Inconsistency: "1a".to_snake_case() == "1a", but "1aa".to_snake_case() == "1_aa".

[nathanfranke]

Tested versions

Reproducible in:
3.5.3.stable using capitalize
4.2.1.stable using capitalize and to_snake_case.

System information

Linux on kernel 6.6.7-arch1-1

Issue description

extends Node

func _ready():
	print("1a".capitalize())
	print("1aa".capitalize())
	print("1a".to_snake_case())
	print("1aa".to_snake_case())

Output:

1a
1 Aa
1a
1_aa

Expected:

1 A
1 Aa
1_a
1_aa

Steps to reproduce

https://gd.tumeo.space/?KYDwLsB2AmDOAEA5A9tYAodAzArpAxvAPoBOwAhtAJ4AUAlAFzoCQADiQJaRg0BEAjOV4A6fOVYcw5ADYcAXsHp0W7LjwHkho8ZJnzFdZW07c+gkWGRFYkcgGtgRMbANHVpjVsvXbDp+RclIA

Minimal reproduction project (MRP)

to_snake_case_4.2.1.zip

[dalexeev]

This is made to handle "2D"/"3D" and "str2var", "var2bytes", etc. (replaced by _to_). See test cases.

godot/core/string/ustring.cpp

Lines 1060 to 1063 in 07b8860

	const bool cond_a = is_prev_lower && is_curr_upper; // aA
	const bool cond_b = (is_prev_upper || is_prev_digit) && is_curr_upper && is_next_lower; // AAa, 2Aa
	const bool cond_c = is_prev_digit && is_curr_lower && is_next_lower; // 2aa
	const bool cond_d = (is_prev_upper || is_prev_lower) && is_curr_digit; // A2, a2

godot/tests/core/string/test_string.h

Lines 1312 to 1337 in 07b8860

	TEST_CASE("[String] Checking case conversion methods") {
	StringCasesTestCase test_cases[] = {
	/* clang-format off */
	{ "2D", "2d", "2d", "2d" },
	{ "2d", "2d", "2d", "2d" },
	{ "2db", "2Db", "2Db", "2_db" },
	{ "Vector3", "vector3", "Vector3", "vector_3" },
	{ "sha256", "sha256", "Sha256", "sha_256" },
	{ "Node2D", "node2d", "Node2d", "node_2d" },
	{ "RichTextLabel", "richTextLabel", "RichTextLabel", "rich_text_label" },
	{ "HTML5", "html5", "Html5", "html_5" },
	{ "Node2DPosition", "node2dPosition", "Node2dPosition", "node_2d_position" },
	{ "Number2Digits", "number2Digits", "Number2Digits", "number_2_digits" },
	{ "get_property_list", "getPropertyList", "GetPropertyList", "get_property_list" },
	{ "get_camera_2d", "getCamera2d", "GetCamera2d", "get_camera_2d" },
	{ "_physics_process", "physicsProcess", "PhysicsProcess", "_physics_process" },
	{ "bytes2var", "bytes2Var", "Bytes2Var", "bytes_2_var" },
	{ "linear2db", "linear2Db", "Linear2Db", "linear_2_db" },
	{ "sha256sum", "sha256Sum", "Sha256Sum", "sha_256_sum" },
	{ "camelCase", "camelCase", "CamelCase", "camel_case" },
	{ "PascalCase", "pascalCase", "PascalCase", "pascal_case" },
	{ "snake_case", "snakeCase", "SnakeCase", "snake_case" },
	{ "Test TEST test", "testTestTest", "TestTestTest", "test_test_test" },
	{ nullptr, nullptr, nullptr, nullptr },
	/* clang-format on */
	};

[Mickeon]

A few small examples for the casing may do the trick, like in capitalize()

[Mickeon]

I updated #86639 with some examples for this, as well. Feel free to suggest examples if necessary.