updated tests/implementations to work with hapi 17 #340
Conversation
|
I've tested with Twitter in a live app and worked like a charm. |
| server.auth.strategy('arcgisonline', 'bell', { | ||
| provider: 'arcgisonline', | ||
| password: 'cookie_encryption_password_secure', | ||
| isSecure: false, |
There was a problem hiding this comment.
This should be true for the purposes of an example IMHO, with a comment that it should only be set to false for development.
There was a problem hiding this comment.
Well, I merely updated to async/await, and did some linting (ie top-level await is not yet possible, hence the wrap)
|
I've got this working happy-path with quite a few providers, including a custom one. However there is a change between this branch and the current bell behavior that I think is worth bringing up. If the request that bell makes to the The behavior used to be (correctly I would think) that a 500 error would be sent back to the client. |
|
FWIW, I tested this with LinkedIn tonight and it worked great. |
|
Working great with Google too. Thanks for the hard work! |
|
Any timeline on when this will be merged? Can't use hapi 17 without bell. |
|
@flipxfx you can just put in your "bell": "makeomatic/bell#hapi17"That will do the trick nicely in the meantime and will allow you to make all the development effort on your side. You can go up to production with this if you are on a small scale project or at least up to staging on anything else. That's still quite a thing. Once done, maintaining a dev branch on v17 until release should be fairly easy and require a minimal amount of work. |
|
I think that @wraithgar is correct about the fact that an error generated by a provider should result on an error 500 and not bubble up to the client. What are your thoughts on this @AVVS ? And are you willing to fix this if you are in agreement over this ? |
|
This thread has been automatically locked due to inactivity. Please open a new issue for related bugs or questions following the new issue template instructions. |
Did some refactoring, all tests pass, though I'd clean up code a bit