Requests: 'NoneType' object has no attribute 'readline'
Tried a quick check of open issues, but wasn't sure what to search for on this.
Here's the code to reproduce:
import shutil
import requests
import os
session = requests.Session()
# it has something to do with content on the 302 from this url
u = 'http://www.amazon.com/gp/redirect.html/ref=gw_m_b_ir?_encoding=UTF8&location=http%3A%2F%2Fphx.corporate-ir.net%2Fphoenix.zhtml%3Fc%3D97664%26p%3Dirol-irhome&source=standards&token=F9CAD8A11D4336B5E0B3C3B089FA066D0A467C1C'
r0 = session.get(u, stream=True, allow_redirects=False)
redirects = session.resolve_redirects(r0, r0.request, stream=True)
for redirect in redirects:
with open(os.devnull, 'wb') as devnull:
shutil.copyfileobj(redirect.raw, devnull)
Gives me the traceback:
(tmp-91018d725433cf2c) gsbrown$ python bug.py
Traceback (most recent call last):
File "bug.py", line 11, in <module>
for redirect in redirects:
File "/Users/gsbrown/.virtualenvs/tmp-91018d725433cf2c/lib/python2.7/site-packages/requests/sessions.py", line 106, in resolve_redirects
resp.content # Consume socket so it can be released
File "/Users/gsbrown/.virtualenvs/tmp-91018d725433cf2c/lib/python2.7/site-packages/requests/models.py", line 781, in content
self._content = bytes().join(self.iter_content(CONTENT_CHUNK_SIZE)) or bytes()
File "/Users/gsbrown/.virtualenvs/tmp-91018d725433cf2c/lib/python2.7/site-packages/requests/models.py", line 703, in generate
for chunk in self.raw.stream(chunk_size, decode_content=True):
File "/Users/gsbrown/.virtualenvs/tmp-91018d725433cf2c/lib/python2.7/site-packages/requests/packages/urllib3/response.py", line 428, in stream
for line in self.read_chunked(amt, decode_content=decode_content):
File "/Users/gsbrown/.virtualenvs/tmp-91018d725433cf2c/lib/python2.7/site-packages/requests/packages/urllib3/response.py", line 590, in read_chunked
self._update_chunk_length()
File "/Users/gsbrown/.virtualenvs/tmp-91018d725433cf2c/lib/python2.7/site-packages/requests/packages/urllib3/response.py", line 532, in _update_chunk_length
line = self._fp.fp.readline()
AttributeError: 'NoneType' object has no attribute 'readline'
The redirect url is significant. It doesn't happen with the httpbin ones.
gilessbrown
All 31 comments
FYI:
$ pip list
pip (1.5.4)
requests (2.12.4)
setuptools (2.2)
wsgiref (0.1.2)
Just a quick note... I can't really help with your issue directly, as it goes beyond my knowledge.
Just thought I'd mention that pip is up to version 9.0.1... pip 1.5.4 is nearly 3 years old. I'd strongly suggest look at why you're using such an old version of pip
StewPoll
on 11 Jan 2017
Ha. I just like old stuff. :) Can you reproduce the bug?
gilessbrown
on 11 Jan 2017
I can when using python 2.7.12
I can't replicate with python 3.5 though.
StewPoll
on 11 Jan 2017
Are you going to tell me to upgrade to Python 3.5? ;)
gilessbrown
on 11 Jan 2017
Hey @gilessbrown, thanks for opening this issue! This behaviour was exposed by a change made in (327512f) which removed exception handling for this case. It currently only exists in the 2.12.x releases, so using Requests 2.11.1 should work for you.
While you're seeing this behaviour in Requests, it actually seems to be how we're handling chunked responses without a body in urllib3. shazow/urllib3#990 introduced an attempt to catch this problem but the check is just slightly off. While it verifies the existence to fp, it doesn't check that fp isn't None.
I think the simple fix here is to change the check in urllib3 to return getattr(self._fp, 'fp', None) is not None which should give us what we actually want. It verifies that fp both exists, and is not the default None value.
If @Lukasa or @sigmavirus24 are in agreement, we can address this over in urllib3.
nateprewitt
on 11 Jan 2017
Also, I'm able to produce this in both Python 2.7 and 3.5.
nateprewitt
on 11 Jan 2017
Aha - My 3.5 runtime was using requests 2.11.x, which is why it worked
StewPoll
on 11 Jan 2017
Thanks for the tip on Request 2.11.1 and for investigating the issue. I think I’ll follow your suggestion.
gilessbrown
on 11 Jan 2017
While you're seeing this behaviour in Requests, it actually seems to be how we're handling chunked responses without a body in urllib3.
