d34bbd6ca3
* src/rtl/hbstrfmt.c
! fixed GPF regression after 2013-02-12 12:49 UTC+0100
when parameter was missing for d, x, X, f types
* utils/hbmk2/hbmk2.*.po
* utils/hbmk2/hbmk2.prg
+ show warning when both -inc and -rebuild[all] options are
unconditionally passed together on command-line or inside
one make file.
These instruct hbmk2 to do all the extra work involved
with an incremental build, but to do it _every time_, making
it effectively _not_ incremental, only leaving temp files
in workdir, slower and thus probably not what one wants.
(EXPERIMENTAL)
Best practice is to put -inc inside the .hbp file, and
use -rebuild[all] on command-line when a rebuild is needed.
* minor cleanup to two existing warnings
+ documented exit code when -run is used
+ further clarified .hbp file format in help
+ contrib/hbbz2/3rd/bz2/bz2.dif
* contrib/hbbz2/3rd/bz2/bz2.hbp
* contrib/hbbz2/3rd/bz2/bzlib_pr.h
! fixed libbzip2 header bug where it cannot link
mixed C/C++ builds with option -DBZ_NO_STDIO enabled
+ config/win/tcc.mk
* include/hbdefs.h
* include/hbfloat.h
* include/hbinit.h
* src/3rd/png/Makefile
* src/common/hbprintf.c
* src/common/hbver.c
* src/rtl/arc4.c
* src/rtl/gtwvt/gtwvt.c
* src/rtl/hbsocket.c
* src/vm/extrap.c
* src/vm/hvm.c
* src/vm/mainstd.c
* utils/hbmk2/hbmk2.prg
+ added rough and experimental Tiny C Compiler support.
Some Harbour features disabled. Some hunks quick and dirty.
Needs: set HB_COMPILER=tcc; set HB_PLATFORM=win; set HB_BUILD_DYN=no
Final Harbour apps fail in GT initialization.
238 lines
8.4 KiB
C
238 lines
8.4 KiB
C
/*
|
|
* $Id$
|
|
*/
|
|
|
|
/*
|
|
* Harbour Project source code:
|
|
* Harbour floating point math macros
|
|
*
|
|
* Copyright 2009 Przemyslaw Czerpak <druzus / at / priv.onet.pl>
|
|
* www - http://harbour-project.org
|
|
*
|
|
* This program 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.
|
|
*
|
|
* This program 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 this software; see the file COPYING.txt. If not, write to
|
|
* the Free Software Foundation, Inc., 59 Temple Place, Suite 330,
|
|
* Boston, MA 02111-1307 USA (or visit the web site http://www.gnu.org/).
|
|
*
|
|
* As a special exception, the Harbour Project gives permission for
|
|
* additional uses of the text contained in its release of Harbour.
|
|
*
|
|
* The exception is that, if you link the Harbour libraries with other
|
|
* files to produce an executable, this does not by itself cause the
|
|
* resulting executable to be covered by the GNU General Public License.
|
|
* Your use of that executable is in no way restricted on account of
|
|
* linking the Harbour library code into it.
|
|
*
|
|
* This exception does not however invalidate any other reasons why
|
|
* the executable file might be covered by the GNU General Public License.
|
|
*
|
|
* This exception applies only to the code released by the Harbour
|
|
* Project under the name Harbour. If you copy code from other
|
|
* Harbour Project or Free Software Foundation releases into a copy of
|
|
* Harbour, as the General Public License permits, the exception does
|
|
* not apply to the code that you add in this way. To avoid misleading
|
|
* anyone as to the status of such modified files, you must delete
|
|
* this exception notice from them.
|
|
*
|
|
* If you write modifications of your own for Harbour, it is your choice
|
|
* whether to permit this exception to apply to your modifications.
|
|
* If you do not wish that, delete this exception notice.
