Added
- base_url parameter to
niquests.Session
orniquests.AsyncSession
. automatically prefix every request emitted with it. (#179)
Fixed
- warning about coroutine method 'aclose' of 'AsyncSession.resolve_redirects' was never awaited.
Fixed
- Resolving lazy responses when emitted through a SOCKS, HTTP or HTTPS proxy. (#170)
Fixed
- Ensure
stream
, andverify
both defaults to yourSession
parameters.
Fixed
- Exception leak from urllib3-future when gathering / resolving lazy responses.
Added
- Automatic Advanced Keep-Alive for HTTP/2 and HTTP/3 over QUIC by sending PING frames.
New Session, and Adapter parameters are now available:
keepalive_delay
, andkeepalive_idle_window
. This greatly improves your daily experience working with HTTP/2+ remote peers.
Fixed
- Unshielded picotls assertion error in Python < 3.10 when trying to fetch the peer intermediate certificate.
Fixed
- Exception leak from urllib3-future when using WebSocket.
- Enforcing HTTP/3 in an AsyncSession. (#152)
- Adapter kwargs fallback to support old Requests extensions.
- Type hint for
Response.extension
linked to the generic interface instead of the inherited ones. - Accessing WS over HTTP/2+ using the synchronous session object.
Misc
- Documentation improvement for in-memory certificates and WebSocket use cases.
Changed
- urllib3-future lower bound version is raised to 2.10.904 to ensure exception are properly translated into urllib3-future ones for WS.
Added
- Support for WebSocket over HTTP/1, HTTP/2 and HTTP/3. It brings a unified API that makes you leverage our powerful features like Happy Eyeballs, SOCKS/HTTP/HTTPS proxies, thread/task safety etc...
- Hook for catching early responses like "103 Early Hints".
Fixed
- Informational responses are fully supported over HTTP/1, HTTP/2 and HTTP/3.
Changed
- urllib3-future lower bound version is raised to 2.10.900.
Added
- Support for HTTP Trailers.
- Help script now yield warnings if update are available for each sub dependencies.
Fixed
- Setting a list of Resolver.
Changed
- urllib3-future lower bound version is raised to 2.9.900 (for http trailer support).
- relax strict kwargs passing in Session adapters (required for some plugins).
Fixed
- DummyLock injected into RequestsCookieJar is not fully compatible, thus breaking AsyncSession on certain scenario. (#136)
Added
- Official support for Python 3.13 This has been tested outside GitHub CI due to httpbin unready state for 3.13[...]
- Support for asynchronous auth callables.
- Support for asynchronous bodies through
AsyncIterable
that yield either bytes or str. - Support for purposely excluding a domain/port from connecting to QUIC/HTTP3 via the
quic_cache_layer
property ofSession
. In order to excludecloudflare.com
from HTTP3 auto-upgrade:from niquests import Session s = Session() s.quic_cache_layer.exclude_domain("cloudflare.com")
Fixed
- auth argument not accepting a function according to static type checkers. (#133)
- RequestsCookieJar having a lock in
AsyncSession
. Its effect has been nullified to improve performances.
Changed
- urllib3-future lower bound version is raised to 2.8.902
Added
- TransferProgress tracking in Response when downloading using
stream=True
based on the Content-Length. (#127) There's no easy way to track the "real" amount of bytes consumed using "iter_content" when the remote is sending a compressed body. This change makes it possible to track the amount of bytes consumed. TheResponse
object now contain a property nameddownload_progress
that is eitherNone
or aTransferProgress
object. - HTTP/2 with prior knowledge over TLS or via an unencrypted connection.
disable_http1
toggle is now available through yourSession
constructor. In consequence, you may leverage all HTTP/2 capabilities like multiplexing using a plain (e.g. non-TLS) socket. You may enable/disable any protocols per Session object (but not all of them at once!). In non-TLS connections, you have to keep one of HTTP/1.1 or HTTP/2 enabled. Otherwise, one of HTTP/1.1, HTTP/2 or HTTP/3. ARuntimeError
may be thrown if no protocol can be used in a given context.
Changed
- Relax main API constraint in get, head, options and delete methods / functions by accepting kwargs.
- urllib3-future lower bound version is raised to 2.8.900
Fixed
- CaseInsensitiveDict repr (to string) causing an unexpected error when upstream have multiple values for a single header. (#129)
Misc
- Minor docs typos (#128)
- Reformated error messages in our OCSP module for better readability.
- Added real test cases for our OCSP module to ensure its reliability.
Fixed
- ReasonFlag not properly translated to readable text when peer or intermediate certificate is revoked.
Fixed
- Support
localhost
as a valid domain for cookies. The standard library does not allow this special domain. Researches showed that a valid domain should have at least two dots (e.g. abc.com. and xyz.tld. but not com.). Public suffixes cannot be used as a cookie domain for security reasons, but aslocalhost
isn't one we are explicitly allowing it. Reported in httpie/cli#602RequestsCookieJar
set a default policy that circumvent that limitation, if you specified a custom cookie policy then this fix won't be applied.
Changed
- Lazy load the OCSP extension in order to improve the import performance.
Removed
- Class variable
disable_thread
inAsyncSession
that is no longer relevant since the native asyncio implementation. (PR #122)
Changed
- Avoid parsing X509 peer certificate in the certificate revocation check process over and over again.
- Avoid iterating over header items redundantly or needlessly.
Fixed
- Fixed encoding data with None values and other objects. This was a regression introduced in our v3. #119
Changed
- Various minor performance improvements.
Fixed
- "Help" program
python -m niquests.help
that depended on h2 while not required anymore. - Minor performance regression in async while checking OCSP when certificate isn't eligible (e.g. no OCSP url provided).
Changed
- urllib3.future lower bound constraint has been raised to version 2.7.905 to ensure inclusion of jh2 instead of h2.
Fixed
- Handling broken environments with a graceful exception with a detailed error message.
Added
- Support for qh3 version 1.0.0 This qh3 release enable a greater flexibility by dropping cryptography. We had to adapt the OCSP code as we relied on cryptography. HTTP/3 experience is greatly improved.
Changed
- urllib3.future lower bound constraint has been raised to version 2.7.904 to ensure support for the last qh3 release.
Fixed
- Improved compatibility with third party mocking tool that are bound to requests.
- OCSP check did not warn if the HTTP server responded with a non 2xx response in strict mode.
Added
- Support for Happy Eyeballs. This feature is disabled by default, you must pass
happy_eyeballs=True
within your session constructor or http adapter in order to leverage this.
Fixed
- Missed close implementation in AsyncSession causing the underlying poolmanager to remain open.
- Additional OCSP requests (following a redirect) did not use specified custom DNS resolver.
Changed
- urllib3.future lower bound constraint has been raised to version 2.7.900 for the newly added happy eyeballs feature.
Added
- Support to verify the peer certificate fingerprint using
verify=...
by passing a string using the following format:verify="sha256_748c76348778cb4a536e7ec12bc9aa559c12770bd1419c7ffe516006e1dea0ec"
. Doing so disable the certificate usual verification and only checks for its fingerprint match.
Fixed
- Multiplexed request in async did not support awaitable in hooks.
- Setting
verify=...
andcert=...
then change it for the same host did not apply to the underlying (existing) connection pool.
Misc
- Overall performance improvements in both async and sync requests.
- Update pre-commit dependencies (ruff, pyupgrade, and mypy).
- Fixed SessionRedirect in tests that produced an incomplete Response instance that wasn't suitable for tests.
Fixed
- A rare error that occurs on PyPy, especially on Windows, complaining about a missing release call.
Misc
- Allow latest dependencies version for httpbin, Flask and werkzeug in tests.
- Remove wheel from test dependencies.
Fixed
- ImportError in an attempt to retrieve
ConnectionInfo
.
Changed
- General performance improvements.
Changed
- General performance improvements.
Changed
- Replaced the thread pool executor for the true asyncio implementation. The asynchronous part is rewritten.
- urllib3.future lower bound constraint has been raised to version 2.6.900 for the newly added asynchronous interfaces.
Added
- Support for awaitable hooks in AsyncSession.
- Extendable AsyncBaseAdapter.
- Support for orjson if installed in environment to dramatically increase performance in JSON deserialization.
- Extra
speedups
to increase overall performances by including optional dependencies such as zstandard, brotli and orjson.
Fixed
- A rare error when leveraging multiplexing when a proxy is used.
- Bypass proxies when looking to validate certificate revocation status in Python < 3.10 with OCSP.
Fixed
- Unintentional performance regression with multiple concurrent multiplexed connection within a single Session.
Fixed
- Unmatched filter for deprecation warning yielded by Cryptography due to some legacy CA available in Windows having a negative serial number.
- Setting
boundary
inContent-Type
header with no value associated (no equal sign) can cause a rare error (multipart). - Rare racing condition while emitting too many request across a multiplexed connections.
- Spawning too many threads while using
AsyncSession
in specific contexts.
Fixed
- Thread-safety issue when leveraging a single multiplexed connection across multiple threads.
