5.1.1.2. Floating Literals

Float literals follow ISO C99 except we allow underscores between digits. The allowed suffixes are:
Felix namesuffix
floatf, F
doubled,D,<none>
ldoublel, L

The mantissa may be give in either decimal or hexadecimal, the latter requires a prefix "0x" or "0X". The exponent may be given as a decimal power using code "e" or "E" or binary power using "p" or "P", the radix is always decimal.

Note that float literals are distinguished from integers by the presence of a decimal point. The decimal point may not be elided, even if an exponent is given. A digit must appear either to the left or right of the decimal point.