There is actually one edge parsing case that Text::xSV handles differently than Text::CSV_XS where I think that Text::xSV's behaviour is more convenient.

Aha, I missed that. and I agree, Text::xSV's is more convenient. I've added that to the chart.