- Apparently consumed content when allow_redirect is set to True when accessing a lazy response that follow redirects.
Changed
- urllib3.future lower bound constraint has been raised to version 2.5.900 in order to leverage the advanced multiplexing scheduler. This upgrade come with a noticeable performance bump.
Added
Session
constructor now accepts bothpool_connections
andpool_maxsize
parameters to scale your pools of connections at will.
Fixed
- Issuing a request with
Session(multiplexed=True)
that weren't eligible (e.g. HTTP/1.1) but was redirected to an eligible server (HTTP/2+) caused a rare error. - An anonymous netrc entry (e.g. username only) could be wrongfully used when it should be discarded. (PR #61)
Added
- Awaitable close method within
AsyncSession
.
Fixed
- Accessing a lazy response (multiplexed enabled) that have multiple redirects did not work appropriately.
Changed
- Response
iter_content
anditer_line
read chunks as they arrive by default. The default chunk size is now-1
.-1
mean to instruct that the chunks can be of variable sizes, depending on how packets arrives. It improves overall performances. - urllib3.future lower bound constraint has been raised to version 2.4.904 in order to accept
-1
as a chunk size.
Fixed
- Connection information kept targeting its original copy, thus always keeping the latest timings inside while expecting the historical ones.
Added
AsyncSession
now returns aAsyncResponse
whenstream
is set to True in order to handle properly streams in an async context.
Fixed
- CaseInsensibleDict did not properly convert HTTPHeaderDict from urllib3 thus only letting the last entry in.
- Redirect chain lead to a non waited coroutine in
AsyncSession
.
Added
- Support for specifying a custom DNS resolver in
Session
. - Support for passing the source address in
Session
. - Support for disabling either IPv4 or IPv6 within a
Session
.
Changed
- PySocks is no longer used for SOCKS proxies. Replaced by python-socks instead.
- urllib3.future minimal version raised to 2.4+ to leverage newly added features.
- Improve compatibility when end-user inadvertently pass a
Timeout
orRetry
instance from the legacyurllib3
instead ofurllib3_future
.
Fixed
- Blocking the event loop when closing the
AsyncSession
usingwith
. - Rare exception on older PyPy interpreters due to Generic having unsupported type variable in
extensions._sync_to_async
module.
Misc
- Project extras aligned with urllib3.future.
- Using nox instead of tox.
- Switch to ruff instead of black and isort.
Fixed
- Overall static typing experience have been improved.
- Code HTTP 425 is now registered as
too_early
in addition to the legacyunordered_collection
.
Removed
- Private module
niquests._internal_utils
has been removed as it no longer serve its purposes.
Added
- Hook
on_upload
that allows you to monitor/track the upload progress. - Model
TransferProgress
that is used inPreparedRequest
as public propertyupload_progress
.
Fixed
- Hooks that does not accept keyword arguments are rejected.
- Applying
max_fetch
toSession.gather(...)
did not prevent the adapter to drain all pending responses. - Closed session having unconsumed multiplexed requests leaked an exception from urllib3.future.
Changed
- Aligned
qh3
version constraint inhttp3
extra with urllib3.future.
Fixed
- Warning filter (ignore) for
DependencyWarning
within urllib3.future wasn't applied in time.
Added
- Maximum of (lazy) response(s) to be resolved when calling
Session.gather(..., max_fetch = ...)
. Specifying a valid int tomax_fetch
will stop right after having resolved the right amount of responses.
Changed
- urllib3.future minimal version raised to 2.3.900 to leverage the fallback top-level package
urllib3_future
.
Fixed
- Error when having accidentally overriden
urllib3.future
by an external dependency. - Undesirable warning yielded by
cryptography
because of a Microsoft root certificate. "Parsed a negative serial number, which is disallowed by RFC 5280." - Take into account newly registered custom CA when using
wassima.register_ca(...)
.
Removed
- Dependency check at runtime for
urllib3
. There's no more check and warnings at runtime for that subject. Ever.
Fixed
- Compatibility with some third-party mock tools.
Changed
- Relax IllegalHeader constraint when value is an integer, or float.
Fixed
- Overall performance improvement.
- Racing condition in
HTTPAdapter
with multiplexed enabled while in threads.
Removed
- Checks for header validity as it is duplicated (upstream does it already).
Warning: Function
check_header_validity
has been removed. It was not public in the first place.
Changed
- Cached proxies from environment to avoid unnecessary scan on each request.
Changed
- Enforced a maximum in-flight request when using multiplexed connections. Default to 124 per connections
so, actually 1240 per Session (default is 10 connections). This can be overriden in our
HTTPAdapter
for advanced users. This limit was changed due to constraint inqh3
, for now we have no way to dynamically set this. We choose the safest lowest common value in h2, and qh3.
Fixed
- Performance issues in
get_environ_proxies()
.
Fixed
- Performance issues in HTTP/2, and HTTP/3, with or without multiplexed connections.
Changed
- Enforced a maximum in-flight request when using multiplexed connections. Default to 200 per connections
so, actually 2000 per Session (default is 10 connections). This can be overriden in our
HTTPAdapter
for advanced users.
Changed
- Changed method
raise_for_status
in classResponse
to return self in order to make the call chainable. Idea taken from upstream psf#6215 - Bump minimal version supported for
urllib3.future
to 2.2.901 for recently introduced added features (bellow).
Added
-
Support for multiplexed connection in HTTP/2 and HTTP/3. Concurrent requests per connection are now a thing, in synchronous code. This feature is the real advantage of using binaries HTTP protocols. It is disabled by default and can be enabled through
Session(multiplexed=True)
, eachResponse
object will be 'lazy' loaded. Accessing anything from returnedResponse
will block the code until target response is retrieved. UseSession.gather()
to efficiently receive responses. You may also give a list of responses that you want to load.Example A) Emitting concurrent requests and loading them via
Session.gather()
from niquests import Session from time import time s = Session(multiplexed=True) before = time() responses = [] responses.append( s.get("https://pie.dev/delay/3") ) responses.append( s.get("https://pie.dev/delay/1") ) s.gather() print(f"waited {time() - before} second(s)") # will print 3s
Example B) Emitting concurrent requests and loading them via direct access
from niquests import Session from time import time s = Session(multiplexed=True) before = time() responses = [] responses.append( s.get("https://pie.dev/delay/3") ) responses.append( s.get("https://pie.dev/delay/1") ) # internally call gather with self (Response) print(responses[0].status_code) # 200! :! Hidden call to s.gather(responses[0]) print(responses[1].status_code) # 200! print(f"waited {time() - before} second(s)") # will print 3s
You have nothing to do, everything from streams to connection pooling are handled automagically!
-
Support for in-memory intermediary/client certificate (mTLS). Thanks for support within
urllib3.future
. Unfortunately this feature may not be available depending on your platform. Passingcert=(a, b, c)
where a or/and b contains directly the certificate is supported. See https://urllib3future.readthedocs.io/en/latest/advanced-usage.html#in-memory-client-mtls-certificate for more information. It is proposed to circumvent recent pyOpenSSL complete removal. -
Detect if a new (stable) version is available when invoking
python -m niquests.help
and propose it for installation. -
Add the possibility to disable a specific protocol (e.g. HTTP/2, and/or HTTP/3) when constructing
Session
. Like so:s = Session(disable_http2=..., disable_http3=...)
both options are set toFalse
, thus letting them enabled. urllib3.future does not permit to disable HTTP/1.1 for now. -
Support passing a single
str
toauth=...
in addition to actually supported types. It will be treated as a Bearer token, by default to theAuthorization
header. It's a shortcut. You may keep your own token prefix in given string (e.g. if not Bearer). -
Added
MultiplexingError
exception for anything related to failure with a multiplexed connection. -
Added async support through
AsyncSession
that utilize an underlying thread pool.from niquests import AsyncSession import asyncio from time import time async def emit() -> None: responses = [] async with AsyncSession(multiplexed=True) as s: responses.append(await s.get("https://pie.dev/get")) responses.append(await s.get("https://pie.dev/head")) await s.gather() print(responses) async def main() -> None: foo = asyncio.create_task(emit()) bar = asyncio.create_task(emit()) await foo await bar if __name__ == "__main__": before = time() asyncio.run(main()) print(time() - before)
Or without
multiplexing
if you want to keep multiple connections open per host per request.from niquests import AsyncSession import asyncio from time import time async def emit() -> None: responses = [] async with AsyncSession() as s: responses.append(await s.get("https://pie.dev/get")) responses.append(await s.get("https://pie.dev/head")) print(responses) async def main() -> None: foo = asyncio.create_task(emit()) bar = asyncio.create_task(emit()) await foo await bar if __name__ == "__main__": before = time() asyncio.run(main()) print(time() - before)
You may disable concurrent threads by setting
AsyncSession.no_thread = True
.
Security
- Certificate revocation verification may not be fired for subsequents requests in a specific condition (redirection).
Fixed
- Static type checker not accepting iterable[str] for data. A fix in urllib3.future allows it since v2.1.902.
