Fix add_root_node() being no-op

[KoBeWi]

The method was non-functional since e9bbb97, for no reason.

[ryevdokimov]

I'm not sure it's a great idea to start revitalizing the EditorScript class since it's borderline entirely deprecated by other classes. This functionality seems like it would fit better as part of EditorInterface.

[ryevdokimov]

After adding the root node, the tab for the scene will still show as [empty].

[KoBeWi]

After adding the root node, the tab for the scene will still show as [empty].

Fixed.

I'm not sure it's a great idea to start revitalizing the EditorScript class since it's borderline entirely deprecated by other classes.

The base functionality of this class is running utility scripts in the editor, which is still relevant. Methods other than _run() are mostly extras that we can't remove for compatibility reasons.

[ryevdokimov]

The base functionality of this class is running utility scripts in the editor, which is still relevant. Methods other than _run() are mostly extras that we can't remove for compatibility reasons.

Would this method still be under the purview of compatibility after not having worked for nearly a decade haha. It has been marked unimplemented since 3.0. I can see a rational for the answer being yes, so I won't press the matter.

I still think it makes sense at least for a similar function to be implemented in EditorInterface, and this one to be marked as deprecated (along with get_scene(), since get_tree().current_scene does the same thing).

[ryevdokimov]

When using this method with scenes, would it make sense for this method to add an inherited instance of a scene instead of just opening a scene? IMO, having it open the scene, seems wrong and can lead to confusion, and accidentally modifying the original scene, especially when you have other options like open_scene_from_path()

vs

I could see this as being supplementary to: #90057, but instead of simultaneously opening a new tab and creating a new inherited scene, it allows you to build plugins to do stuff like this after pressing 'Add new scene'

[KoBeWi]

This PR is supposed to make the method functional again the way it was designed, any additional changes can be done later.

When using this method with scenes, would it make sense for this method to add an inherited instance of a scene instead of just opening a scene?

It does not open a scene, you need to be already in empty scene to use it.

[ryevdokimov]

This PR is supposed to make the method functional again the way it was designed, any additional changes can be done later.

Functional to what standard? The way the engine is used is pretty different from version 1.1 (the actual last time this method worked) to version 4. We'd have to understand first why the method was commented out in the first place to have full understanding of its intent. I would argue at this point you might as well treat it as if you are adding a new function, and its use-cases should be examined as such.

It does not open a scene, you need to be already in empty scene to use it.

The empty scene gets replaced with the loaded scene, as if you opened it. If you begin to make modifications, you are now modifying the original scene. This is typically not what people want when adding scenes as nodes into the tree and could lead to issues.

[KoBeWi]

I made it create inherited scene when adding an instance.

[ryevdokimov]

The tab should probably be marked as unsaved similar to as if you were creating a new scene through the GUI.

[ryevdokimov]

Works great now.

Along with #90057 this will fully implement:

godotengine/godot-proposals#3907, giving users the option to create a new inherited scene either over an empty scene or straight into a new tab.

[akien-mga]

Thanks!