|
|
*
|
|
*/
|
|
|
|
#ifndef HB_FLOAT_H_
|
|
#define HB_FLOAT_H_
|
|
|
|
#ifndef _GNU_SOURCE
|
|
# define _GNU_SOURCE
|
|
#endif
|
|
|
|
#ifndef __EXTENSIONS__
|
|
# define __EXTENSIONS__
|
|
#endif
|
|
|
|
/* workaround for some missing C99 math macros in SunOS GCC
|
|
* used in C++ mode
|
|
*/
|
|
#if ! defined( __C99FEATURES__ ) && defined( __GNUC__ ) && defined( __sun__ )
|
|
# define __C99FEATURES__
|
|
#endif
|
|
|
|
#if defined( __TINYC__ )
|
|
#define __NO_ISOCEXT
|
|
#endif
|
|
|
|
#include "hbapi.h"
|
|
|
|
#if defined( __DJGPP__ )
|
|
# include <libm/math.h>
|
|
/* _LIB_VERSION_TYPE _LIB_VERSION = _XOPEN_; */
|
|
#else
|
|
# include <math.h>
|
|
# if defined( __BORLANDC__ ) || defined( __WATCOMC__ ) || defined( _MSC_VER ) || defined( HB_OS_MINIX )
|
|
# include <float.h>
|
|
# elif defined( HB_OS_SUNOS )
|
|
# include <ieeefp.h> /* for finite() */
|
|
# endif
|
|
#endif
|
|
|
|
|
|
#if defined( HB_LONG_DOUBLE_OFF ) && ! defined( __NO_LONGDOUBLE__ )
|
|
# define __NO_LONGDOUBLE__
|
|
#endif
|
|
|
|
|
|
#define _HB_NUM_NAN 1
|
|
#define _HB_NUM_PINF 2
|
|
#define _HB_NUM_NINF 4
|
|
|
|
|
|
#if defined( __BORLANDC__ ) && 0
|
|
/* do not use Borland C _fpclass[l]() function.
|
|
* it switches internal logic used for floating point calculation
|
|
* in this compiler reducing the precision to 'float' type.
|
|
*/
|
|
# ifdef __NO_LONGDOUBLE__
|
|
# define hb_fpclassify( d ) _fpclass( d )
|
|
# else
|
|
# define hb_fpclassify( d ) ( sizeof( d ) == sizeof( double ) ? \
|
|
_fpclass( d ) : _fpclassl( d ) )
|
|
# endif
|
|
#endif
|
|
|
|
|
|
/* on some platforms signbit() needs _GNU_SOURCE defined.
|
|
* If it's not available on given platform then it can be replaced by
|
|
* 'value < 0' but in such case -0.0 will be shown as 0.0
|
|
*/
|
|
#if defined( _ISOC99_SOURCE ) || defined( _STDC_C99 ) || defined( signbit )
|
|
|
|
# define hb_signbit( d ) signbit( d )
|
|
|
|
#elif defined( __BORLANDC__ ) && defined( hb_fpclassify )
|
|
|
|
# define hb_signbit( d ) ( ( hb_fpclassify( d ) & ( _FPCLASS_NINF | _FPCLASS_NZ ) ) != 0 )
|
|
|
|
#elif 0 /* TODO: add other C compilers here (check their version number) */
|
|
#else
|
|
|
|
# define hb_signbit( d ) ( d < 0 )
|
|
|
|
#endif
|
|
|
|
|
|
|
|
#if defined( _ISOC99_SOURCE ) || defined( _STDC_C99 ) || \
|
|
( defined( isfinite ) && defined( isnan ) && defined( isinf ) )
|
|
|
|
/* use C99 macros */
|
|
# define hb_isfinite( d ) isfinite( d )
|
|
# define HB_NUMTYPE( v, d ) do { \
|
|
v = ( isfinite( d ) ? 0 : \
|
|
( isnan( d ) ? _HB_NUM_NAN : \
|
|
( isinf( d ) < 0 ? _HB_NUM_NINF : \
|
|
_HB_NUM_PINF ) ) ); \
|
|
} while( 0 )
|
|
|
|
#elif ( defined( __GNUC__ ) || \
|
|
defined( __SUNPRO_C ) || defined( __SUNPRO_CC ) ) && \
|
|
( defined( _BSD_SOURCE ) || defined( _SVID_SOURCE ) || \
