Ah, the pitfalls of UTF. Did not know that and don't like it at all, I used \d as shorthand for 0-9 a lot. My old camel book still says "\d - A digit, same as [0-9]"

I think I would have prefered a new character for this class to a break of backwards compatibility. The other solution that the internal number conversion also recognizes thai and other numbers would be even better theoretically, but that may not be easy with UTF