parsing issue for hexadecimal

[rschumi0]

I discovered a parsing issue for the hexadecimal representation of floats or doubles. When a literal is starting with 0x1.a ... 0x1.f then there is an error, but they get correctly parsed when they start with 0x1.0 ... 0x1.9.

float a = 0x1.c4c52cp-1f; // produces an error, 0x1.1c52cp-1f works

(And Is there a reason why hex and octal literals are supported, but not binary ones?)