gcc/libdecnumber/decimal32.h

104 lines
4.0 KiB
C

/* Decimal 32-bit format module header for the decNumber C Library
Copyright (C) 2005 Free Software Foundation, Inc.
Contributed by IBM Corporation. Author Mike Cowlishaw.
This file is part of GCC.
GCC is free software; you can redistribute it and/or modify it under
the terms of the GNU General Public License as published by the Free
Software Foundation; either version 2, or (at your option) any later
version.
GCC is distributed in the hope that it will be useful, but WITHOUT ANY
WARRANTY; without even the implied warranty of MERCHANTABILITY or
FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
for more details.
You should have received a copy of the GNU General Public License
along with GCC; see the file COPYING. If not, write to the Free
Software Foundation, 51 Franklin Street, Fifth Floor, Boston, MA
02110-1301, USA. */
#if !defined(DECIMAL32)
#define DECIMAL32
#define DEC32NAME "decimal32" /* Short name */
#define DEC32FULLNAME "Decimal 32-bit Number" /* Verbose name */
#define DEC32AUTHOR "Mike Cowlishaw" /* Who to blame */
/* parameters for decimal32s */
#define DECIMAL32_Bytes 4 /* length */
#define DECIMAL32_Pmax 7 /* maximum precision (digits) */
#define DECIMAL32_Emax 96 /* maximum adjusted exponent */
#define DECIMAL32_Emin -95 /* minimum adjusted exponent */
#define DECIMAL32_Bias 101 /* bias for the exponent */
#define DECIMAL32_String 15 /* maximum string length, +1 */
/* highest biased exponent (Elimit-1) */
#define DECIMAL32_Ehigh (DECIMAL32_Emax+DECIMAL32_Bias-DECIMAL32_Pmax+1)
#ifndef DECNUMDIGITS
#define DECNUMDIGITS DECIMAL32_Pmax /* size if not already defined */
#endif
#ifndef DECNUMBER
#include "decNumber.h" /* context and number library */
#endif
/* Decimal 32-bit type, accessible by bytes */
typedef struct
{
uint8_t bytes[DECIMAL32_Bytes]; /* decimal32: 1, 5, 6, 20 bits */
} decimal32;
/* special values [top byte excluding sign bit; last two bits are
don't-care for Infinity on input, last bit don't-care for NaN] */
#if !defined(DECIMAL_NaN)
#define DECIMAL_NaN 0x7c /* 0 11111 00 NaN */
#define DECIMAL_sNaN 0x7e /* 0 11111 10 sNaN */
#define DECIMAL_Inf 0x78 /* 0 11110 00 Infinity */
#endif
/* Macros for accessing decimal32 fields. These assume the argument
is a reference (pointer) to the decimal32 structure */
/* Get sign */
#define decimal32Sign(d) ((unsigned)(d)->bytes[0]>>7)
/* Get combination field */
#define decimal32Comb(d) (((d)->bytes[0] & 0x7c)>>2)
/* Get exponent continuation [does not remove bias] */
#define decimal32ExpCon(d) ((((d)->bytes[0] & 0x03)<<4) \
| ((unsigned)(d)->bytes[1]>>4))
/* Set sign [this assumes sign previously 0] */
#define decimal32SetSign(d, b) { \
(d)->bytes[0]|=((unsigned)(b)<<7);}
/* Set exponent continuation [does not apply bias] */
/* This assumes range has been checked and exponent previously 0; */
/* type of exponent must be unsigned */
#define decimal32SetExpCon(d, e) { \
(d)->bytes[0]|=(uint8_t)((e)>>4); \
(d)->bytes[1]|=(uint8_t)(((e)&0x0F)<<4);}
/* ------------------------------------------------------------------ */
/* Routines */
/* ------------------------------------------------------------------ */
#ifdef IN_LIBGCC2
#define decimal32FromString __decimal32FromString
#define decimal32ToString __decimal32ToString
#define decimal32ToEngString __decimal32ToEngString
#define decimal32FromNumber __decimal32FromNumber
#define decimal32ToNumber __decimal32ToNumber
#endif
/* String conversions. */
decimal32 *decimal32FromString (decimal32 *, char *, decContext *);
char *decimal32ToString (decimal32 *, char *);
char *decimal32ToEngString (decimal32 *, char *);
/* decNumber conversions. */
decimal32 *decimal32FromNumber (decimal32 *, decNumber *, decContext *);
decNumber *decimal32ToNumber (decimal32 *, decNumber *);
#endif