Script-class-aware Inspector & related controls.

[willnationsdev]

Updates EditorResourcePicker so that global script classes are properly recognized and handled.

Related Issues:

• Add first-class custom resource support godot-proposals#18

Related PRs:

• Allow exporting custom resources from/to GDScript, VisualScript, C#, and PluginScript. #48201
  • implements the Editor / Inspector tooling

Sibling PRs:

• Add GDScript resource export. #62411
• Enable QuickOpen to see scripted resources #62417
• Add C# resource export. #63126

Considerations:

• None. Everything should work fine as far as I know. Have tested with regular properties, array properties, make unique, external resources, embedded resources, etc.

[mhilbrunner]

PR meeting note: we really want this to be in Godot 4.0, but @vnen needs to have a look.

[willnationsdev]

@mhilbrunner Sure thing. I'm rebasing all the various PRs atm to make sure they are all up-to-date.

[reduz]

I will approve this for now, but the way we are obtaining the class information in EditorData is very hacky. We really need to do a through discussion and cleanup of all the EditorNode structures for 4.1.

[YuriSizov]

Seems nice otherwise. Definitely shouldn't make things worse, and @akien-mga has reported it working well together with #62411 is his limited testing.

[willnationsdev]

I've made the requested adjustments. Should be good to go for merge at this point afaik.

[akien-mga]

Tested, works well.

The code changes also seem sensible. As reduz pointed out the lack of proper registration of global script classes makes this all a bit awkward but that's a pre-existing design issue that should be solved in a future release.

[akien-mga]

Thanks!

[YuriSizov]

Do you mean to say that I should explicitly use a break inside of an if statement rather than making the any_type_matches variable part of the loop condition, as I have done?

@willnationsdev <...> There was a long reply here explaining what I'd meant, but it seems I was a bit confused and mistaken. Now I'm seeing what you've done originally. 😅

It was fine, execution-wise, but yeah, I think it's better to do this check explicitly in the body. Clearer that way. However, when you've refactored your code after my request, I think you've made a mistake. Your checks are if (!any_type_matches) { break; } now, when they have to be the opposite (and same for paste_valid) and break if there are matches.

[willnationsdev]

Oh, crap...lol. Guess I'll need to submit a follow-up PR.