- Unattended override of manually provided Authorization if
.netrc
existed with an eligible entry. Taken from closed PR psf#6555 and initially raised in psf#3929
Added
- oheaders property in
Request
, andPreparedRequest
in addition toResponse
.
Fixed
- oheaders from a Response contains
Set-Cookie
entries when it should not. - Static type checker not accepting list[str] in values for argument param.
- Static type checker not accepting Iterable[bytes] for data.
- Function proxy_bypass_registry for Windows may be fooled by insufficient control on our end. Patch taken from idle upstream PR psf#6302
- SSLError message related to the certificate revocation could print
None
instead ofunspecified
for the reason.
Changed
- Allow setting
None
in max_size for SharableLimitedDict to remove limits. - Using
RLock
instead ofLock
in SharableLimitedDict, and InMemoryRevocationStatus classes.
Misc
- Missing assert statements for test test_header_validation.
- Unrelated warnings are now silent in our test suite.
- Unexpected warning now trigger an error in our test suite.
- Removed
tests.compat
. - Removed
test-readme
,flake8
, andpublish
from Makefile.
Added
- Extra-dist install
http3
to force install HTTP/3 support in your environment if not present. - Extra-dist install
ocsp
to force install certificate revocation support in your environment if not present.
Fixed
- Static type checker not accepting list[str] in values for argument data.
Misc
- Changed the documentation theme by furo.
Added
-
IPv6 support in the
NO_PROXY
environment variable or in the proxies (key no_proxy) argument. Patch taken from idle upstream PR psf#5953 -
Preemptively register a website to be HTTP/3 capable prior to the first TLS over TCP handshake. You can do so by doing like:
from niquests import Session s = Session() s.quic_cache_layer.add_domain("cloudflare.com")
-
Passed data will be converted to form-data if headers have a Content-Type header and is set to
multipart/form-data
. Otherwise, by default, it is still urlencoded. If you specified a boundary, it will be used, otherwise, a random one will be generated.
Fixed
- Fixed
Transfer-Encoding
wrongfully added to headers when body is actually of length 0. Due to ambiguous return ofsuper_len
in niquests internals. - Fixed accepting three-valued tuple for Timeout (connect, read, total) in addition of known (connect, read) tuple.
Misc
- Static typing has been improved to provide a better development experience.
Added
-
Certificate revocation verification via the OCSP protocol.
This feature is broadly available and is enabled by default when
verify=True
. We decided to follow what browsers do by default, so Niquests follows by being non-strict. OCSP responses are expected to arrive in less than 200ms, otherwise ignored (e.g. OCSP is dropped). Niquests keeps in-memory the results until the size exceed 2,048 entries, then an algorithm choose an entry to be deleted (oldest request or the first one that ended in error).You can at your own discretion enable strict OCSP checks by passing the environment variable
NIQUESTS_STRICT_OCSP
with anything inside but0
. In strict mode the maximum delay for response passes from 200ms to 1,000ms and raises an error or explicit warning.In non-strict mode, this security measure will be deactivated automatically if your usage is unreasonable. e.g. Making a hundred of requests to a hundred of domains, thus consuming resources that should have been allocated to browser users. This was made available for users with a limited target of domains to get a complementary security measure.
Unless in strict-mode, the proxy configuration will be respected when given, as long as it specify a plain
http
proxy. This is meant for people who want privacy.This feature may not be available if the
cryptography
package is missing from your environment. Verify the availability after Niquests upgrade by runningpython -m niquests.help
.There is several downside of using OCSP, Niquests knows it. It is not a silver bullet solution. But better than nothing. It does not apply to HTTPS proxies themselves. For now.
-
Add property
ocsp_verified
in bothPreparedRequest
, andResponse
to have a clue on the post handshake verification.Will be
None
if no verification took place,True
if the verification leads to a confirmation from the OCSP server that the certificate is valid,False
otherwise.
Changed
- Bump lower version requirement for
urllib3.future
to 2.1.900 to ensure compatibility with newer features. - Internal in-memory QUIC capabilities is now thread safe and limited to 12,288 entries.
- Pickling a
Session
object no-longer dump adapters or the QUIC in-memory capabilities, they are reset on setstate.
Fixed
conn_info
was unset if the response came after a redirect.
Changed
- niquests.help show more information about direct dependencies.
- urllib3.future minimal version raised to 2.0.936 due to an important fixe on the QUIC layer.
- wassima minimal version raised to 1.0.1 in order to support certifi as fallback in rare case.
Fixed
- Method
head
not acceptingallow_redirect
named argument. PreparedRequest
did not uppercase the HTTP verb when constructed manually.
Changed
- Minor code refactor in the
text
, andjson
methods within theResponse
class to be more concise.
Added
- Static type annotations thorough the whole package.
cert
argument for client authentication with certificate can now pass the password/passphrase using a 3-values tuple (cert, key, password). The three parameters in the tuple must be of typestr
.verify
argument behavior has been extended and now accept your CA bundle asstr
instead of a path. It also accepts your CA bundle asbytes
directly. This help when you do not have access to the fs.- Operating system truststore will be used instead of
certifi
. Root CAs are automatically grabbed from your computer configuration. - Oriented-object headers. Access them through the new property
oheaders
in yourResponse
. - Propagated the argument
retries
inniquests.api
for all functions. - Added argument
retries
in theSession
constructor. - Property
conn_info
to thePreparedRequest
andResponse
that hold a reference to aConnectionInfo
. This class exposes the following properties:certificate_der
(bytes),certificate_dict
(dict) as provided by the standard library (ssl),destination_address
(tuple[ipAddress, portNumber]),cipher
(str),tls_version
(TLSVersion), andhttp_version
. - Two hooks, namely
pre_send
andpre_request
. Thepre_request
event is fired just after the initial construction of aPreparedRequest
instance. Finally, thepre_send
will be triggered just after picking a (live) connection for your request. The two events receive aPreparedRequest
instance.
Changed
- Calling the method
json
fromResponse
when no encoding was provided no longer relies on internal encoding inference. We fall back oncharset-normalizer
with a limited set of charsets allowed (UTF-8/16/32 or ASCII). - No longer will the
text
method fromResponse
return str if content cannot be decoded. It returns None instead. - If specified charset in content-type does not exist (LookupError) the
text
method fromResponse
will rely on charset detection. - If specified charset in content-type is not made for text decoding (e.g. base64), the
text
method fromResponse
returns None. - With above four changes, the
json
method will raiseRequestsJSONDecodeError
when the payload (body) cannot be decoded. - Passing invalid
files
description no longer just skip invalid entries, it raisesValueError
from now on. - Non-str HTTP-Verb are refused.
- Passing
files
with minimal description (meaning no tuple but just the fp) no longer guess its name whenfp.name
return bytes. - No longer will the default timeout be unset, thus making you waiting indefinitely.
Functions
get
,head
, andoptions
ships with a default of 30 seconds. Thenput
,post
,patch
anddelete
uses a default of 120 seconds. Finally, therequest
function also have 120 seconds. - Basic authorization username and password are now encoded using utf-8 instead of latin-1 prior to being base64 encoded.
Removed
- Property
apparent_encoding
in favor of a discrete internal inference. - Support for the legacy
chardet
detector in case it was present in environment. Extrachardet_on_py3
is now unavailable. - requests.compat no longer hold reference to chardet.
- Deprecated
requests.packages
that was meant to avoid breakage from people importingurllib3
orchardet
within this package. They were vendored in early versions of Requests. A long time ago. - Deprecated function
get_encodings_from_content
from utils. - Deprecated function
get_unicode_from_response
from utils. - BasicAuth middleware no-longer support anything else than
bytes
orstr
for username and password. requests.compat
is stripped of every reference that no longer vary between supported interpreter version.- Charset fall back ISO-8859-1 when content-type is text and no charset was specified.
- Main function
get
,post
,put
,patch
,delete
, andhead
no longer accept kwargs. They have a fixed list of typed argument. It is no longer possible to specify non-supported additional keyword argument from aSession
instance or directly throughrequests.api
functions. e.g. functiondelete
no-longer acceptjson
, orfiles
arguments. as per RFCs specifications. You can still override this behavior through therequest
function. - Mixin classes
RequestEncodingMixin
, andRequestHooksMixin
due to OOP violations. Now deported directly into child classes. - Function
unicode_is_ascii
as it is part of the stablestr
stdlib on Python 3 or greater. - Alias function
session
forSession
context manager that was kept for BC reasons since the v1. - pyOpenSSL/urllib3 injection in case built-in ssl module does not have SNI support as it is not the case anymore for every supported interpreters.
- Constant
DEFAULT_CA_BUNDLE_PATH
, and submodulecerts
due to droppingcertifi
. - Function
extract_zipped_paths
because rendered useless as it was made to handle an edge case wherecertifi
is "zipped". - Extra
security
when installing this package. It was previously emptied in the previous major. - Warning emitted when passing a file opened in text-mode instead of binary. urllib3.future can overrule the content-length if it detects an error. You should not encounter broken request being sent.
- Support for
simplejson
if was present in environment. - Submodule
compat
.
