StringName comparisons give wrong and inconsistent results

[lilizoey]

Godot version

4.0.2

System information

Linux 5.10.167-2-MANJARO

Issue description

Comparing StringNames will give wrong and inconsistent results, running the provided code will give the results:

a < b false
a > b false
a <= b true
a >= b true
a == b false

The documentation for < on StringName says:

Returns true if the left String comes before right in Unicode order, which roughly matches the alphabetical order. Useful for sorting.

However then "a < b" should be true, not false.

In addition a <= b and b <= a is true, yet a == b is false. Which is inconsistent.

Steps to reproduce

run

func _ready():
	var a = &"a"
	var b = &"b"
	print("a < b ", a < b)
	print("a > b ", a > b)
	print("a <= b ", a <= b)
	print("a >= b ", a >= b)
	print("a == b ", a == b)

Minimal reproduction project

N/A

[AThousandShips]

The documentation is apparently wrong from the implementation, should be updated

[KoBeWi]

The operator is implemented like this

godot/core/string/string_name.h

Lines 117 to 119 in 06f5b09

	_FORCE_INLINE_ bool operator<(const StringName &p_name) const {
	return _data < p_name._data;
	}

_data is a pointer. It's wrongly implemented (unless there is a reason why it's like that).

[AThousandShips]

Internally StringNames are compared by memory to keep it fast it seems, when stored in RBMap, the fact that the documentation talks about String and not StringName would imply to me that it was copied from String without considering the actual implementation

[AThousandShips]

I assume StringName was/are used in RBMap etc, sorting the pointers should give uniqueness and is faster than comparing by the content, seems to be overlooked

[AThousandShips]

In fact the wording for String and StringName are equal for the ordering operators, including calling it String in StringName, however the equality operator differs, and makes a note about how it is faster:

Returns true if the StringName and right refer to the same name. Comparisons between StringNames are much faster than regular String comparisons.

[AThousandShips]

The implementation is weird though as it doesn't give correct results, also there isn't any other comparison operators for StringName than < so might be doing conversion under the hood somewhere

[AThousandShips]

For now if you want the alphabetic ordering a workaround should be to cast them to String

I suspect there's some conversion going on as there's no errors caused by the fact that StringName doesn't have any of the remaining operators, and a decision would have to be made on what ordering to use, alphabetical or pointer based

[AThousandShips]

Working on a fix for this