Invalid set index error

Nanites

Hi,

I have 3 nodes:

A (Node2D)
	->B (Node2D)
		->C (Area2D)

C has this script:

extends Area2D

var parent 
var parentsparent

A has this:

func _ready():
	get_node("B/C").parent = get_node("B")
	get_node("B/C").parentparent = self

this gives the "Invalid set index 'parent' (on base: 'Area2D') with value of type 'Node2D (test_b.gd)'. From my understanding this should work? C gets loaded before B and that before A, so when A's _ready() runs it should be able to access C's variables?

The purpose of that btw is to test if saving the reference is (noteworthy) faster then using get_parent()

Rukus

I think you have it backward. It works down the hierarchy A enters the tree and waits and add childeren once C is ready, at the end, the root A is ready, Put scripts in base nodes and refrence children. I do: If a node adds child node (probably refrenced from somwhere else) I add to it's variable root_node=self so there is no confusion the if I do get_parent in node C in your case it would be...B work from the base of the hierarchy not the middle

Rukus

Btw faster in code? Or for you to type?

Nanites

@Rukus said: A enters the tree and waits and add childeren once C is ready, at the end, the root A is ready, Put scripts in base nodes and refrence children.

But isn't that what I do? C is a child of B, which is a child of A. When A is ready, so should be C, right?

@Rukus said: Btw faster in code? Or for you to type?

Faster to execute. I have a lot of calls to different nodes and trying to optimize it, for example I figured out for item in get_tree().get_nodes_in_group is way faster to access them then for item in node.get_children()

newmodels

I see your problem. You have a typo.

get_node("B/C").parentparent = self should be get_node("B/C").parentsparent = self

cybereality

There might be some difference in speed, but it's probably negligible since you are only doing it once on ready. In any case, get_parent() should intuitively be faster since that is C++ code and not GDScript. Feel free to test, as sometimes counter-intuitive stuff works, but it seems like you might be optimizing the wrong thing.

Nanites

Thanks!

I corrected the typo but it didn't changed the error (it was on the previous "parent" node).

Even when get_parent is faster, I'd like to understand why this code is not working - for some reason it can't access the parent variables (I checked if this has something to do what is assigned to them, but = 1 gives the same error just with "type int"). Shouldn't this just work?

Nanites

Fixed it, the problem was that I attached the script to C scene itself, while the C that was A's child was made local. Don't code at 2am ;)

But the takeaway from this experiment is, @cybereality doing it this way is significantly faster then calling get_parent(). Now this will only be of any significance for the niche cases where you call get_parent a lot, but every optimization counts I think.