Yeah sure, one probably could. For this proposal specifically, I think anything Unicode-related is well outside the scope.

Even having support for digit separators is somewhat fishy, but it's possible to justify by saying "at least one should be able to use this function to parse C++ floating-point literals in C++."