12 'User-Agent': 'Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.0 Chrome/30.0.1599.101 Safari/537.36',
13 'Referer': 'https://www.baidu.com',
14 'Connection': 'keep-alive'
15 }
16 request=urllib2.Request('http://www.suning.com.cn',headers=loginHeaders)
17 response = urllib2.urlopen(request)
18 page=''
19 page= response.read()
20 print response.info()
21 print page输出报文：
注意日志中划线部分，可以看到请求报文其他头部，例如User-agent已被修改成功，但connection仍然保持close
- Connection: close
- header: Connection: close
1 send: 'GET / HTTP/1.1\r\nAccept-Encoding: identity\r\nHost: www.suning.com.cn\r\nConnection: close\r\nUser-Agent: Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.0 Chrome/30.0.1599.101 Safari/537.36\r\n\r\n'
2 reply: 'HTTP/1.1 200 OK\r\n'
3 header: Connection: close
4 header: Transfer-Encoding: chunked
5 header: Expires: Thu, 19 Nov 1981 08:52:00 GMT
6 header: Date: Sun, 15 May 2016 06:09:43 GMT
7 header: Content-Type: text/html; charset=utf-8
8 header: Server: nginx/1.2.9
9 header: Vary: Accept-Encoding
10 header: X-Powered-By: ThinkPHP
11 header: Set-Cookie: PHPSESSID=tv7ced9sbu7lph60o8ilfj4207; path=/
12 header: Cache-Control: private
13 header: Pragma: no-cache
14 <!doctype html>
15 <html lang="zh-CN">
16 <head>
17 <meta charset="utf-8">
18 <meta http-equiv="X-UA-Compatible" content="IE=edge,chrome=1" />2、httplib2写法代码
换成httplib2协议的代码，当然这也是urllib2不支持keep-alive的解决办法之一，另一个方法是Requests。
1 #测试9 使用httplib2测试Connection=keep-alive
2 import httplib2
3
4 ghttp = httplib2.Http()
5 httplib2.debuglevel=1
6 loginHeaders={
7 'User-Agent': 'Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.0 Chrome/30.0.1599.101 Safari/537.36',
8 'Connection': 'keep-alive'
9 }
10
11 response ,page= ghttp.request('http://www.suning.com.cn',headers=loginHeaders )
12 print page.decode('utf-8')可以看到输出中，长连接设置成功。
- header: Connection: Keep-Alive
1 connect: (www.suning.com.cn, 80) ************
2 send: 'GET / HTTP/1.1\r\nHost: www.suning.com.cn\r\nconnection: keep-alive\r\naccept-encoding: gzip, deflate\r\nuser-agent: Mozilla/5.0 (Windows NT 5.1) AppleWebKit/537.36 (KHTML, like Gecko) Maxthon/4.0 Chrome/30.0.1599.101 Safari/537.36\r\n\r\n'
3 reply: 'HTTP/1.1 200 OK\r\n'
4 header: Connection: Keep-Alive
5 header: Transfer-Encoding: chunked
6 header: Expires: Thu, 19 Nov 1981 08:52:00 GMT
7 header: Date: Sun, 15 May 2016 06:15:16 GMT
8 header: Content-Type: text/html; charset=utf-8
9 header: Server: nginx/1.2.9
10 header: Vary: Accept-Encoding
11 header: X-Powered-By: ThinkPHP
12 header: Set-Cookie: PHPSESSID=ep580j8bq3uba0ud3fe3rgu5i5; path=/
13 header: Cache-Control: private
14 header: Pragma: no-cache3、分析原因
还是上urllib2的源码吧，可以看到在do_open核心方法中，connection被写死成了close。 至于原因就是上面那一堆注释，大概意思是addinfourl这个类一旦启用长链接，可以读取到上次交互未读完的应答报文，为了防止此类情况，所以强制性将Connection写死成close。 def do_open(self, http_class, req, **http_conn_args): …… # We want to make an HTTP/1.1 request, but the addinfourl # class isn't prepared to deal with a persistent connection. # It will try to read all remaining data from the socket, # which will block while the server waits for the next request. # So make sure the connection gets closed after the (only) # request. headers["Connection"] = "close" headers = dict((name.title(), val) for name, val in headers.items()) if req._tunnel_host: tunnel_headers = {} proxy_auth_hdr = "Proxy-Authorization" if proxy_auth_hdr in headers: tunnel_headers[proxy_auth_hdr] = headers[proxy_auth_hdr] # Proxy-Authorization should not be sent to origin # server. del headers[proxy_auth_hdr] h.set_tunnel(req._tunnel_host, headers=tunnel_headers) try: h.request(req.get_method(), req.get_selector(), req.data, headers)