Widget state persistance

[thecalcc]

SDBELGA-847

TODO:

• Look into possible different solutions for closedThroughAction
• Add comment for the functionality
• Look into state getting reset, when:
  a) widget is pinned and you click to close it
  b) widget is pinned or not, you switch back and forth between views, and switch widgets
• Final check against develop
• Update UIF so disabled take effect (version)
• Add a playwright test
• Verify persistence works across all widgets in React Authoring

What this PR aims to address: When a user switches back and forth between full view width and normal width, if there's a pinned widget that has state changes, the changes should persist between the two view modes. In this PR that's being handled using a HOC (WidgetStatePersistanceHOC) and a variable (closedThroughAction). The HOC handles the logic for state persistence using widgetState variable and ref (to get child component state - current widget). Inside authoring when a widget gets loaded we handle the initialState of the widget using closedThroughAction variable - If a widget has been closed by the user on purpose, once the same widget is reopened, state does get reset. If that's not the case (switching between the two article view modes) we count that as the widget not being closed through a direct user interaction, thus retrieving the old state of the widget and loading it with the widget rendering.

Side issues this PR fixes: Widget pinned state in the different scenarios, and widget persistence using preferencesService.

Additional solutions: Another solution for setTimeout for the initialState in authoring-react - Add a variable renderCount and based on the count delete the localStorage value, then reset renderCount. Problems with that - adds more complexity, if logic changes we'd need to adjust that on top of everything.

[tomaskikutis]

• Add a playwright test
• Add documentation as a comment somewhere describing the issue and your solution

[tomaskikutis]

take a look at 2 remaining unresolved comments

[thecalcc]

"Versions & History" and "Comments" widget are not the usual implementations of widgets.

Nested components with multiple levels of state management, and more. I'd say time wasted is not worth it now, we can make a ticket to check it when working on authoring-react.

[thecalcc]

Lint failing because UIF isn't yet updated

[tomaskikutis]

Since the object is only neccessary for mutation to work, it would be better to call the object property "value", because the real name is already used in the constant.

I don't think closedOnRender is a good name either. What does it have to do with rendering? closedIntentionally sounds better. Also comment on how it may get closed unitentionally.

[tomaskikutis]

Also comment on how it may get closed unitentionally.