This ticket is for considering pros and cons of different approaches to custom QML components with Qt Graphs (2D). Few places where these are already used:

• QXYSeries::pointMarker
• QBarSeries::barComponent
• QAbstractAxis::labelComponent (tbd)

Few ways how these could be implemented:

1) Utilize QML properties directly. Basically we could say that QAbstractAxis::labelComponent must be Text{} element, so AxisRenderer can fill the value directly.

Pros: Minimal amount of supporting code or magic properties.
Cons: Has limitations. For example what if user wants to add margin into Text? Or label to be a Button instead of Text?

2) Special data class. Example of this is BarComponent, so user is expected to use "barComponent: BarComponent { id: comp; ... }" and gets access the all required data with comp id.

Pros: Easy to document what data is available and add more.
Cons: Special QML elements need to be used.

3) Custom dynamic properties. Example of this is QXYSeries::pointMarker. User can use any QML component there and just add "property bool selected: false" etc. there. PointRenderer then checks if these properties exist with e.g. "if (marker->property("selected").isValid())"

Pros: No need to use specific data element.
Cons: How are these special properties documented? Currently QXYSeries::pointMarker doesn't document what properties it supports. Is there performance overhead with dynamic properties & checking them?