Adding support for RF6902 for path requests.

[rhysrevans3]

Related Issue(s):

• Is JSON Merge PATCH RFC7396 compatible with STAC? #13

Proposed Changes:

1. Adding support for RFC 6902 to PATCH.
2. Updating RFC 7386 support to RFC 7396 in PATCH.

PR Checklist:

• This PR has no breaking changes.
• I have added my changes to the CHANGELOG or a CHANGELOG entry is not required.

[rhysrevans3]

I've added the original application/json content type back as to not cause any breaking changes. This is not technically correct for RFC 7396.

[vincentsarago]

@m-mohr would you be wiling to review this?

[m-mohr]

This primarily needs to be in accordance with OGC APIs corresponding part. If we have an issue in STAC, OGC may also have one. Can someone check with them? I'd probably be fine with this if OGC also adopts this. I'd try to avoid diverging from OGC.

I'm traveling/on vacation the next 4 weeks so would only be able to review in November, I think.

[rhysrevans3]

@m-mohr I've created an issue in the OGC Features repo and had confirmation that OGC is not exclusive to RFC 7396 and that the addition of RFC 6902 won't diverge from OGC. They haven't said they will add RFC 6902 but instead will add a disclaimer that JSON Merge Patch is not exclusive. Do you think a similar disclaimer should be added to STAC's PATCH?

[m-mohr]

I think we should explicitly and clearly say that the "old" way is the default, but that the "new" way can be added in addition and be distinguished via content negotiation (media types). That's also how I understood Peter in the OGC issue.

[rhysrevans3]

I've updated the description to state that merge-patch is the default and json-patch is optional. Merge-patch is also the method for base application/json content type with merge-patch requiring application/json-patch+json. Is there anywhere else that this would need to be added?

[m-mohr]

Thanks, good start. I think it should also be clearly started in the README.md

[rhysrevans3]

Good point I've update the PATCH text to RFC 6902 is optional and additional to RFC 7396.

[m-mohr]

Looks good to me (from a theoretical perspective).

Would like to see another review from a practitioner, server and/or client implementor.