Hang on. Chunked responses without a body are not a thing. The only time they're allowed is in response to a HEAD request, where the body must not be sent. Otherwise, if Transfer-Encoding: chunked is set then there must be a chunk body.
I'd say the server is at fault here.
Lukasa
on 11 Jan 2017
I'd say the server is at fault here.
I can't reproduce this with 2.12.4 on Python 2.7.12 ...
What I am seeing, however, is that the URL provided redirects to
https://www.amazon.com/gp/redirect.html/ref=gw_m_b_ir?_encoding=UTF8&location=http%3A%2F%2Fphx.corporate-ir.net%2Fphoenix.zhtml%3Fc%3D97664%26p%3Dirol-irhome&source=standards&token=F9CAD8A11D4336B5E0B3C3B089FA066D0A467C1C
Which results in another 302 with Transfer-Encoding: chunked. If I access the content on that response, I get '\n'.
>>> r.headers
{'x-amz-id-1': 'V7C8AQ04FPXYP69MGHPS', 'Content-Encoding': 'gzip', 'Transfer-Encoding': 'chunked', 'Strict-Transport-Security': 'max-age=47474747; includeSubDomains; preload', 'Vary': 'Accept-Encoding,User-Agent', 'Server': 'Server', 'Connection': 'keep-alive', 'Location': 'http://phx.corporate-ir.net/phoenix.zhtml?c=97664&p=irol-irhome', 'Date': 'Wed, 11 Jan 2017 12:48:59 GMT', 'p3p': 'policyref="https://www.amazon.com/w3c/p3p.xml",CP="CAO DSP LAW CUR ADM IVAo IVDo CONo OTPo OUR DELi PUBi OTRi BUS PHY ONL UNI PUR FIN COM NAV INT DEM CNT STA HEA PRE LOC GOV OTC "', 'Content-Type': 'text/html; charset=ISO-8859-1', 'x-frame-options': 'SAMEORIGIN'}
>>> r.content
'\n'
After that, I get a 200. So in a virtualenv with 2.12.4 on Python 2.7.12 (and on Fedora 25) I get no issues. I'm curious what's different in what y'all are using that you're seeing this.
sigmavirus24
on 11 Jan 2017
I'd say the server is at fault here.
I completely agree the server isn't compliant here, but we also just said we'd like to be tolerant of things like this in #3794, which is even more out of spec. The server is definitely returning garbage though, so maybe we choose not to address this.
@sigmavirus24 I'm currently able to reproduce this on Python 2.7.12 and 3.5.2 on Mac OS 10.12.2, Ubuntu 12.04, and Travis.
The repro won't fail if you start it at the second hop (the https url), so it seems to require this specific set of responses. It's definitely related to the transfer-encoding because chaining a similar set of responses ((http)302->(https)302->(separate server)200) from httpbin won't trigger the failure.
This "bug" was introduced in Requests 2.7.0 (urllib3 1.10.4) but masked by two separate try/except AssertionError blocks. The first was removed in 2.8.0 in c6c8d64 but this didn't expose the issue because the content exception block was still catching it. 327512f removed the second safeguard which is why 2.12.x is now showing this. I backported 327512f and was able to confirm this only started happening after the chunk transfer work in urllib3 1.10.4. Something causes the underlying socket closed before we can read it for the redirect, but I wasn't able to immediately find what.
At this point this is probably too much digging for a pretty uncommon edge case, but I'll let you two decide.
nateprewitt
on 11 Jan 2017
I understand that streaming from the intermediate redirects is a little exotic. For the record I want to note that the server behaviour is not especially exotic. With tweaked sample to pass the url ...
