Regression: Unicode superscript no longer works

[pko-trackunit]

Hi,

It seems that Unicode superscript is no longer treated as a power of 2, 3, etc.

import tech.units.indriya.format.SimpleUnitFormat;
import tech.units.indriya.unit.Units;
import javax.measure.MetricPrefix;

public class IndriyaSuperscriptBug {
    public static void main(String[] args) throws Exception {
        var fmt = SimpleUnitFormat.getInstance();

        // Caret notation works correctly
        var mm2_caret = fmt.parse("mm^2");
        var cm2_caret = fmt.parse("cm^2");
        System.out.println("100 mm^2 -> cm^2 = " + mm2_caret.getConverterToAny(cm2_caret).convert(100));

        // Unicode superscript is broken — prefix applied linearly instead of squared
        var mm2_super = fmt.parse("mm²");
        var cm2_super = fmt.parse("cm²");
        System.out.println("100 mm² -> cm² = " + mm2_super.getConverterToAny(cm2_super).convert(100));

        var mm2_correct = MetricPrefix.MILLI(Units.METRE).pow(2);
        System.out.println("parse(\"mm²\").equals(MILLI(m).pow(2)) = " + mm2_super.equals(mm2_correct));
    }
}

indriya 2.2.2 outputs:

100 mm^2 -> cm^2 = 1.0
100 mm² -> cm² = 1.0
parse("mm²").equals(MILLI(m).pow(2)) = true

indriya 2.2.3 outputs:

100 mm^2 -> cm^2 = 1.0
100 mm² -> cm² = 10.0
parse("mm²").equals(MILLI(m).pow(2)) = false

Is it expected? I checked the changelog, but didn't see anything obvious that could cause this. And I would hope a patch release would not have such breaking changes.