feat: primitive children prop mapping

[dpilch]

Existing Text prop value functionality is preserved for now. All existing Studio apps need to update Text property value to label, then Text prop value functionality can be removed.

label was chosen because it does not have any conflicts with existing primitive props.

• Create a mapping from a "synthetic" prop to children for primitives that will likely have scalars as children.
• change Primitives enum to Primitive to follow enum style

See packages/studio-ui-codegen-react/lib/react-component-with-children-renderer.ts for mapping logic.

[alharris-at]

Changes look good to me, but I'd want to confirm that we've never seen that 'value' prop used in a text before we release.

[alharris-at]

I'm a bit concerned about this - as far as I understood it these complexTests came from a real export from studio ui - if they're publishing the value then we shouldn't be changing this behavior.

Could you confirm with @frimfram if this has been modified at all from the raw output of CLI?

[frimfram]

I would suggest that whenever we are "removing" something in a PR, that PR should contain new tests with "realistic" JSONs that get generated from Studio UI Figma import. That would make sure that we don't just work with assumptions and/or fake jsons that were created outside of Studio UI.

[dpilch]

This would be a breaking change I believe the current UI is not compatible with this. In my opinion this is a change that we should make before re:invent because we won't be able to make breaking changes easily after re:invent. I will check with the UI team to get consensus on the functionality.

[dpilch]

To add an example:

Text and Button would have different approaches to put text as a child.

Text:

{
  "componentType": "Text",
  "properties": {
    "value": {
      "value": "text"
    }
  }
}

<Text>text</Text>

Button:

{
  "componentType": "Button",
  "properties": {
    "children": {
      "value": "text"
    }
  }
}

<Button children="text"></Button>

[alharris-at]

Yeah, definitely agreed that the change makes sense, and harrison and chris have verbally confirmed that the children thing is being used rather than value, but we should just confirm w/ Rene that we're okay with this change, since he brought it up during backlog refinement yesterday.

[dpilch]

Sounds good, waiting on response from René.

[dpilch]

I've updated this PR with WIP primitive children prop mapping.

[dpilch]

Got the ok on this list from René.

[dpilch]

Looking into the failed integ test now.