Why is there such a difference between the URL encoding behavior of requests and urllib,urllib2,urllib3? __ Code

#!/usr/bin/env python
#coding: Utf-8
Import Requests,urllib,urllib2,urllib3,urlparse

url = "Http://xxx.com/index.php?Q=u=%OS%26%20"
Print "Original:", url
#print requests.get (URL). Content
Print "-------------------"
url_parts = urlparse.urlparse (URL)
Print "splited:", Url_parts.query
Print "-------------------"
params = Dict (URLPARSE.PARSE_QSL (url_parts.query,true))
Print "parsed:", "q=" +params[' Q ']
Print "-------------------"
url_dealed = Urlparse.urlunsplit (url_parts.scheme,url_parts.netloc\
, Url_parts.path,urllib.urlencode (params) \
, url_parts.fragment))
Print "unsplited (dealed_final):", url_dealed
Print "-------------------"

#requests. Get (url= "http://xxx.com/index.php?Q=u=%OS%26%20"). Content
#requests. Get (url= "http://xxx.com/index.php", params). Content
Print "Requests:", "http://xxx.com/index.php?Q=u=%25OS%2526%2520"

#urllib. Urlopen (url= "http://xxx.com/index.php?Q=u=%OS%26%20"). Read ()
Print "Urllib:", "http://xxx.com/index.php?Q=u=%OS%26%20"

#urllib2. Urlopen (url= "http://xxx.com/index.php?Q=u=%OS%26%20"). Read ()
Print "URLLIB2:", "http://xxx.com/index.php?Q=u=%OS%26%20"

#urllib3. Proxymanager (' http://localhost:8888/'). Request ("Get", "http://xxx.com/index.php?Q=u%3D%25OS%26+"). Data
Print "URLLIB3:", "http://xxx.com/index.php?Q=u=%OS%26%20"

#url_dealed = "http://xxx.com/index.php?Q=u%3D%25OS%26+"
#requests. Get (url_dealed). Content
#urllib3. Proxymanager (' http://localhost:8888/'). Request ("Get", url_dealed). Data
#urllib2. Urlopen (url_dealed). Read ()
#urllib. Urlopen (url_dealed). Read ()
Print "All requests dealed_final:", "http://xxx.com/index.php?Q=u%3D%25OS%26+"
Borrow the number to send a learning exchange posts, to find the way great God Popular Science four python library default encoding behavior differences:
Today, test injection, the browser URL posted to the requests parameters in the contract has not been expected to respond to the decision to grab a bag diff, did not think really requests did the hands and feet.
1. Requests request http://xxx.com/index.php?Q=u=%25OS%2526%2520 will encode the% of Q value as%25.
2. Requests request after the split-code-assembly processing http://xxx.com/index.php?Q=u%3D%25OS%26+ remain intact.
3. URLLIB,URLLIB2,URLLIB3 the URL request in 1,2 is intact.
The good requests is the encapsulated URLLIB3, why change this default behavior. Collection Thanks (0) share to: 0 10 reply Answer | 2016-03-03 20:55

2# Invisible Man really busy (focus on security research and development and vulnerabilities) | 2016-03-03 20:59

This is a pit requests code is very powerful in general my words with the Urllib this method:
Urllib.quote (string[, safe])
Encodes a string. Parameter safe specifies characters that do not need to be encoded. As for why can only see the source code slightly ... 3# Null_z | 2016-03-03 21:18

If you're stitching up your own URL and then use requests to not encode
If the dictionary is passed in as a parameter, the requests is encoded once
Requests is not a standard library. 4# Larry (How to sign the end MT i) | 2016-03-03 21:28

With the second floor, today also want to UrlEncode, feel can also use Urllib.quote_plus (string[, safe) can be encoded in the URL HTML form of the query string, and then stitching their own 5# fate0 (I will wait for you in the future) | 2016-03-03 22:58

https://github.com/kennethreitz/requests/blob/master/requests%2Futils.py#L443

Because the OS is not a legitimate 16-in-6#, a small slag | 2016-03-04 08:37

I have also been in the tangle of writing exp sometimes use a similar%u0027 to circumvent some restrictions, but with this coded URL directly error 7# zxx | 2016-03-04 11:16

From requests import Request, session
s = Session ()
Url= "Http://xxx.com/index.php?Q=u=%OS%26%20"
req = Request (' Get ', url)
prepped = S.prepare_request (req)
Prepped.url = Prepped.url.replace ('%25 ', '% ')
RESP = s.send (prepped) 8# Modnar | 2016-03-04 11:46

@fate0 # We couldn ' t unquote the given URI, so let ' s try quoting it, but
# There May is unquoted '% ' s in the URI. We need to make sure they ' re
# properly quoted so they does not cause issues elsewhere
In order to not decode the 16, so simply rough coding%, but also affect the other can decode the 16, a bit incomprehensible ... Traditional browsers have not done so, and requests does not offer a switch. 9# Modnar | 2016-03-04 11:49

@ The invisible man is really busy. This can be addressed by requests to the Urllib.urlencode (Dict (URLPARSE.PARSE_QSL (urlparse.urlparse (URL). query,true)), Decoding is also encoded after the%26 and%20 reserved ~ 10# Invisible Man really busy (focus on security research and development and vulnerabilities) | 2016-03-04 17:20

@ModNar Indeed    was careful to read the documentation for   , but there was no such flag