(requestsbug$ python bug.py http://localad.homedepot.com/
Traceback (most recent call last):
File "bug.py", line 8, in <module>
bug.main()
File "/Users/gsbrown/bug.py", line 28, in main
stream_redirects(url)
File "/Users/gsbrown/bug.py", line 21, in stream_redirects
for redirect in redirects:
File "/Users/gsbrown/.virtualenvs/requestsbug/lib/python2.7/site-packages/requests/sessions.py", line 106, in resolve_redirects
resp.content # Consume socket so it can be released
File "/Users/gsbrown/.virtualenvs/requestsbug/lib/python2.7/site-packages/requests/models.py", line 781, in content
self._content = bytes().join(self.iter_content(CONTENT_CHUNK_SIZE)) or bytes()
File "/Users/gsbrown/.virtualenvs/requestsbug/lib/python2.7/site-packages/requests/models.py", line 703, in generate
for chunk in self.raw.stream(chunk_size, decode_content=True):
File "/Users/gsbrown/.virtualenvs/requestsbug/lib/python2.7/site-packages/requests/packages/urllib3/response.py", line 428, in stream
for line in self.read_chunked(amt, decode_content=decode_content):
File "/Users/gsbrown/.virtualenvs/requestsbug/lib/python2.7/site-packages/requests/packages/urllib3/response.py", line 590, in read_chunked
self._update_chunk_length()
File "/Users/gsbrown/.virtualenvs/requestsbug/lib/python2.7/site-packages/requests/packages/urllib3/response.py", line 532, in _update_chunk_length
line = self._fp.fp.readline()
AttributeError: 'NoneType' object has no attribute 'readline'
Extra info about the previous run:
```
(requestsbug)$ pip list --format=columns
Package Version
pip 9.0.1
requests 2.12.4
setuptools 2.2
(requestsbug)$ python --version
Python 2.7.10```
gilessbrown
on 11 Jan 2017
And the downgrade does not generate the traceback:
(requestsbug)$ pip install requests==2.11.1
Collecting requests==2.11.1
Using cached requests-2.11.1-py2.py3-none-any.whl
Installing collected packages: requests
Found existing installation: requests 2.12.4
Uninstalling requests-2.12.4:
Successfully uninstalled requests-2.12.4
Successfully installed requests-2.11.1
(requestsbug)$ python bug.py http://localad.homedepot.com/
(requestsbug)$
There's nothing unexpected about streaming the intermediate response. We're fine with you doing that. What's wrong is the form of that intermediate response. Allow me to reproduce it for you here (using the original example):
HTTP/1.1 302 MovedTemporarily
Server: Server
Date: Wed, 11 Jan 2017 17:06:17 GMT
Content-Type: text/html; charset=ISO-8859-1
Transfer-Encoding: chunked
Connection: keep-alive
Strict-Transport-Security: max-age=47474747; includeSubDomains; preload
x-amz-id-1: PF9SK6WE9FGKS0FV29ED
p3p: policyref="https://www.amazon.com/w3c/p3p.xml",CP="CAO DSP LAW CUR ADM IVAo IVDo CONo OTPo OUR DELi PUBi OTRi BUS PHY ONL UNI PUR FIN COM NAV INT DEM CNT STA HEA PRE LOC GOV OTC "
x-frame-options: SAMEORIGIN
Location: http://phx.corporate-ir.net/phoenix.zhtml?c=97664&p=irol-irhome
Vary: Accept-Encoding,User-Agent
Content-Encoding: gzip
\x1f\x8b\x08\x00\x00\x00\x00\x00\x00\x03\xe2\x02\x00\x00\x00\xff\xff\x03\x00\x93\x06\xd72\x01\x00\x00\x00
(I have had to hex encode the body to print it out, because it's not valid ASCII)
Note that this response says Transfer-Encoding: chunked, and then provides a non-chunk-encoded body. That body does not obey chunked transfer encoding. This is an error condition. The traceback is not wrong. There is no way to safely read this body. That means your attempt to redirect that body into a file must fail.
I cannot stress this enough: there is no way Requests can "tolerate" this if you ask us for the body because we cannot read the body. It is an invalid HTTP/1.1 response, and we cannot tell when the body is supposed to end. The problem here is not the traceback. If anything, the fact that it "works" in the older case is the actual bug, because we are pretending to the user that everything is fine when in fact it is very much not fine.
If you are going to handle redirects yourself, then you must be prepared to do what Requests does when it receives a bad body on a redirect response, which is swallow them. It is not our job to swallow them for you when you indicated you want the actual response. We are giving you the response: it is the server's fault that the response is invalid. Please take it up with them.
Lukasa
on 11 Jan 2017
I should note that I would be prepared to accept a PR on urllib3 to provide a better exception in this case, because really we should be emitting a urllib3-specific exception here.
