Merge pull request #45957 from lupoDharkael/natural-comp

[3.2] Add natural string comparison

doc/classes/String.xml

@@ -621,6 +621,19 @@
 				Returns the MD5 hash of the string as a string.
 			</description>
 		</method>
+		<method name="naturalnocasecmp_to">
+			<return type="int">
+			</return>
+			<argument index="0" name="to" type="String">
+			</argument>
+			<description>
+				Performs a case-insensitive [i]natural order[/i] comparison to another string. Returns [code]-1[/code] if less than, [code]1[/code] if greater than, or [code]0[/code] if equal. "less than" or "greater than" are determined by the [url=https://en.wikipedia.org/wiki/List_of_Unicode_characters]Unicode code points[/code] of each string, which roughly matches the alphabetical order. Internally, lowercase characters will be converted to uppercase during the comparison.
+				When used for sorting, natural order comparison will order suites of numbers as expected by most people. If you sort the numbers from 1 to 10 using natural order, you will get [code][1, 2, 3, ...][/code] instead of [code][1, 10, 2, 3, ...][/code].
+				[b]Behavior with different string lengths:[/b] Returns [code]1[/code] if the "base" string is longer than the [code]to[/code] string or [code]-1[/code] if the "base" string is shorter than the [code]to[/code] string. Keep in mind this length is determined by the number of Unicode codepoints, [i]not[/i] the actual visible characters.
+				[b]Behavior with empty strings:[/b] Returns [code]-1[/code] if the "base" string is empty, [code]1[/code] if the [code]to[/code] string is empty or [code]0[/code] if both strings are empty.
+				To get a boolean result from a string comparison, use the [code]==[/code] operator instead. See also [method nocasecmp_to] and [method casecmp_to].
+			</description>
+		</method>
 		<method name="nocasecmp_to">
 			<return type="int">
 			</return>