Fixed
- An invalid content-type definition would cause the charset being evaluated to
True
, thus making the program crash. - Given
proxies
could be mutated when environment proxies were evaluated and injected. This package should not modify your inputs. For context see psf#6118 - A server could specify a
Location
header that does not comply to HTTP specifications and could lead to an unexpected exception. We try to fall back to Unicode decoding if the typical and expected Latin-1 would fail. If that fails too, a proper exception is raised. For context see psf#6026 - Top-level init now specify correctly the exposed api. Fixes mypy error
.. does not explicitly export attribute ..
.
Bugfixes
- Fix QUIC cache when using requests.request without persistent Session
Dependencies
- urllib3.future minimal version supported raised to 2.0.932 (ship with critical fixes)
Dependencies
⚠️ Switch urllib3 for urllib3.future⚠️ This may contain minor breaking changes, so we advise careful testing and reviewing.
Features
- Support for HTTP/2.
- Support for HTTP/3 over QUIC.
- Added property
http_version
that return an integer (11, 20, and 30) withinResponse
to identify which protocol has been negotiated. - Added
quic_cache_layer
mutable mapping optional argument toSession
for optional external caching of QUIC server capabilities.
Improvements
- HTTP/1.1 no longer relies on the standard library
http.client
buth11
. - Response representation
__repr__
now yield the HTTP version.
Removed
- Warning filter for
charset_normalizer
as it was removed long ago.
Security
-
Versions of Requests between v2.3.0 and v2.30.0 are vulnerable to potential forwarding of
Proxy-Authorization
headers to destination servers when following HTTPS redirects.When proxies are defined with user info (https://user:pass@proxy:8080), Requests will construct a
Proxy-Authorization
header that is attached to the request to authenticate with the proxy.In cases where Requests receives a redirect response, it previously reattached the
Proxy-Authorization
header incorrectly, resulting in the value being sent through the tunneled connection to the destination server. Users who rely on defining their proxy credentials in the URL are strongly encouraged to upgrade to Requests 2.31.0+ to prevent unintentional leakage and rotate their proxy credentials once the change has been fully deployed.Users who do not use a proxy or do not supply their proxy credentials through the user information portion of their proxy URL are not subject to this vulnerability.
Full details can be read in our Github Security Advisory and CVE-2023-32681.
Dependencies
-
⚠️ Added support for urllib3 2.0.⚠️ This may contain minor breaking changes so we advise careful testing and reviewing https://urllib3.readthedocs.io/en/latest/v2-migration-guide.html prior to upgrading.
Users who wish to stay on urllib3 1.x can pin to
urllib3<2
.
Improvements
- Requests now defers chunked requests to the urllib3 implementation to improve standardization. (#6226)
- Requests relaxes header component requirements to support bytes/str subclasses. (#6356)
Dependencies
- Requests now supports charset_normalizer 3.x. (#6261)
Bugfixes
- Updated MissingSchema exception to suggest https scheme rather than http. (#6188)
Improvements
- Speed optimization in
iter_content
with transition toyield from
. (#6170)
Dependencies
- Added support for chardet 5.0.0 (#6179)
- Added support for charset-normalizer 2.1.0 (#6169)
Deprecations
⚠️ Requests has officially dropped support for Python 2.7.⚠️ (#6091)- Requests has officially dropped support for Python 3.6 (including pypy3.6). (#6091)
Improvements
- Wrap JSON parsing issues in Request's JSONDecodeError for payloads without
an encoding to make
json()
API consistent. (#6097) - Parse header components consistently, raising an InvalidHeader error in all invalid cases. (#6154)
- Added provisional 3.11 support with current beta build. (#6155)
- Requests got a makeover and we decided to paint it black. (#6095)
Bugfixes
- Fixed bug where setting
CURL_CA_BUNDLE
to an empty string would disable cert verification. All Requests 2.x versions before 2.28.0 are affected. (#6074) - Fixed urllib3 exception leak, wrapping
urllib3.exceptions.SSLError
withrequests.exceptions.SSLError
forcontent
anditer_content
. (#6057) - Fixed issue where invalid Windows registry entries caused proxy resolution to raise an exception rather than ignoring the entry. (#6149)
- Fixed issue where entire payload could be included in the error message for JSONDecodeError. (#6036)
Bugfixes
- Fixed parsing issue that resulted in the
auth
component being dropped from proxy URLs. (#6028)
Improvements
-
Officially added support for Python 3.10. (#5928)
-
Added a
requests.exceptions.JSONDecodeError
to unify JSON exceptions between Python 2 and 3. This gets raised in theresponse.json()
method, and is backwards compatible as it inherits from previously thrown exceptions. Can be caught fromrequests.exceptions.RequestException
as well. (#5856) -
Improved error text for misnamed
InvalidSchema
andMissingSchema
exceptions. This is a temporary fix until exceptions can be renamed (Schema->Scheme). (#6017) -
Improved proxy parsing for proxy URLs missing a scheme. This will address recent changes to
urlparse
in Python 3.9+. (#5917)
Bugfixes
-
Fixed defect in
extract_zipped_paths
which could result in an infinite loop for some paths. (#5851) -
Fixed handling for
AttributeError
when calculating length of files obtained byTarfile.extractfile()
. (#5239) -
Fixed urllib3 exception leak, wrapping
urllib3.exceptions.InvalidHeader
withrequests.exceptions.InvalidHeader
. (#5914) -
Fixed bug where two Host headers were sent for chunked requests. (#5391)
-
Fixed regression in Requests 2.26.0 where
Proxy-Authorization
was incorrectly stripped from all requests sent withSession.send
. (#5924) -
Fixed performance regression in 2.26.0 for hosts with a large number of proxies available in the environment. (#5924)
-
Fixed idna exception leak, wrapping
UnicodeError
withrequests.exceptions.InvalidURL
for URLs with a leading dot (.) in the domain. (#5414)
Deprecations
- Requests support for Python 2.7 and 3.6 will be ending in 2022. While we don't have exact dates, Requests 2.27.x is likely to be the last release series providing support.
Improvements
-
Requests now supports Brotli compression, if either the
brotli
orbrotlicffi
package is installed. (#5783) -
Session.send
now correctly resolves proxy configurations from both the Session and Request. Behavior now matchesSession.request
. (#5681)
Bugfixes
- Fixed a race condition in zip extraction when using Requests in parallel from zip archive. (#5707)
Dependencies
-
Instead of
chardet
, use the MIT-licensedcharset_normalizer
for Python3 to remove license ambiguity for projects bundling requests. Ifchardet
is already installed on your machine it will be used instead ofcharset_normalizer
to keep backwards compatibility. (#5797)You can also install
chardet
while installing requests by specifying[use_chardet_on_py3]
extra as follows:pip install "requests[use_chardet_on_py3]"
Python2 still depends upon the
chardet
module. -
Requests now supports
idna
3.x on Python 3.idna
2.x will continue to be used on Python 2 installations. (#5711)
Deprecations
-
The
requests[security]
extra has been converted to a no-op install. PyOpenSSL is no longer the recommended secure option for Requests. (#5867) -
Requests has officially dropped support for Python 3.5. (#5867)
Bugfixes
- Requests now treats
application/json
asutf8
by default. Resolving inconsistencies betweenr.text
andr.json
output. (#5673)
Dependencies
- Requests now supports chardet v4.x.
Improvements
- Added support for NETRC environment variable. (#5643)
Dependencies
- Requests now supports urllib3 v1.26.
Deprecations
- Requests v2.25.x will be the last release series with support for Python 3.5.
- The
requests[security]
extra is officially deprecated and will be removed in Requests v2.26.0.
Improvements
-
pyOpenSSL TLS implementation is now only used if Python either doesn't have an
ssl
module or doesn't support SNI. Previously pyOpenSSL was unconditionally used if available. This applies even if pyOpenSSL is installed via therequests[security]
extra (#5443) -
Redirect resolution should now only occur when
allow_redirects
is True. (#5492) -
No longer perform unnecessary Content-Length calculation for requests that won't use it. (#5496)
Improvements
- Remove defunct reference to
prefetch
in Session__attrs__
(#5110)
Bugfixes
- Requests no longer outputs password in basic auth usage warning. (#5099)
Dependencies
- Pinning for
chardet
andidna
now uses major version instead of minor. This hopefully reduces the need for releases every time a dependency is updated.
Dependencies
- Requests now supports urllib3 v1.25.2. (note: 1.25.0 and 1.25.1 are incompatible)
Deprecations
- Requests has officially stopped support for Python 3.4.
Dependencies
- Requests now supports idna v2.8.
Bugfixes
- Fixed bug with unintended Authorization header stripping for redirects using default ports (http/80, https/443).
Bugfixes
- Content-Type header parsing is now case-insensitive (e.g. charset=utf8 v Charset=utf8).
- Fixed exception leak where certain redirect urls would raise uncaught urllib3 exceptions.
- Requests removes Authorization header from requests redirected from https to http on the same hostname. (CVE-2018-18074)
should_bypass_proxies
now handles URIs without hostnames (e.g. files).
