Improve area/body_shape_entered/exited signals parameter names and doc

[MaxLap]

This is improving the documentation (and, I believe, the signal templates) for these signals:

• area_shape_entered
• area_shape_exited
• body_shape_entered
• body_shape_exited

Here are the improvements, I made a separate commit per improvement:

• Some doc used body_id or area_id instead of the _rid that is written elsewhere
• I added a _index suffix to the body_shape, area_shape and local_shape parameters. This has 2 benefits:
  • More intuitive when reading the name
  • Allows to then use var body_shape: Node = body.shape_owner_get_owner(body_shape_index). Before, body_shape would have been taken by the int parameter.
• Added code to get the actual node from those _index arguments.
  It took me a while to find it when I needed it, and when I did, I wasn't sure if that was valid for more recent Godot versions until I tried it.

All of those changes would also be great improvements for Godot 3.3 documentation (and maybe farther back, but I am not familiar with older versions)

[mhilbrunner]

Hey! Thanks for contributing.

1. Can you squash the commits into one, please? :)
2. The commit message and PR title should clarify that this also changes the signal/naming, not only the documentation.

[MaxLap]

Sure, squashing is done.

[mhilbrunner]

Thanks for the squash and fixup!

All of those changes would also be great improvements for Godot 3.3 documentation (and maybe farther back, but I am not familiar with older versions)

It is IMO unlikely to be backported to 3.x as all these changes break compatability by changing method argument names, which we generally try to avoid in minor releases AFAIK.

(This PR is somewhat relevant to #16863)

[MaxLap]

Alright, thanks for the feedback.

I'm surprised that the name of arguments is considered to be part of the signature. Languages usually only work by position, unless they have named arguments. I mean, I can change the parameter name of a signal's function and everything seems to still work.

Anwyays, it's very minimal so probably not worth the time for the possible merge conflicts

[mhilbrunner]

No, thats actually a good point - I spent too much time with Python recently, where this is more of a concern. :)

I'll add a backport request label for now optimistically, let's see if someone disagrees :P

[pouleyKetchoupp]

Looks great, thanks!

Just left a comment for a minor improvement in the documentation, otherwise it should be good to go.

[pouleyKetchoupp]

Also, some class names have changed on master, so we won't be able to cherry-pick the commit directly. Feel free to open a separate PR with the equivalent changes on the 3.x branch.

[MaxLap]

Added the change and squashed everything together

[pouleyKetchoupp]

@MaxLap Thanks! My comment was probably not clear enough, but I meant to suggest the same change for all the occurences of actual node with either [CollisionShape2D] or [CollisionShape3D].

[MaxLap]

Lol, my bad. Your comment was perfectly fine and I understood it, but I forgot in 5 seconds. Not my proudest moment xD

Done. Changed all "actual node" this time :)

[pouleyKetchoupp]

No worries :) I've added comments on a few last things to fix before merging.

[MaxLap]

Good catches. Fixed the 3D and the missing "node".

[pouleyKetchoupp]

Looks all good!

[pouleyKetchoupp]

Congrats on your first merged PR!

Please let me know if you're up for doing the 3.x version, otherwise I can take care of it.

[MaxLap]

I can do it tomorrow. Is there any diff other than CollisionShape3D being just CollisionShape?

[pouleyKetchoupp]

Thanks! Some file names are also different, but I think that should be it.

[MaxLap]

Here is the PR for 3.x: #53848