Problem/Motivation

Currently our entry point is a JSON API -ish response to the /jsonapi endpoint. However, it could be worth adopting https://tools.ietf.org/html/draft-nottingham-json-home-06 for it and stop pretending that it's valid JSON API.

It's worth noting that moving to this new solution will break BC.

Proposed resolution

Discuss the value of this, given that we need to break BC to implement it.

Comments

e0ipso created an issue. See original summary.

Wim Leers’s picture

Ohhh!

      A "resources" member, whose value is an object that describes the
      resources associated with the API.  Its member names are link
      relation types (as defined by [RFC5988]), and their values are
      Resource Objects (Section 4).

This sounds very sensible :)

This also taught me about the existence of https://tools.ietf.org/html/rfc7240, which I'd never seen before.

dawehner’s picture

Currently our entry point is a JSON API -ish response to the /jsonapi endpoint.

Do you believe people use that for their actual sites? At least for me this sounds like something you do during development time, which could be argued not any BC break.

dubcanada’s picture

So the proposal is to get rid of /jsonapi/ and respond to any requests containing the application/json-home Content-Type?

arifkhn46’s picture

It think it should be configurable.

e0ipso’s picture

JSON API is zero configuration. I think it should remain so.

The plan is to replace only the entry point (/jsonapi). Every thing else remains the same.

e0ipso’s picture

Priority: Normal » Minor
dawehner’s picture

JSON API is zero configuration. I think it should remain so.

I love this idea, ALSO because this makes it possible to not have this conversation in core and maybe keep jsonapi_extra in contrib land.

e0ipso’s picture

Status: Active » Fixed

This was done at some point.

Status: Fixed » Closed (fixed)

Automatically closed - issue fixed for 2 weeks with no activity.