Dependencies
- Requests now supports urllib3 v1.24.
Deprecations
- Requests has officially stopped support for Python 2.6.
Bugfixes
- Fixed issue where status_codes.py's
init
function failed trying to append to a__doc__
value ofNone
.
Improvements
- Warn user about possible slowdown when using cryptography version < 1.3.4
- Check for invalid host in proxy URL, before forwarding request to adapter.
- Fragments are now properly maintained across redirects. (RFC7231 7.1.2)
- Removed use of cgi module to expedite library load time.
- Added support for SHA-256 and SHA-512 digest auth algorithms.
- Minor performance improvement to
Request.content
. - Migrate to using collections.abc for 3.7 compatibility.
Bugfixes
- Parsing empty
Link
headers withparse_header_links()
no longer return one bogus entry. - Fixed issue where loading the default certificate bundle from a zip
archive would raise an
IOError
. - Fixed issue with unexpected
ImportError
on windows system which do not supportwinreg
module. - DNS resolution in proxy bypass no longer includes the username and password in the request. This also fixes the issue of DNS queries failing on macOS.
- Properly normalize adapter prefixes for url comparison.
- Passing
None
as a file pointer to thefiles
param no longer raises an exception. - Calling
copy
on aRequestsCookieJar
will now preserve the cookie policy correctly.
Dependencies
- We now support idna v2.7.
- We now support urllib3 v1.23.
Improvements
- Error messages for invalid headers now include the header name for easier debugging
Dependencies
- We now support idna v2.6.
Improvements
- Running
$ python -m requests.help
now includes the installed version of idna.
Bugfixes
- Fixed issue where Requests would raise
ConnectionError
instead ofSSLError
when encountering SSL problems when using urllib3 v1.22.
Bugfixes
requests.help
no longer fails on Python 2.6 due to the absence ofssl.OPENSSL_VERSION_NUMBER
.
Dependencies
- We now support urllib3 v1.22.
Bugfixes
- Fix an error in the packaging whereby the
*.whl
contained incorrect data that regressed the fix in v2.17.3.
Improvements
Response
is now a context manager, so can be used directly in awith
statement without first having to be wrapped bycontextlib.closing()
.
Bugfixes
- Resolve installation failure if multiprocessing is not available
- Resolve tests crash if multiprocessing is not able to determine the number of CPU cores
- Resolve error swallowing in utils set_environ generator
Improvements
- Improved
packages
namespace identity support, for monkeypatching libraries.
Improvements
- Improved
packages
namespace identity support, for monkeypatching libraries.
Improvements
- Improved
packages
namespace identity support, for monkeypatching libraries.
Improvements
- Removal of the 301 redirect cache. This improves thread-safety.
- Improvements to
$ python -m requests.help
.
- Introduction of the
$ python -m requests.help
command, for debugging with maintainers!
- Further restored the
requests.packages
namespace for compatibility reasons.
- Further restored the
requests.packages
namespace for compatibility reasons.
No code modification (noted below) should be necessary any longer.
- Restored the
requests.packages
namespace for compatibility reasons. - Bugfix for
urllib3
version parsing.
Note: code that was written to import against the
requests.packages
namespace previously will have to import code that
rests at this module-level now.
For example:
from requests.packages.urllib3.poolmanager import PoolManager
Will need to be re-written to be:
from requests.packages import urllib3
urllib3.poolmanager.PoolManager
Or, even better:
from urllib3.poolmanager import PoolManager
- Unvendor ALL the things!
- Everyone makes mistakes.
Improvements
- Introduction of the
Response.next
property, for getting the nextPreparedResponse
from a redirect chain (whenallow_redirects=False
). - Internal refactoring of
__version__
module.
Bugfixes
- Restored once-optional parameter for
requests.utils.get_environ_proxies()
.
Bugfixes
- Changed a less-than to an equal-to and an or in the dependency markers to widen compatibility with older setuptools releases.
Bugfixes
- Changed the dependency markers to widen compatibility with older pip releases.
Improvements
- It is now possible to pass
no_proxy
as a key to theproxies
dictionary to provide handling similar to theNO_PROXY
environment variable. - When users provide invalid paths to certificate bundle files or
directories Requests now raises
IOError
, rather than failing at the time of the HTTPS request with a fairly inscrutable certificate validation error. - The behavior of
SessionRedirectMixin
was slightly altered.resolve_redirects
will now detect a redirect by callingget_redirect_target(response)
instead of directly queryingResponse.is_redirect
andResponse.headers['location']
. Advanced users will be able to process malformed redirects more easily. - Changed the internal calculation of elapsed request time to have higher resolution on Windows.
- Added
win_inet_pton
as conditional dependency for the[socks]
extra on Windows with Python 2.7. - Changed the proxy bypass implementation on Windows: the proxy bypass check doesn't use forward and reverse DNS requests anymore
- URLs with schemes that begin with
http
but are nothttp
orhttps
no longer have their host parts forced to lowercase.
Bugfixes
- Much improved handling of non-ASCII
Location
header values in redirects. FewerUnicodeDecodeErrors
are encountered on Python 2, and Python 3 now correctly understands that Latin-1 is unlikely to be the correct encoding. - If an attempt to
seek
file to find out its length fails, we now appropriately handle that by aborting our content-length calculations. - Restricted
HTTPDigestAuth
to only respond to auth challenges made on 4XX responses, rather than to all auth challenges. - Fixed some code that was firing
DeprecationWarning
on Python 3.6. - The dismayed person emoticon (
/o\\
) no longer has a big head. I'm sure this is what you were all worrying about most.
Miscellaneous
- Updated bundled urllib3 to v1.21.1.
- Updated bundled chardet to v3.0.2.
- Updated bundled idna to v2.5.
- Updated bundled certifi to 2017.4.17.
Features
- Only load the
idna
library when we've determined we need it. This will save some memory for users.
Miscellaneous
- Updated bundled urllib3 to 1.20.
- Updated bundled idna to 2.2.
Bugfixes
- Fixed an issue with JSON encoding detection, specifically detecting big-endian UTF-32 with BOM.
Bugfixes
- Fixed regression from 2.12.2 where non-string types were rejected in the basic auth parameters. While support for this behaviour has been re-added, the behaviour is deprecated and will be removed in the future.
Bugfixes
- Fixed regression from v2.12.1 for URLs with schemes that begin with "http". These URLs have historically been processed as though they were HTTP-schemed URLs, and so have had parameters added. This was removed in v2.12.2 in an overzealous attempt to resolve problems with IDNA-encoding those URLs. This change was reverted: the other fixes for IDNA-encoding have been judged to be sufficient to return to the behaviour Requests had before v2.12.0.
Bugfixes
- Fixed several issues with IDNA-encoding URLs that are technically invalid but which are widely accepted. Requests will now attempt to IDNA-encode a URL if it can but, if it fails, and the host contains only ASCII characters, it will be passed through optimistically. This will allow users to opt-in to using IDNA2003 themselves if they want to, and will also allow technically invalid but still common hostnames.
- Fixed an issue where URLs with leading whitespace would raise
InvalidSchema
errors. - Fixed an issue where some URLs without the HTTP or HTTPS schemes would still have HTTP URL preparation applied to them.
- Fixed an issue where Unicode strings could not be used in basic auth.
- Fixed an issue encountered by some Requests plugins where
constructing a Response object would cause
Response.content
to raise anAttributeError
.
Bugfixes
- Updated setuptools 'security' extra for the new PyOpenSSL backend in urllib3.
Miscellaneous
- Updated bundled urllib3 to 1.19.1.
Improvements
- Updated support for internationalized domain names from IDNA2003 to IDNA2008. This updated support is required for several forms of IDNs and is mandatory for .de domains.
- Much improved heuristics for guessing content lengths: Requests will
no longer read an entire
StringIO
into memory. - Much improved logic for recalculating
Content-Length
headers forPreparedRequest
objects. - Improved tolerance for file-like objects that have no
tell
method but do have aseek
method. - Anything that is a subclass of
Mapping
is now treated like a dictionary by thedata=
keyword argument. - Requests now tolerates empty passwords in proxy credentials, rather than stripping the credentials.
- If a request is made with a file-like object as the body and that request is redirected with a 307 or 308 status code, Requests will now attempt to rewind the body object so it can be replayed.
Bugfixes
- When calling
response.close
, the call toclose
will be propagated through to non-urllib3 backends. - Fixed issue where the
ALL_PROXY
environment variable would be preferred over scheme-specific variables likeHTTP_PROXY
. - Fixed issue where non-UTF8 reason phrases got severely mangled by falling back to decoding using ISO 8859-1 instead.
- Fixed a bug where Requests would not correctly correlate cookies set when using custom Host headers if those Host headers did not use the native string type for the platform.
Miscellaneous
- Updated bundled urllib3 to 1.19.
- Updated bundled certifi certs to 2016.09.26.
