Use namespaces instead of underscore prefix for core bind classes

[neikeq]

I propose this as a replacement for the underscore thing.
This PR doesn't remove the existing extra code that is needed to support classes with underscores (#26922 lists most of them).

The namespace names are just placeholders. I don't know what's our naming convention for namespaces.
There is a lot of ::ClassDB::bind_method lines in core_bind.cpp, which could be avoided if we move core_bind::ClassDB to its own namespace so it doesn't bother the other classes. Edit: Moved to its own namespace.

[fire]

Are there any problems from the c-api end? Aka gdnative?

[neikeq]

@fire I don't think there is any, but I didn't check. Most if not all the code that works around these underscores is just to strip it (or add it in case of a lookup).

[reduz]

I am surprised, never expected anything like this to work, if it really does feel free to go ahead and merge.

[neikeq]

It may be safer to leave this for 4.0. I don't think it breaks anything, but I don't know what some GDNative language bindings out there are doing, so just to be safe...

[neikeq]

The only annoyance I found is that we can't use VARIANT_ENUM_CAST because the template specializations it declares must be in the same namespace where the template was originally declared (which is why I moved all the VARIANT_ENUM_CAST to the bottom outside the namespace in core_bind.h).

[aaronfranke]

Does this change anything on the front end?

[ptrojahn]

I created #30153 which only removes the checks.

[vnen]

Any news in this regard? I'm now rewriting GDScript code and it would be nice to not have to deal with those pesky underscores.

[neikeq]

I'll be revisiting this soon, but right now I'm focusing on other tasks.

[Xrayez]

Inspired by this PR, I've already done something similar while developing a custom module. But stumbled upon another bug while regenerating documentation and then generating the mono glue, see Xrayez/godot-geomtools#3.

[neikeq]

@Xrayez The issue is with your PolyDecompType enum. Enums in namespaces needs the changes I made to type info in this PR.

[Xrayez]

Oh ok, so it would be nice for these changes to be cherry-picked to 3.2, else I'd have to make similar changes and/or revert to how it's now.

[Xrayez]

Just saying that I have some other module with having enums in namespaces which is an issue in and of itself (which is worth treating via separate PR perhaps which can be cherry-picked), for instance see: Xrayez/godot-anl@642b470#diff-142e17f094b421b5777b79be43c94633R336-R339. I've just had to duplicate enums eventually.

[aaronfranke]

@neikeq This was last rebased over a year ago, what's the status of this?

[neikeq]

It should be easy to re-do it for the master branch, but I haven't had time for that yet. Maybe next month.

[neikeq]

I've added this to my TODO list for July.

[akien-mga]

Superseded by #51627.