Lukasa
on 11 Jan 2017
Hrm. Further investigation suggests that the chunked transfer encoding here is actually fine. I managed to pull out some specific chunked encoding from curl that works just fine, as does the httplib code. So, new question: why is the urllib3 code getting this wrong?
Lukasa
on 11 Jan 2017
Ack, I suspect I know why. We're consuming the content twice. If you eat the data via response.raw, resolve_redirects goes to consume it again and gets all wibbly and confused because it doesn't know that you already did. This can be seen by modifying the originally posted code to (note the additional line at the bottom):
import requests
import os
session = requests.Session()
# it has something to do with content on the 302 from this url
u = 'http://www.amazon.com/gp/redirect.html/ref=gw_m_b_ir?_encoding=UTF8&location=http%3A%2F%2Fphx.corporate-ir.net%2Fphoenix.zhtml%3Fc%3D97664%26p%3Dirol-irhome&source=standards&token=F9CAD8A11D4336B5E0B3C3B089FA066D0A467C1C'
r0 = session.get(u, stream=True, allow_redirects=False)
redirects = session.resolve_redirects(r0, r0.request, stream=True)
for redirect in redirects:
with open(os.devnull, 'wb') as devnull:
shutil.copyfileobj(redirect.raw, devnull)
redirect._content = b''
So, iter_content flags this by setting response._content_consumed = True when it has finished iterating.
I suspect the real culprit here is urllib3's custom chunking code, which does not check whether it believes the response is complete or not (in this case, it is, and so it should be short-circuit returning). I think that's a bug on urllib3.
Lukasa
on 11 Jan 2017
See shazow/urllib3#1088. In the meantime, I recommend a workaround when using response.raw when iterating redirects that you set _content_consumed to True if you have, in fact, consumed the content. Closing in favour of the urllib3 bug.
Lukasa
on 11 Jan 2017
Hi @Lukasa, your analysis of it being caused by the double consumption of raw (first by the copyfileobj, and then by the response.content in resolve_redirects) matches what I saw in the debugger so I'm pretty confident your work-around of setting ._content will work for me. Thanks for your deep dive into this.
gilessbrown
on 11 Jan 2017
Could someone change this subject to "'NoneType' object has no attribute 'readline'" ?
I ran into the same issue, and the SEO on this pushed it down a bit on search results.
From my tests so far, I believe the ._content_consumed flag needs to be set to true if anything is read off the socket during a redirect resolve (ie a header not just "content"). My workaround is to place a "#TODO" pointing to this issue in my code, and consuming all the content + setting the flag at that point.
jvanasco
on 2 May 2017
@jvanasco Reading the headers should not require _content_consumed to be True. If it did, the _content_consumed flag would be entirely redundant, because we never create a Response object without having already read the headers.
Lukasa
on 2 May 2017
hm. I'm not sure what happened. I tracked my error down to a custom redirect resolver. while the block does consume content, I was getting an error on cases that exited out before then. the only thing accessed on those was the headers/status.
i added these 2 lines to the top of the block, and all the edges passed:
...
touched = resp.content
resp._content_consumed = True
...
i'm far too busy to look deeper into this, but I'll make a note and try to investigate on my free time.
oh! thanks for the input and making the title more SEO friendly!
jvanasco
on 4 May 2017
Howdy, I'm not clear – was this resolved? I'm on requests==2.18.4 with python 3.5.1 and still seeing this issue.
Specifically,
File "/Users/foo/Code/dev/myapp/brain/network.py", line 56, in expand_url
return get(url)
File "/Users/foo/Code/dev/myapp/brain/network.py", line 27, in get
return requests.get(url, headers=get_headers(), timeout=20, allow_redirects=True)
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/api.py", line 72, in get
return request('get', url, params=params, **kwargs)
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/api.py", line 58, in request
return session.request(method=method, url=url, **kwargs)
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/sessions.py", line 508, in request
resp = self.send(prep, **send_kwargs)
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/sessions.py", line 640, in send
history = [resp for resp in gen] if allow_redirects else []
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/sessions.py", line 640, in <listcomp>
history = [resp for resp in gen] if allow_redirects else []
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/sessions.py", line 218, in resolve_redirects
**adapter_kwargs
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/sessions.py", line 658, in send
r.content
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/models.py", line 823, in content
self._content = bytes().join(self.iter_content(CONTENT_CHUNK_SIZE)) or bytes()
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/requests/models.py", line 745, in generate
for chunk in self.raw.stream(chunk_size, decode_content=True):
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/urllib3/response.py", line 432, in stream
for line in self.read_chunked(amt, decode_content=decode_content):
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/urllib3/response.py", line 598, in read_chunked
self._update_chunk_length()
File "/Users/foo/Code/dev/venv/lib/python3.5/site-packages/urllib3/response.py", line 540, in _update_chunk_length
line = self._fp.fp.readline()
AttributeError: 'NoneType' object has no attribute 'readline'
Thank you for any insight.