Bugfixes
- Fixed a bug when using
iter_content
withdecode_unicode=True
for streamed bodies would raiseAttributeError
. This bug was introduced in 2.11. - Strip Content-Type and Transfer-Encoding headers from the header block when following a redirect that transforms the verb from POST/PUT to GET.
Improvements
- Added support for the
ALL_PROXY
environment variable. - Reject header values that contain leading whitespace or newline characters to reduce risk of header smuggling.
Bugfixes
- Fixed occasional
TypeError
when attempting to decode a JSON response that occurred in an error case. Now correctly returns aValueError
. - Requests would incorrectly ignore a non-CIDR IP address in the
NO_PROXY
environment variables: Requests now treats it as a specific IP. - Fixed a bug when sending JSON data that could cause us to encounter obscure OpenSSL errors in certain network conditions (yes, really).
- Added type checks to ensure that
iter_content
only accepts integers andNone
for chunk sizes. - Fixed issue where responses whose body had not been fully consumed
would have the underlying connection closed but not returned to the
connection pool, which could cause Requests to hang in situations
where the
HTTPAdapter
had been configured to use a blocking connection pool.
Miscellaneous
- Updated bundled urllib3 to 1.16.
- Some previous releases accidentally accepted non-strings as acceptable header values. This release does not.
New Features
- SOCKS Proxy Support! (requires PySocks;
$ pip install requests[socks]
)
Miscellaneous
- Updated bundled urllib3 to 1.15.1.
Improvements
- Change built-in CaseInsensitiveDict (used for headers) to use OrderedDict as its underlying datastore.
Bugfixes
- Don't use redirect_cache if allow_redirects=False
- When passed objects that throw exceptions from
tell()
, send them via chunked transfer encoding instead of failing. - Raise a ProxyError for proxy related connection issues.
Bugfixes
- Resolve regression introduced in 2.9.0 that made it impossible to send binary strings as bodies in Python 3.
- Fixed errors when calculating cookie expiration dates in certain locales.
Miscellaneous
- Updated bundled urllib3 to 1.13.1.
Minor Improvements (Backwards compatible)
- The
verify
keyword argument now supports being passed a path to a directory of CA certificates, not just a single-file bundle. - Warnings are now emitted when sending files opened in text mode.
- Added the 511 Network Authentication Required status code to the status code registry.
Bugfixes
- For file-like objects that are not sought to the very beginning, we now send the content length for the number of bytes we will actually read, rather than the total size of the file, allowing partial file uploads.
- When uploading file-like objects, if they are empty or have no
obvious content length we set
Transfer-Encoding: chunked
rather thanContent-Length: 0
. - We correctly receive the response in buffered mode when uploading chunked bodies.
- We now handle being passed a query string as a bytestring on Python 3, by decoding it as UTF-8.
- Sessions are now closed in all cases (exceptional and not) when using the functional API rather than leaking and waiting for the garbage collector to clean them up.
- Correctly handle digest auth headers with a malformed
qop
directive that contains no token, by treating it the same as if noqop
directive was provided at all. - Minor performance improvements when removing specific cookies by name.
Miscellaneous
- Updated urllib3 to 1.13.
Bugfixes
- Update certificate bundle to match
certifi
2015.9.6.2's weak certificate bundle. - Fix a bug in 2.8.0 where requests would raise
ConnectTimeout
instead ofConnectionError
- When using the PreparedRequest flow, requests will now correctly
respect the
json
parameter. Broken in 2.8.0. - When using the PreparedRequest flow, requests will now correctly handle a Unicode-string method name on Python 2. Broken in 2.8.0.
Minor Improvements (Backwards Compatible)
- Requests now supports per-host proxies. This allows the
proxies
dictionary to have entries of the form{'<scheme>://<hostname>': '<proxy>'}
. Host-specific proxies will be used in preference to the previously-supported scheme-specific ones, but the previous syntax will continue to work. Response.raise_for_status
now prints the URL that failed as part of the exception message.requests.utils.get_netrc_auth
now takes anraise_errors
kwarg, defaulting toFalse
. WhenTrue
, errors parsing.netrc
files cause exceptions to be thrown.- Change to bundled projects import logic to make it easier to unbundle requests downstream.
- Changed the default User-Agent string to avoid leaking data on Linux: now contains only the requests version.
Bugfixes
- The
json
parameter topost()
and friends will now only be used if neitherdata
norfiles
are present, consistent with the documentation. - We now ignore empty fields in the
NO_PROXY
environment variable. - Fixed problem where
httplib.BadStatusLine
would get raised if combiningstream=True
withcontextlib.closing
. - Prevented bugs where we would attempt to return the same connection back to the connection pool twice when sending a Chunked body.
- Miscellaneous minor internal changes.
- Digest Auth support is now thread safe.
Updates
- Updated urllib3 to 1.12.
This is the first release that follows our new release process. For more, see our documentation.
Bugfixes
- Updated urllib3 to 1.10.4, resolving several bugs involving chunked transfer encoding and response framing.
Bugfixes
- Fix regression where compressed data that was sent as chunked data was not properly decompressed. (#2561)
Bugfixes
- Remove VendorAlias import machinery introduced in v2.5.2.
- Simplify the PreparedRequest.prepare API: We no longer require the user to pass an empty list to the hooks keyword argument. (c.f. #2552)
- Resolve redirects now receives and forwards all of the original arguments to the adapter. (#2503)
- Handle UnicodeDecodeErrors when trying to deal with a unicode URL that cannot be encoded in ASCII. (#2540)
- Populate the parsed path of the URI field when performing Digest Authentication. (#2426)
- Copy a PreparedRequest's CookieJar more reliably when it is not an instance of RequestsCookieJar. (#2527)
Bugfixes
- CVE-2015-2296: Fix handling of cookies on redirect. Previously a cookie without a host value set would use the hostname for the redirected URL exposing requests users to session fixation attacks and potentially cookie stealing. This was disclosed privately by Matthew Daley of BugFuzz. This affects all versions of requests from v2.1.0 to v2.5.3 (inclusive on both ends).
- Fix error when requests is an
install_requires
dependency andpython setup.py test
is run. (#2462) - Fix error when urllib3 is unbundled and requests continues to use the vendored import location.
- Include fixes to
urllib3
's header handling. - Requests' handling of unvendored dependencies is now more restrictive.
Features and Improvements
- Support bytearrays when passed as parameters in the
files
argument. (#2468) - Avoid data duplication when creating a request with
str
,bytes
, orbytearray
input to thefiles
argument.
Bugfixes
- Revert changes to our vendored certificate bundle. For more context see (#2455, #2456, and https://bugs.python.org/issue23476)
Features and Improvements
- Add sha256 fingerprint support. (shazow/urllib3#540)
- Improve the performance of headers. (shazow/urllib3#544)
Bugfixes
- Copy pip's import machinery. When downstream redistributors remove requests.packages.urllib3 the import machinery will continue to let those same symbols work. Example usage in requests' documentation and 3rd-party libraries relying on the vendored copies of urllib3 will work without having to fallback to the system urllib3.
- Attempt to quote parts of the URL on redirect if unquoting and then quoting fails. (#2356)
- Fix filename type check for multipart form-data uploads. (#2411)
- Properly handle the case where a server issuing digest authentication challenges provides both auth and auth-int qop-values. (#2408)
- Fix a socket leak. (shazow/urllib3#549)
- Fix multiple
Set-Cookie
headers properly. (shazow/urllib3#534) - Disable the built-in hostname verification. (shazow/urllib3#526)
- Fix the behaviour of decoding an exhausted stream. (shazow/urllib3#535)
Security
- Pulled in an updated
cacert.pem
. - Drop RC4 from the default cipher list. (shazow/urllib3#551)
Behavioural Changes
- Only catch HTTPErrors in raise_for_status (#2382)
Bugfixes
- Handle LocationParseError from urllib3 (#2344)
- Handle file-like object filenames that are not strings (#2379)
- Unbreak HTTPDigestAuth handler. Allow new nonces to be negotiated (#2389)
Improvements
- Allow usage of urllib3's Retry object with HTTPAdapters (#2216)
- The
iter_lines
method on a response now accepts a delimiter with which to split the content (#2295)
Behavioural Changes
- Add deprecation warnings to functions in requests.utils that will be removed in 3.0 (#2309)
- Sessions used by the functional API are always closed (#2326)
- Restrict requests to HTTP/1.1 and HTTP/1.0 (stop accepting HTTP/0.9) (#2323)
Bugfixes
- Only parse the URL once (#2353)
- Allow Content-Length header to always be overridden (#2332)
- Properly handle files in HTTPDigestAuth (#2333)
- Cap redirect_cache size to prevent memory abuse (#2299)
- Fix HTTPDigestAuth handling of redirects after authenticating successfully (#2253)
- Fix crash with custom method parameter to Session.request (#2317)
- Fix how Link headers are parsed using the regular expression library (#2271)
Documentation
- Add more references for interlinking (#2348)
- Update CSS for theme (#2290)
- Update width of buttons and sidebar (#2289)
- Replace references of Gittip with Gratipay (#2282)
- Add link to changelog in sidebar (#2273)
Bugfixes
- Unicode URL improvements for Python 2.
