trivial-float-bits

You can use trivial-float-bits to convert between a float and its representation as an integer using implementation-specific functions.

(single-float-bits 1s0)
=> 1065353216

(double-float-bits 1d0)
=> 0, 1072693248

(make-single-float 1065353216)
=> 1.0

(make-double-float 0 1072693248)
=> 1.0D0

Conversion is done using low-level functions provided by the Lisp implementation. If the Lisp implementation is unsupported, we fall back to casting via FFI.

Note that implementations differ in whether they preserve the sign bit as the sign of the returned integer. This library normalizes the functions so they all take, and receive, only unsigned integers.

License

MIT