|
|
@ -1,8 +1,5 @@ |
|
|
|
// cl_float_to_FF_pointer().
|
|
|
|
|
|
|
|
// Ensure that union ffloatjanus has a 'float' alternative.
|
|
|
|
#define FAST_FLOAT
|
|
|
|
|
|
|
|
// General includes.
|
|
|
|
#include "cl_sysdep.h"
|
|
|
|
|
|
|
@ -15,9 +12,9 @@ namespace cln { |
|
|
|
|
|
|
|
cl_private_thing cl_float_to_FF_pointer (const float x) |
|
|
|
{ |
|
|
|
var union ffloatjanus val_; |
|
|
|
val_.machine_float = x; |
|
|
|
var ffloat val = val_.eksplicit; |
|
|
|
var union { ffloat eksplicit; float machine_float; } u; |
|
|
|
u.machine_float = x; |
|
|
|
var ffloat val = u.eksplicit; |
|
|
|
var uintL exp = (val >> FF_mant_len) & (bit(FF_exp_len)-1); // e
|
|
|
|
if (exp == 0) // e=0 ?
|
|
|
|
// vorzeichenbehaftete 0.0 oder subnormale Zahl
|
|
|
|