- Re-order JSON param for backwards compat.
- Automatically defrag authentication schemes from host/pass URIs. (#2249)
Improvements
Bugfixes
- Avoid getting stuck in a loop (#2244)
- Multiple calls to iter* fail with unhelpful error. (#2240, #2241)
Documentation
- Correct redirection introduction (#2245)
- Added example of how to send multiple files in one request. (#2227)
- Clarify how to pass a custom set of CAs (#2248)
- Now has a "security" package extras set,
$ pip install requests[security]
- Requests will now use Certifi if it is available.
- Capture and re-raise urllib3 ProtocolError
- Bugfix for responses that attempt to redirect to themselves forever (wtf?).
Behavioral Changes
Connection: keep-alive
header is now sent automatically.
Improvements
- Support for connect timeouts! Timeout now accepts a tuple (connect, read) which is used to set individual connect and read timeouts.
- Allow copying of PreparedRequests without headers/cookies.
- Updated bundled urllib3 version.
- Refactored settings loading from environment -- new Session.merge_environment_settings.
- Handle socket errors in iter_content.
API Changes
- New
Response
propertyis_redirect
, which is true when the library could have processed this response as a redirection (whether or not it actually did). - The
timeout
parameter now affects requests with bothstream=True
andstream=False
equally. - The change in v2.0.0 to mandate explicit proxy schemes has been
reverted. Proxy schemes now default to
http://
. - The
CaseInsensitiveDict
used for HTTP headers now behaves like a normal dictionary when references as string or viewed in the interpreter.
Bugfixes
- No longer expose Authorization or Proxy-Authorization headers on redirect. Fix CVE-2014-1829 and CVE-2014-1830 respectively.
- Authorization is re-evaluated each redirect.
- On redirect, pass url as native strings.
- Fall-back to autodetected encoding for JSON when Unicode detection fails.
- Headers set to
None
on theSession
are now correctly not sent. - Correctly honor
decode_unicode
even if it wasn't used earlier in the same response. - Stop advertising
compress
as a supported Content-Encoding. - The
Response.history
parameter is now always a list. - Many, many
urllib3
bugfixes.
Bugfixes
- Fixes incorrect parsing of proxy credentials that contain a literal or encoded '#' character.
- Assorted urllib3 fixes.
API Changes
- New exception:
ContentDecodingError
. Raised instead ofurllib3
DecodeError
exceptions.
Bugfixes
- Avoid many many exceptions from the buggy implementation of
proxy_bypass
on OS X in Python 2.6. - Avoid crashing when attempting to get authentication credentials from ~/.netrc when running as a user without a home directory.
- Use the correct pool size for pools of connections to proxies.
- Fix iteration of
CookieJar
objects. - Ensure that cookies are persisted over redirect.
- Switch back to using chardet, since it has merged with charade.
- Updated CA Bundle, of course.
- Cookies set on individual Requests through a
Session
(e.g. viaSession.get()
) are no longer persisted to theSession
. - Clean up connections when we hit problems during chunked upload, rather than leaking them.
- Return connections to the pool when a chunked upload is successful, rather than leaking it.
- Match the HTTPbis recommendation for HTTP 301 redirects.
- Prevent hanging when using streaming uploads and Digest Auth when a 401 is received.
- Values of headers set by Requests are now always the native string type.
- Fix previously broken SNI support.
- Fix accessing HTTP proxies using proxy authentication.
- Unencode HTTP Basic usernames and passwords extracted from URLs.
- Support for IP address ranges for no_proxy environment variable
- Parse headers correctly when users override the default
Host:
header. - Avoid munging the URL in case of case-sensitive servers.
- Looser URL handling for non-HTTP/HTTPS urls.
- Accept unicode methods in Python 2.6 and 2.7.
- More resilient cookie handling.
- Make
Response
objects pickleable. - Actually added MD5-sess to Digest Auth instead of pretending to like last time.
- Updated internal urllib3.
- Fixed @Lukasa's lack of taste.
- Updated included CA Bundle with new mistrusts and automated process for the future
- Added MD5-sess to Digest Auth
- Accept per-file headers in multipart file POST messages.
- Fixed: Don't send the full URL on CONNECT messages.
- Fixed: Correctly lowercase a redirect scheme.
- Fixed: Cookies not persisted when set via functional API.
- Fixed: Translate urllib3 ProxyError into a requests ProxyError derived from ConnectionError.
- Updated internal urllib3 and chardet.
API Changes:
- Keys in the Headers dictionary are now native strings on all Python versions, i.e. bytestrings on Python 2, unicode on Python 3.
- Proxy URLs now must have an explicit scheme. A
MissingSchema
exception will be raised if they don't. - Timeouts now apply to read time if
Stream=False
. RequestException
is now a subclass ofIOError
, notRuntimeError
.- Added new method to
PreparedRequest
objects:PreparedRequest.copy()
. - Added new method to
Session
objects:Session.update_request()
. This method updates aRequest
object with the data (e.g. cookies) stored on theSession
. - Added new method to
Session
objects:Session.prepare_request()
. This method updates and prepares aRequest
object, and returns the correspondingPreparedRequest
object. - Added new method to
HTTPAdapter
objects:HTTPAdapter.proxy_headers()
. This should not be called directly, but improves the subclass interface. httplib.IncompleteRead
exceptions caused by incorrect chunked encoding will now raise a RequestsChunkedEncodingError
instead.- Invalid percent-escape sequences now cause a Requests
InvalidURL
exception to be raised. - HTTP 208 no longer uses reason phrase
"im_used"
. Correctly uses"already_reported"
. - HTTP 226 reason added (
"im_used"
).
Bugfixes:
- Vastly improved proxy support, including the CONNECT verb. Special thanks to the many contributors who worked towards this improvement.
- Cookies are now properly managed when 401 authentication responses are received.
- Chunked encoding fixes.
- Support for mixed case schemes.
- Better handling of streaming downloads.
- Retrieve environment proxies from more locations.
- Minor cookies fixes.
- Improved redirect behaviour.
- Improved streaming behaviour, particularly for compressed data.
- Miscellaneous small Python 3 text encoding bugs.
.netrc
no longer overrides explicit auth.- Cookies set by hooks are now correctly persisted on Sessions.
- Fix problem with cookies that specify port numbers in their host field.
BytesIO
can be used to perform streaming uploads.- More generous parsing of the
no_proxy
environment variable. - Non-string objects can be passed in data values alongside files.
- Simple packaging fix
- Simple packaging fix
- 301 and 302 redirects now change the verb to GET for all verbs, not just POST, improving browser compatibility.
- Python 3.3.2 compatibility
- Always percent-encode location headers
- Fix connection adapter matching to be most-specific first
- new argument to the default connection adapter for passing a block argument
- prevent a KeyError when there's no link headers
- Fixed cookies on sessions and on requests
- Significantly change how hooks are dispatched - hooks now receive all the arguments specified by the user when making a request so hooks can make a secondary request with the same parameters. This is especially necessary for authentication handler authors
- certifi support was removed
- Fixed bug where using OAuth 1 with body
signature_type
sent no data - Major proxy work thanks to @Lukasa including parsing of proxy authentication from the proxy url
- Fix DigestAuth handling too many 401s
- Update vendored urllib3 to include SSL bug fixes
- Allow keyword arguments to be passed to
json.loads()
via theResponse.json()
method - Don't send
Content-Length
header by default onGET
orHEAD
requests - Add
elapsed
attribute toResponse
objects to time how long a request took. - Fix
RequestsCookieJar
- Sessions and Adapters are now picklable, i.e., can be used with the multiprocessing library
- Update charade to version 1.0.3
The change in how hooks are dispatched will likely cause a great deal of issues.
- CHUNKED REQUESTS
- Support for iterable response bodies
- Assume servers persist redirect params
- Allow explicit content types to be specified for file data
- Make merge_kwargs case-insensitive when looking up keys
- Fix file upload encoding bug
- Fix cookie behavior
- Proxy fix for HTTPAdapter.
- Cert verification exception bug.
- Proxy fix for HTTPAdapter.
- Massive Refactor and Simplification
- Switch to Apache 2.0 license
- Swappable Connection Adapters
- Mountable Connection Adapters
- Mutable ProcessedRequest chain
- /s/prefetch/stream
- Removal of all configuration
- Standard library logging
- Make Response.json() callable, not property.
- Usage of new charade project, which provides python 2 and 3 simultaneous chardet.
- Removal of all hooks except 'response'
- Removal of all authentication helpers (OAuth, Kerberos)
This is not a backwards compatible change.
- Improved mime-compatible JSON handling
- Proxy fixes
- Path hack fixes
- Case-Insensitive Content-Encoding headers
- Support for CJK parameters in form posts
- Python 3.3 Compatibility
- Simply default accept-encoding
- Bugfixes
- No more iter_content errors if already downloaded.
- Fix for OAuth + POSTs
- Remove exception eating from dispatch_hook
- General bugfixes
- Incredible Link header support :)
- Support for (key, value) lists everywhere.
