Re: [DISCUSS] Rest Catalog 419 Response Code

2025-02-24 Thread rdb...@gmail.com
Yeah, I don't think that this response was used. We thought that it was needed but it can probably be safely removed as I'm not aware of any implementation that sent or handled it. If that's the right thing to do because there are other more standard mechanisms for sending more information about a

Re: [DISCUSS] Rest Catalog 419 Response Code

2025-02-24 Thread Alex Dutra
Hi all, Thanks Sung for your support – I will start a discussion thread soon on the topic of expected client behavior for both 401 and 419 response codes. And out of curiosity, I looked at how the AuthenticationTimeoutResponse type was currently handled in iceberg-core, and was surprised to see t

Re: [DISCUSS] Rest Catalog 419 Response Code

2025-02-22 Thread Sung Yun
Thank you Daniel, Dmitri, Alex and Yufei for sharing your thoughts. I'm in favor of updating the REST Spec to be more clear about these status codes. Here's a PR[1] that follows the RFC language to make expected and recommended behaviors of the server and the client more explicit (thanks Dmitri f

Re: [DISCUSS] Rest Catalog 419 Response Code

2025-02-21 Thread Yufei Gu
I'm fine with 419 being considered an optional status code, in which case: 1. Servers can choose not to implement it. 2. Clients shouldn't rely solely on it for token expiration. However, the current REST spec doesn't clarify this. We should make it more explicit. We could try to deprecate

Re: [DISCUSS] Rest Catalog 419 Response Code

2025-02-21 Thread Alex Dutra
Hi all, I think there are two aspects in this issue that should be discussed separately: 1. What are the pros and cons of having a custom HTTP response code for authentication issues, as opposed to using 401? 2. How should clients deal with authentication issues, both 401 and 419? *Rega

Re: [DISCUSS] Rest Catalog 419 Response Code

2025-02-20 Thread Dmitri Bourlatchkov
Thank you for your response, Dan! I agree with all your points. On the other hand, I guess readers of the Catalog REST API spec may still be uncertain about when specific error codes MAY or SHOULD occur (using the traditional RFC language). Do you think it might be worth clarifying the REST spec

Re: [DISCUSS] Rest Catalog 419 Response Code

2025-02-20 Thread Daniel Weeks
Hey Sung, My interpretation is that it's up to the REST Server to decide whether to send a 419 or 401 response code (I don't think it's a mandate). The use case for 419 would be that the client has client credentials or can re-authenticate via some other mechanism and could reattempt the request.

[DISCUSS] Rest Catalog 419 Response Code

2025-02-19 Thread Sung Yun
Hi folks, While working with the Polaris community on an issue[1], I decided to bring this discussion to the Iceberg community mailing list as I believe that the status code 419 in the REST Catalog Open API Spec may have become a source of confusion for the community. In the Iceberg REST Catalog