- Added official support for Python 3.8 and Django 3.0.
- Added settings for expected audience and issuer claims.
- Documentation updates.
- Updated package/python version support (check the README to see what new versions are supported and what old ones are no longer supported!)
- Added Chilean Spanish language support.
- Added Russian language support.
- Added
JTI_CLAIM
setting to allow storing token identifiers under a different claim.
- We now return HTTP 401 for user not found or inactive.
- Restricted
setup.py
config to Python 3 only.
- Included translation files in release package.
- Updated
python-jose
version requirement.
- Fixed
KeyError
inTokenObtainSerializer.validate
.
- Added request pass-through on
django.contrib.auth.authenticate
call inTokenObtainSerializer
. - Updated
TokenObtainSerializer
to usefail
API from parent class.
- Added language support for Brazilian Portuguese.
- Added support for automatic username lookup in
TokenUser
.
- Removed Python 2 support.
- Fixed crash when empty AUTHORIZATION header is sent.
- Fixed testing DB transaction issues.
- Simplified/improved testing and dev setup.
- Switched to using bumpversion for release process.
- Removed official support for Python 3.4.
- Added support for Python 3.7.
- Added support for Django 2.1.
- Added support for DRF 3.9.
- Fixed issue with
WWW-Authenticate
header not being included in 401 responses.
- Added missing method
get
onToken
base class.
- Simplified some blacklist app code.
- Resolved possible race condition.
- Added
TokenObtainSerializer.get_token
method to facilitate customization of token claims. - Added
TokenVerifyView
to allow verification of HMAC-signed tokens by API users who have no access to the signing key. - Renamed
AUTH_HEADER_TYPE
setting toAUTH_HEADER_TYPES
. This setting now contains either a single valid auth header type or a list or tuple of valid auth header types. If authentication fails, and more than one string is present in this tuple or list, the first item in the list will be used to build the "WWW-Authenticate" header in the response.
- Moved handling of TokenError exceptions from inside of serializer
validate
methods into token viewpost
methods.
- Added support for refresh token rotation via
ROTATE_REFRESH_TOKENS
andBLACKLIST_AFTER_ROTATION
settings. See README for details. - Added
BlacklistMixin.blacklist
method to make it easier to blacklist tokens regardless of whether or not they are present in the outstanding token list. - In token blacklist app, changed
OutstandingToken.jti
field to char field to better reflect JWT spec. - Renamed
AUTH_TOKEN_CLASS
setting toAUTH_TOKEN_CLASSES
. This setting now specifies a list of token classes (or class paths) which are used to verify tokens which are submitted for authorization. This will hopefully help anyone wishing to gradually migrate between using different token types. - Removed support for extensible JWT backends. We're just going to use PyJWT exclusively to simplify things.
- Added support for more crypto algorithms. All HMAC and RSA variants from PyJWT now supported.
- Renamed
SECRET_KEY
setting toSIGNING_KEY
. - The renamed
SIGNING_KEY
setting now acts doubly as a symmetric signing/verification key for HMAC algorithms and as a private key for RSA algorithms. - Added
VERIFYING_KEY
setting for use with RSA algorithms. - Removed undocumented
TOKEN_BACKEND_CLASS
setting.
- Switched to using PyJWT as the underlying library for signing and verifying tokens.