Python Mitmproxy出现奇怪的Web驱动程序问题
下面是我的代码的浓缩版本,显示了webdriver和mitmproxy的一个问题:Python Mitmproxy出现奇怪的Web驱动程序问题,python,selenium-webdriver,timeout,webdriver,mitmproxy,Python,Selenium Webdriver,Timeout,Webdriver,Mitmproxy,下面是我的代码的浓缩版本,显示了webdriver和mitmproxy的一个问题: from libmproxy import controller from libmproxy import proxy import os from multiprocessing import Process from selenium import webdriver from selenium.webdriver.common.proxy import * import time # traffic
from libmproxy import controller
from libmproxy import proxy
import os
from multiprocessing import Process
from selenium import webdriver
from selenium.webdriver.common.proxy import *
import time
# traffic inception
class HttpMitmReporter(controller.Master):
def __init__(self, proxy_server):
controller.Master.__init__(self, proxy_server)
def run(self):
try:
return controller.Master.run(self)
except KeyboardInterrupt:
self.shutdown()
# setup browser with proxy settings etc.
def setup_browser():
global driver
proper_setup = False
proxy_con = "127.0.0.1:8088"
my_proxy = Proxy({'proxyType': ProxyType.MANUAL,
'httpProxy': proxy_con,
'ftpProxy': proxy_con,
'sslProxy': proxy_con,
'noProxy': ''})
while not proper_setup:
try:
driver = webdriver.Firefox(proxy=my_proxy)
driver.set_page_load_timeout(20)
driver.set_script_timeout(20)
proper_setup = True
except Exception as exception:
print exception
try:
driver.quit()
except Exception:
time.sleep(3)
driver = None
# start proxy
config = proxy.ProxyConfig(
cacert=os.path.expanduser("~/.mitmproxy/mitmproxy-ca.pem")
)
server = proxy.ProxyServer(config, 8088)
reporter = HttpMitmReporter(server)
mitm_proxy = Process(target=reporter.run)
mitm_proxy.start()
setup_browser()
# urls to test
urls = [
'http://support.microsoft.com/common/survey.aspx?showpage=1&scid=sw%3Bde%3B3663&altStyle=narrow&renderOption=OverrideDefault&fr=1',
'http://outlook.com/',
]
# visit sites
for url in urls:
print url
try:
driver.get(url)
elements = driver.find_elements_by_tag_name('a')
except Exception as ex:
print ex
driver.quit()
setup_browser()
# terminate browser and proxy
driver.quit()
mitm_proxy.terminate()
如果执行此操作,您将看到firefox正在加载forever outlook.com,并且不会触发超时。该问题仅在mitmproxy中出现。我已经测试了另一个代理,它们处理这些URL,但我不能使用它们,因为它们要么不能检查SSL流量,要么只是提供原始数据输出。然而,我发现mitmproxy应该是最好的解决方案,如果它能与webdriver一起正常工作的话。
此外,我已经用mitmproxy在shell->同一问题上开始测试了这一点
与webdriver一起使用Java->同一问题
此外,这种情况有时会发生在其他URL上,如fc2.com 下面是对我的配置的一点概述:
-Linux Mint Petra x64
-python 2.7
-Mitmproxy 0.10.1
-硒2.41.0
-火狐28 有人知道如何解决这个问题吗?
为什么mitmproxy会发生这种情况?
此外,我对使用selenium检查http和https流量的替代方案感兴趣 使用FlowMaster更新:
from libmproxy import flow
from libmproxy import proxy
import os
from multiprocessing import Process
from selenium import webdriver
from selenium.webdriver.common.proxy import *
import time
# traffic inception
class HttpMitmReporter(flow.FlowMaster):
def run(self):
try:
flow.FlowMaster.run(self)
except KeyboardInterrupt:
self.shutdown()
def handle_request(self, r):
f = flow.FlowMaster.handle_request(self, r)
if f:
r.reply()
return f
def handle_response(self, r):
f = flow.FlowMaster.handle_response(self, r)
if f:
r.reply()
return f
# setup browser with proxy settings etc.
def setup_browser():
global driver
proper_setup = False
proxy_con = "127.0.0.1:8088"
my_proxy = Proxy({'proxyType': ProxyType.MANUAL,
'httpProxy': proxy_con,
'ftpProxy': proxy_con,
'sslProxy': proxy_con,
'noProxy': ''})
while not proper_setup:
try:
driver = webdriver.Firefox(proxy=my_proxy)
driver.set_page_load_timeout(20)
driver.set_script_timeout(20)
proper_setup = True
except Exception as exception:
print exception
try:
driver.quit()
except Exception:
time.sleep(3)
driver = None
# start proxy
config = proxy.ProxyConfig(
cacert=os.path.expanduser("~/.mitmproxy/mitmproxy-ca.pem")
)
state = flow.State()
server = proxy.ProxyServer(config, 8088)
reporter = HttpMitmReporter(server, state)
mitm_proxy = Process(target=reporter.run)
mitm_proxy.start()
setup_browser()
# urls to test
urls = [
'http://support.microsoft.com/common/survey.aspx?showpage=1&scid=sw%3Bde%3B3663&altStyle=narrow&renderOption=OverrideDefault&fr=1',
'http://outlook.com/',
]
# visit sites
for url in urls:
print url
try:
driver.get(url)
elements = driver.find_elements_by_tag_name('a')
except Exception as ex:
print ex
driver.quit()
setup_browser()
# terminate browser and proxy
driver.quit()
mitm_proxy.terminate()
看起来您错误地实例化了mitmproxy。看一看,您肯定需要FlowMaster实例,而不是controller.Master。我按照您的建议使用FlowMaster更改了代码(请参阅更新),但同样的问题正在发生。顺便说一句,Firefox已更改为29版。如果我使用外部浏览器,https不起作用(安装了CA),同样的问题也会发生。但是,如果我使用控制器istead,则https流正在工作,并且外部FF中没有问题。然而,即使没有任何代理,selenium atm也有太多问题。因此,我已经编写了自己的casperjs包装器,它工作得更好。