Add support for auto-refreshing token without refresh token#394
Add support for auto-refreshing token without refresh token#394denizdogan wants to merge 1 commit intorequests:masterfrom denizdogan:master
Conversation
|
This looks amazing and probably a documentation about it and a couple of unit tests will be ideal... |
|
@JonathanHuot Heheh, it's been a while indeed! I had completely forgot about this PR. I think at this point, it's better that you fill in the blanks, or maybe just rewrite it entirely :) Thanks |
|
This is exactly the functionality I'm looking for. Is there any chance of this being merged in the near future? |
|
@rohanliston Personally, I will not be working on adding docs or tests to this PR, I don't really use Python much anymore at all. If anyone really wants this to get merged, I would suggest maybe making a new PR with the same changes, plus some docs and a unit test. |
|
@denizdogan Thanks. I'm happy to submit a fresh PR with tests/docs, but before doing so I'd like to know if the maintainers are likely to review/merge it since there appear to be questions around the status of this project. @JonathanHuot are you able to comment on the above? |
|
@rohanliston, I can review it when ready and integrate it when ready. |
|
@JonathanHuot @rohanliston i can take some time this Sunday to review. |
|
@jtroussard The project I was working on didn't go ahead and I'm now on something else. I don't think I'll have the bandwidth to come back to this, unfortunately. |
|
@rohanliston Thanks for the heads up. |
|
The torch has been passed to #526 |
This is a basic idea on how to fix #260 since nothing has happened to the issue in three years. Let's get a discussion going, because this is a hindrance to me in almost every OAuth2 implementation I make using this library and it's clear that lots of other people have the same issue.
So the idea is to add a new constructor parameter which I currently call
auto_refresh_type, it can be either"refresh_token"(which will do exactly what it does today) or"access_token"(which will just get a new access token as you normally do).It's not a beautiful design by any means, but I intentionally tried to keep the diff as minimal as possible without any major refactorings, which would introduce breaking changes and make it more difficult to get this functionality out there.