zefoo
on 25 Feb 2018
Hey @zefoo, this is an issue in urllib3 which is a layer beneath Requests. The issue is being tracked in shazow/urllib3#1088, but I don't know if any progress has been made. I would recommend upgrading to the latest version of urllib3 and if the issue persists, follow up there. Thanks!
nateprewitt
on 25 Feb 2018
Using requests with any common user-agent I hit the same bug at: http://forbes.com:
import requests
session = requests.Session()
resp = session.get('http://forbes.com', headers={'User-Agent': 'Mozilla/5.0'})
The stacktrace:
AttributeError Traceback (most recent call last)
<ipython-input-8-a73f5204685b> in <module>()
----> 1 resp = session.get('http://forbes.com', headers={'User-Agent': 'Mozilla/5.0'})
~/.virtualenvs/test/lib/python3.6/site-packages/requests/sessions.py in get(self, url, **kwargs)
519
520 kwargs.setdefault('allow_redirects', True)
--> 521 return self.request('GET', url, **kwargs)
522
523 def options(self, url, **kwargs):
~/.virtualenvs/test/lib/python3.6/site-packages/requests/sessions.py in request(self, method, url, params, data, headers, cookies, files, auth, timeout, allow_redirects, proxies, hooks, stream
, verify, cert, json)
506 }
507 send_kwargs.update(settings)
--> 508 resp = self.send(prep, **send_kwargs)
509
510 return resp
~/.virtualenvs/test/lib/python3.6/site-packages/requests/sessions.py in send(self, request, **kwargs)
638
639 # Resolve redirects if allowed.
--> 640 history = [resp for resp in gen] if allow_redirects else []
641
642 # Shuffle things around if there's history.
~/.virtualenvs/test/lib/python3.6/site-packages/requests/sessions.py in <listcomp>(.0)
638
639 # Resolve redirects if allowed.
--> 640 history = [resp for resp in gen] if allow_redirects else []
641
642 # Shuffle things around if there's history.
~/.virtualenvs/test/lib/python3.6/site-packages/requests/sessions.py in resolve_redirects(self, resp, req, stream, timeout, verify, cert, proxies, yield_requests, **adapter_kwargs)
216 proxies=proxies,
217 allow_redirects=False,
--> 218 **adapter_kwargs
219 )
220
~/.virtualenvs/test/lib/python3.6/site-packages/requests/sessions.py in send(self, request, **kwargs)
656
657 if not stream:
--> 658 r.content
659
660 return r
~/.virtualenvs/test/lib/python3.6/site-packages/requests/models.py in content(self)
821 self._content = None
822 else:
--> 823 self._content = bytes().join(self.iter_content(CONTENT_CHUNK_SIZE)) or bytes()
824
825 self._content_consumed = True
~/.virtualenvs/test/lib/python3.6/site-packages/requests/models.py in generate()
743 if hasattr(self.raw, 'stream'):
744 try:
--> 745 for chunk in self.raw.stream(chunk_size, decode_content=True):
746 yield chunk
747 except ProtocolError as e:
~/.virtualenvs/test/lib/python3.6/site-packages/urllib3/response.py in stream(self, amt, decode_content)
430 """
431 if self.chunked and self.supports_chunked_reads():
--> 432 for line in self.read_chunked(amt, decode_content=decode_content):
433 yield line
434 else:
~/.virtualenvs/test/lib/python3.6/site-packages/urllib3/response.py in read_chunked(self, amt, decode_content)
596 with self._error_catcher():
597 while True:
--> 598 self._update_chunk_length()
599 if self.chunk_left == 0:
600 break
~/.virtualenvs/test/lib/python3.6/site-packages/urllib3/response.py in _update_chunk_length(self)
538 if self.chunk_left is not None:
539 return
--> 540 line = self._fp.fp.readline()
541 line = line.split(b';', 1)[0]
542 try:
AttributeError: 'NoneType' object has no attribute 'readline'
This is the wget output:
--2018-05-31 20:39:04-- http://forbes.com/
Resolving forbes.com... 151.101.2.49
Connecting to forbes.com|151.101.2.49|:80... connected.
