|
@ -1,4 +1,4 @@ |
|
|
/* $OpenBSD: math.h,v 1.35 2016/03/17 20:55:35 jca Exp $ */ |
|
|
|
|
|
|
|
|
/* $OpenBSD: math.h,v 1.36 2018/03/10 20:52:58 kettenis Exp $ */ |
|
|
/* |
|
|
/* |
|
|
* ==================================================== |
|
|
* ==================================================== |
|
|
* Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved. |
|
|
* Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved. |
|
@ -273,6 +273,8 @@ int finite(double); |
|
|
double gamma_r(double, int *); |
|
|
double gamma_r(double, int *); |
|
|
double lgamma_r(double, int *); |
|
|
double lgamma_r(double, int *); |
|
|
|
|
|
|
|
|
|
|
|
void sincos(double, double *, double *); |
|
|
|
|
|
|
|
|
/* |
|
|
/* |
|
|
* IEEE Test Vector |
|
|
* IEEE Test Vector |
|
|
*/ |
|
|
*/ |
|
@ -382,6 +384,8 @@ int isnanf(float); |
|
|
float gammaf_r(float, int *); |
|
|
float gammaf_r(float, int *); |
|
|
float lgammaf_r(float, int *); |
|
|
float lgammaf_r(float, int *); |
|
|
|
|
|
|
|
|
|
|
|
void sincosf(float, float *, float *); |
|
|
|
|
|
|
|
|
/* |
|
|
/* |
|
|
* Float version of IEEE Test Vector |
|
|
* Float version of IEEE Test Vector |
|
|
*/ |
|
|
*/ |
|
@ -460,6 +464,13 @@ long double fminl(long double, long double); |
|
|
long double fmal(long double, long double, long double); |
|
|
long double fmal(long double, long double, long double); |
|
|
#endif /* __ISO_C_VISIBLE >= 1999 */ |
|
|
#endif /* __ISO_C_VISIBLE >= 1999 */ |
|
|
|
|
|
|
|
|
|
|
|
/* |
|
|
|
|
|
* Long double versions of BSD math library entry points |
|
|
|
|
|
*/ |
|
|
|
|
|
#if __BSD_VISIBLE |
|
|
|
|
|
void sincosl(long double, long double *, long double *); |
|
|
|
|
|
#endif |
|
|
|
|
|
|
|
|
/* |
|
|
/* |
|
|
* Library implementation |
|
|
* Library implementation |
|
|
*/ |
|
|
*/ |
|
|