Elevate Control.get_rect to CanvasItem.get_rect

[HubbleCommand]

As the title says, I'd like to open the discussion to elevating the get_rect function of Control to CanvasItem, so that nodes inheriting from Node2D could implement it as well.

Some Node2Ds like Sprite2D do have a public method for it, and other nodes like Polygon2D and Line2D keep this information internally for editor use.

While in many cases it can be better to work with the Physics server, that can also be excessive. For example, I wanted to get the Rect of a Polygon2D for some basic overlap checks. Would have been overkill to create the resource in the physics server etc. just for a rare check. While it is very easy to determine the Rect in many cases, it would be nice to have it directly from the engine. Even the example for Sprite2D to just determine clicks seems like a handy enough reason to me, albeit not particularly important.

I'm also interested in the reason get_rect wasn't made available in Node2D, if any. I looked through some discussions, issues, and PRs, and didn't find anything.

This is something I would be willing to work on if it's desired.

[AThousandShips]

Not every CanvasItem has a well defined rect, just like Node3D doesn't expose a get_aabb but VisualInstance3D does, however every Control is supposed to support an area

[HubbleCommand]

A Rect is just a position and a size, even if the size is zero for some nodes by their nature (things like Marker, RemoteTransform, or AudioListener). A sizable chunk of Node2Ds do have a size.

I mean, I guess the same thing can be said about Node3D. In either case, I would intuitively assume something that exists in 2D or 3D space to have some size, it feels like the exception that it wouldn't. It's not as abstracted or incorporeal as Node is.

Can I assume that's the reason then? While many do take space, it's that it's not a guarantee?

[AThousandShips]

So then you agree that many Node2Ds don't have a valid Rect2 distinct from just their position? Just like many Node3Ds doesn't have a valid AABB distinct from their position? Unlike Control nodes which all can be said to be able to have a well defined Rect2?

[HubbleCommand]

Size zero is still valid. We're not talking negative size voodoo.

As long as I know the reason, it's good for me.