HTTP request sent, awaiting response...
HTTP/1.1 301 Moved Permanently
Server: Varnish
Retry-After: 0
Content-Length: 0
Location: https://forbes.com/
Accept-Ranges: bytes
Date: Fri, 01 Jun 2018 03:39:04 GMT
Via: 1.1 varnish
Connection: close
X-Served-By: cache-pao17432-PAO
X-Cache: HIT
X-Cache-Hits: 0
X-Timer: S1527824344.292472,VS0,VE0
Access-Control-Allow-Credentials: true
Location: https://forbes.com/ [following]
--2018-05-31 20:39:04-- https://forbes.com/
Connecting to forbes.com|151.101.2.49|:443... connected.
HTTP request sent, awaiting response...
HTTP/1.1 302 Found
Server: Varnish
Accept-Ranges: bytes
location: https://www.forbes.com/forbes/welcome/?toURL=https://forbes.com/&refURL=&referrer=
X-Frame-Options: SAMEORIGIN
X-Cicero-Cache: MISS
Content-Security-Policy: upgrade-insecure-requests
Strict-Transport-Security: max-age=0; includeSubDomains
Accept-Ranges: bytes
Date: Fri, 01 Jun 2018 03:39:04 GMT
Via: 1.1 varnish
X-Served-By: cache-pao17451-PAO
X-Cache: MISS
X-Cache-Hits: 0
X-Timer: S1527824344.332493,VS0,VE269
Vary: X-ABtesting
Access-Control-Allow-Credentials: true
Connection: close
Location: https://www.forbes.com/forbes/welcome/?toURL=https://forbes.com/&refURL=&referrer= [following]
--2018-05-31 20:39:04-- https://www.forbes.com/forbes/welcome/?toURL=https://forbes.com/&refURL=&referrer=
Resolving www.forbes.com... 151.101.190.49
Connecting to www.forbes.com|151.101.190.49|:443... connected.
HTTP request sent, awaiting response...
HTTP/1.1 200 OK
Content-Type: text/html;charset=utf-8
Content-Language: en-US
Server:
Backend: templates
X-YourTtl: 300.000
X-Frame-Options: SAMEORIGIN
X-Cicero-Cache: HIT 4
Content-Security-Policy: upgrade-insecure-requests
Strict-Transport-Security: max-age=10886400; includeSubDomains; preload
Accept-Ranges: bytes
Date: Fri, 01 Jun 2018 03:39:04 GMT
Via: 1.1 varnish
X-Served-By: cache-pao17438-PAO
X-Cache: MISS
X-Cache-Hits: 0
X-Timer: S1527824345.632429,VS0,VE68
Vary: Accept-Encoding, X-ABtesting
Access-Control-Allow-Credentials: true
Set-Cookie: client_id=442dfb7dc29c1468da4ee0432b04ab7cb1b; Path=/; Domain=.forbes.com; Expires=Sun, 31 May 2020 03:39:04 GMT
Connection: close
Length: unspecified [text/html]
This is for:
requests==2.18.4
urllib3==1.22
However, I've even tried with urlib3-dev which has the fix for urllib3/urllib3#1088 merged and the issue still persists.
mjuarezm
on 1 Jun 2018
Hi @mjuarezm, if you can verify that you're using the dev version of urllib3, both with the warning we emit "RequestsDependencyWarning: urllib3 (dev) or chardet (3.0.4) doesn't match a supported version!" and with the contents of requests.packages.urllib3.__version__ being "dev", you'll want to open a new ticket since this is 18 months old. I'm unable to reproduce the issue with what you've provided with the current master branch of urllib3, so we'll need more information about your system.
nateprewitt
on 1 Jun 2018
@nateprewitt I'm sorry, I was installing requests from master branch using pip and didn't realize that it was automatically uninstalling urllib3-dev and installing urllib3. I made sure to use urllib3-dev and I get a status code 200 as expected. Thanks!
mjuarezm
on 1 Jun 2018
This is a python2 python3 version incompatible error. out.readline is only py2 PEP usage.
BichengWang
on 25 Feb 2020
Related issues
xsren
·
3Comments
8key
·
3Comments
remram44
·
4Comments
avinassh
·
4Comments
brainwane
·
3Comments