Framework: add schema for sharing/publicize

[gwwar]

This PR adds a schema for sharing/publicize so we can avoid data shape changes as described in #3101.

Thanks @aduth's refactors in #3417, we only need to persist connections. Other transient data like fetchingConnections will never be persisted.

Testing Instructions

• In the console set localStorage.debug to calypso:state
• Make sure that your blog has some connections ( Connected to Facebook/Twitter/etc ).
• Navigate to the Editor
• Refresh the page.
• The app behaves normally and loads the persisted state/current-user from IndexedDB storage

To Validate that we can't get into an inconsistent state:

• Add a non-existent required property to connectionsSchema, like foo:

export const connectionsSchema = {
    type: 'object',
    patternProperties: {
        '^\\d+$': {
            type: 'object',
            required: [ 'ID', 'site_ID', 'foo' ], //add a non-existent prop to test invalidations
            properties: {
                ID: { type: 'integer' },
                site_ID: { type: 'integer' },
                //...
            }
        }
    },
    additionalProperties: false
};

• Navigate to the Editor
• Refresh the page
• You should see a console warning that validation failed
• The Editor should load.

cc @aduth @mtias @rralian @blowery

[aduth]

I'm assuming line length is what drove you to using ternary in this case and condition blocks in the others, though from a legibility perspective, I think we should be consistent (on the block?).

[gwwar]

👍 We can stick with blocks. Long ternaries tend to be more error prone.

[gwwar]

Updated in ba8b593

[blowery]

Perhaps connectionsBySiteId should be a selector instead of stored data?

[gwwar]

Perhaps connectionsBySiteId should be a selector instead of stored data?

Interesting, I'll play around with that. I would like to avoid a larger schema object if possible.

[aduth]

Perhaps connectionsBySiteId should be a selector instead of stored data?

Ah yeah, I knew there was something about this subtree I meant to bring in line with our recommendations. We should definitely make this change.

@gwwar : Do you think it should be made separately?

[gwwar]

🙇 Thanks @aduth I'll update this PR after #3417 lands

[gwwar]

cc @aduth @rralian @blowery almost forgot about this one. It's ready for another look if you have time.

[aduth]

Getting a schema validation warning here. This value is null in my case.

[gwwar]

Thanks @aduth I'll take another look

[gwwar]

@aduth I've updated the schema to allow null types for all external properties.

[aduth]

Looks good 👍

(Side-note: I'm really excited for state persistence to gain more adoption / become enabled in production 😄 )

[gwwar]

Thanks for the review! 😄 Devs should be a bit more aware of it when #3853 enables it in stage. I'll continue to chip away at these schema PRs in my spare time.