Keeping plone.restapi and plone server API in sync (was: Doubt about schema / content type serialization)

[bloodbare]

Why its serialized the content types and schema based on the fieldsets and not on the schema. I mean , the expected json representation is linked to the fielsets , and would make more sense to base on schemas and its invalidation and requiremenets.

@tisto

[tisto]

@bloodbare can you please provide an example of what you mean? I'm not 100% sure I understand everything your wrote. :)

[bloodbare]

So the problem is on schema serialization, that we are not showing any schema information. We have like a dual layer of grouping properies, fieldsets and schemas (Behaviors for example).

Then there is another time the same problem I already pointed that having two separate json serializers (plone.jsonserializer and plone.restapi) makes really hard to maintain and plone.jsonserializer was rendering the fields on the fieldset not on the properties... reading actual plone.restapi it looks like is different now and we are unsync. We may need to end having two different API's.

[tisto]

This issue is about two things:

1. A common API spec for plone server and plone.restapi
2. A common serializer for plone server and plone.restapi

plone server uses its own API specification format. We should take another look at swagger etc. to see if it is possible to support traversal with those standard tools. The latest discussion around making connexion framework agnostic also looks interesting: spec-first/connexion#394

cc @bloodbare