- Digest Authentication improvements.
- Ensure proxy exclusions work properly.
- Clearer UnicodeError exceptions.
- Automatic casting of URLs to strings (fURL and such)
- Bugfixes.
- Long awaited fix for hanging connections!
- Packaging fix
- GSSAPI/Kerberos authentication!
- App Engine 2.7 Fixes!
- Fix leaking connections (from urllib3 update)
- OAuthlib path hack fix
- OAuthlib URL parameters fix.
- Use simplejson if available.
- Do not hide SSLErrors behind Timeouts.
- Fixed param handling with urls containing fragments.
- Significantly improved information in User Agent.
- client certificates are ignored when verify=False
- Zero dependencies (once again)!
- New: Response.reason
- Sign querystring parameters in OAuth 1.0
- Client certificates no longer ignored when verify=False
- Add openSUSE certificate support
- Allow passing a file or file-like object as data.
- Allow hooks to return responses that indicate errors.
- Fix Response.text and Response.json for body-less responses.
- Removal of Requests.async in favor of grequests
- Allow disabling of cookie persistence.
- New implementation of safe_mode
- cookies.get now supports default argument
- Session cookies not saved when Session.request is called with return_response=False
- Env: no_proxy support.
- RequestsCookieJar improvements.
- Various bug fixes.
- New
Response.json
property. - Ability to add string file uploads.
- Fix out-of-range issue with iter_lines.
- Fix iter_content default size.
- Fix POST redirects containing files.
- EXPERIMENTAL OAUTH SUPPORT!
- Proper CookieJar-backed cookies interface with awesome dict-like interface.
- Speed fix for non-iterated content chunks.
- Move
pre_request
to a more usable place. - New
pre_send
hook. - Lazily encode data, params, files.
- Load system Certificate Bundle if
certify
isn't available. - Cleanups, fixes.
- Attempt to use the OS's certificate bundle if
certifi
isn't available. - Infinite digest auth redirect fix.
- Multi-part file upload improvements.
- Fix decoding of invalid %encodings in URLs.
- If there is no content in a response don't throw an error the second time that content is attempted to be read.
- Upload data on redirects.
- POST redirects now break RFC to do what browsers do: Follow up with a GET.
- New
strict_mode
configuration to disable new redirect behavior.
- Private SSL Certificate support
- Remove select.poll from Gevent monkeypatching
- Remove redundant generator for chunked transfer encoding
- Fix: Response.ok raises Timeout Exception in safe_mode
- Generate chunked ValueError fix
- Proxy configuration by environment variables
- Simplification of iter_lines.
- New trust_env configuration for disabling system/environment hints.
- Suppress cookie errors.
- encode_uri = False
- Allow '=' in cookies.
- Response body with 0 content-length fix.
- New async.imap.
- Don't fail on netrc.
- Honor netrc.
- HEAD requests don't follow redirects anymore.
- raise_for_status() doesn't raise for 3xx anymore.
- Make Session objects picklable.
- ValueError for invalid schema URLs.
- Vastly improved URL quoting.
- Additional allowed cookie key values.
- Attempted fix for "Too many open files" Error
- Replace unicode errors on first pass, no need for second pass.
- Append '/' to bare-domain urls before query insertion.
- Exceptions now inherit from RuntimeError.
- Binary uploads + auth fix.
- Bugfixes.
- PYTHON 3 SUPPORT!
- Dropped 2.5 Support. (Backwards Incompatible)
Response.content
is now bytes-only. (Backwards Incompatible)- New
Response.text
is unicode-only. - If no
Response.encoding
is specified andchardet
is available,Response.text
will guess an encoding. - Default to ISO-8859-1 (Western) encoding for "text" subtypes.
- Removal of decode_unicode. (Backwards Incompatible)
- New multiple-hooks system.
- New
Response.register_hook
for registering hooks within the pipeline. Response.url
is now Unicode.
- SSL verify=False bugfix (apparent on windows machines).
- Asynchronous async.send method.
- Support for proper chunk streams with boundaries.
- session argument for Session classes.
- Print entire hook tracebacks, not just exception instance.
- Fix response.iter_lines from pending next line.
- Fix but in HTTP-digest auth w/ URI having query strings.
- Fix in Event Hooks section.
- Urllib3 update.
- danger_mode for automatic Response.raise_for_status()
- Response.iter_lines refactor
- verify ssl is default.
- Packaging fix.
- SSL CERT VERIFICATION!
- Release of Cerifi: Mozilla's cert list.
- New 'verify' argument for SSL requests.
- Urllib3 update.
- iter_lines last-line truncation fix
- Force safe_mode for async requests
- Handle safe_mode exceptions more consistently
- Fix iteration on null responses in safe_mode
- Socket timeout fixes.
- Proxy Authorization support.
- Response.iter_lines!
- Prefetch bugfix.
- Added license to installed version.
- Converted auth system to use simpler callable objects.
- New session parameter to API methods.
- Display full URL while logging.
- New Unicode decoding system, based on over-ridable Response.encoding.
- Proper URL slash-quote handling.
- Cookies with
[
,]
, and_
allowed.
- URL Request path fix
- Proxy fix.
- Timeouts fix.
- Keep-alive support!
- Complete removal of Urllib2
- Complete removal of Poster
- Complete removal of CookieJars
- New ConnectionError raising
- Safe_mode for error catching
- prefetch parameter for request methods
- OPTION method
- Async pool size throttling
- File uploads send real names
- Vendored in urllib3
- Digest authentication bugfix (attach query data to path)
- Response.content = None if there was an invalid response.
- Redirection auth handling.
- Session Hooks fix.
- Digest Auth fix.
- PATCH Fix.
- Move away from urllib2 authentication handling.
- Fully Remove AuthManager, AuthObject, &c.
- New tuple-based auth system with handler callbacks.
- Sessions are now the primary interface.
- Deprecated InvalidMethodException.
- PATCH fix.
- New config system (no more global settings).
- Session parameter bugfix (params merging).
- Offline (fast) test suite.
- Session dictionary argument merging.
- Automatic decoding of unicode, based on HTTP Headers.
- New
decode_unicode
setting. - Removal of
r.read/close
methods. - New
r.faw
interface for advanced response usage.* - Automatic expansion of parameterized headers.
- Beautiful
requests.async
module, for making async requests w/ gevent.
- GET/HEAD obeys allow_redirects=False.
- Enhanced status codes experience
\o/
- Set a maximum number of redirects (
settings.max_redirects
) - Full Unicode URL support
- Support for protocol-less redirects.
- Allow for arbitrary request types.
- Bugfixes
- New callback hook system
- New persistent sessions object and context manager
- Transparent Dict-cookie handling
- Status code reference object
- Removed Response.cached
- Added Response.request
- All args are kwargs
- Relative redirect support
- HTTPError handling improvements
- Improved https testing
- Bugfixes
- International Domain Name Support!
- Access headers without fetching entire body (
read()
) - Use lists as dicts for parameters
- Add Forced Basic Authentication
- Forced Basic is default authentication type
python-requests.org
default User-Agent header- CaseInsensitiveDict lower-case caching
- Response.history bugfix
- PATCH Support
- Support for Proxies
- HTTPBin Test Suite
- Redirect Fixes
- settings.verbose stream writing
- Querystrings for all methods
- URLErrors (Connection Refused, Timeout, Invalid URLs) are treated as
explicitly raised
r.requests.get('hwe://blah'); r.raise_for_status()
- Improved Redirection Handling
- New 'allow_redirects' param for following non-GET/HEAD Redirects
- Settings module refactoring
- Response.history: list of redirected responses
- Case-Insensitive Header Dictionaries!
- Unicode URLs
- Urllib2 HTTPAuthentication Recursion fix (Basic/Digest)
- Internal Refactor
- Bytes data upload Bugfix
- Request timeouts
- Unicode url-encoded data
- Settings context manager and module
- Automatic Decompression of GZip Encoded Content
- AutoAuth Support for Tupled HTTP Auth
- Cookie Changes
- Response.read()
- Poster fix
- Automatic Authentication API Change
- Smarter Query URL Parameterization
- Allow file uploads and POST data together
New Authentication Manager System
: - Simpler Basic HTTP System
- Supports all built-in urllib2 Auths
- Allows for custom Auth Handlers
- Python 2.5 Support
- PyPy-c v1.4 Support
- Auto-Authentication tests
- Improved Request object constructor
New HTTPHandling Methods
: - Response.\_\_nonzero\_\_ (false if bad HTTP Status)
- Response.ok (True if expected HTTP Status)
- Response.error (Logged HTTPError if bad HTTP Status)
- Response.raise\_for\_status() (Raises stored HTTPError)
- Still handles request in the event of an HTTPError. (Issue #2)
- Eventlet and Gevent Monkeypatch support.
- Cookie Support (Issue #1)
- Added file attribute to POST and PUT requests for multipart-encode file uploads.
- Added Request.url attribute for context and redirects
- Birth!
- Frustration
- Conception