|
|
defined( _XOPEN_SOURCE ) )
|
|
|
|
/* use BSD floating point functions */
|
|
|
|
# define hb_isfinite( d ) finite( d )
|
|
# define HB_NUMTYPE( v, d ) do { \
|
|
v = ( finite( d ) ? 0 : \
|
|
( isnan( d ) ? _HB_NUM_NAN : \
|
|
( isinf( d ) < 0 ? _HB_NUM_NINF : \
|
|
_HB_NUM_PINF ) ) ); \
|
|
} while( 0 )
|
|
# if ! defined( __NO_LONGDOUBLE__ ) && ! defined( HB_OS_SUNOS )
|
|
# define HB_NUMTYPEL( v, d ) do { \
|
|
v = ( finitel( d ) ? 0 : \
|
|
( isnanl( d ) ? _HB_NUM_NAN : \
|
|
( isinfl( d ) < 0 ? _HB_NUM_NINF : \
|
|
_HB_NUM_PINF ) ) ); \
|
|
} while( 0 )
|
|
# endif
|
|
|
|
#elif defined( __BORLANDC__ )
|
|
|
|
# define hb_isfinite( d ) _finite( d )
|
|
# if defined( hb_fpclassify )
|
|
# define HB_NUMTYPE( v, d ) do { \
|
|
int t = hb_fpclassify( d ); \
|
|
v = ( ( t & ( _FPCLASS_UNSUP | _FPCLASS_SNAN | _FPCLASS_QNAN ) ) ? _HB_NUM_NAN : \
|
|
( ( t & _FPCLASS_NINF ) ? _HB_NUM_NINF : \
|
|
( ( t & _FPCLASS_PINF ) ? _HB_NUM_PINF : 0 ) ) ); \
|
|
} while( 0 )
|
|
# else
|
|
# define HB_NUMTYPE( v, d ) do { \
|
|
v = ( _finite( d ) ? 0 : \
|
|
( _isnan( d ) ? _HB_NUM_NAN : \
|
|
_HB_NUM_PINF ) ); \
|
|
} while( 0 )
|
|
# if ! defined( __NO_LONGDOUBLE__ )
|
|
# define HB_NUMTYPEL( v, d ) do { \
|
|
v = ( _finitel( d ) ? 0 : \
|
|
( _isnanl( d ) ? _HB_NUM_NAN : \
|
|
_HB_NUM_PINF ) ); \
|
|
} while( 0 )
|
|
# endif
|
|
# endif
|
|
|
|
#elif 0 /* TODO: add other C compilers here (check their version number) */
|
|
#else
|
|
|
|
# if defined( __RSXNT__ ) || defined( __EMX__ ) || \
|
|
defined( __XCC__ ) || defined( __POCC__ ) || \
|
|
defined( __MINGW32__ ) || defined( HB_OS_HPUX ) || defined( HB_OS_MINIX )
|
|
# define hb_isfinite( d ) isfinite( d )
|
|
# elif defined( _MSC_VER )
|
|
# define hb_isfinite( d ) _finite( ( double ) d )
|
|
# elif defined( __BORLANDC__ ) || defined( __WATCOMC__ )
|
|
# define hb_isfinite( d ) _finite( d )
|
|
# elif defined( __GNUC__ ) || defined( __DJGPP__ ) || defined( __LCC__ ) || \
|
|
defined( HB_OS_SUNOS )
|
|
# define hb_isfinite( d ) finite( d )
|
|
# endif
|
|
|
|
# if defined( hb_isfinite )
|
|
# define HB_NUMTYPE( v, d ) do { \
|
|
v = hb_isfinite( d ) ? 0 : _HB_NUM_NAN ; \
|
|
} while( 0 )
|
|
# else
|
|
# define hb_isfinite( d ) HB_TRUE
|
|
# define HB_NUMTYPE( v, d ) do { \
|
|
int iTODO; \
|
|
v = hb_isfinite( d ) ? 0 : _HB_NUM_NAN ; \
|
|
} while( 0 )
|
|
# endif
|
|
|
|
#endif
|
|
|
|
#if ! defined( HB_NUMTYPEL )
|
|
# define HB_NUMTYPEL( v, d ) HB_NUMTYPE( v, d )
|
|
#endif
|
|
|
|
|
|
/* NOTE: Workaround for Pellec C 5.00 not having an 'inf' (HUGE_VAL)
|
|
in '-Tarm-coff' mode. [vszakats] */
|
|
#if defined( __POCC__ ) && defined( HB_OS_WIN_CE )
|
|
#undef HUGE_VAL
|
|
#define HUGE_VAL ( 1.0 / ( 1.0, 0.0 ) )
|
|
#endif
|
|
|
|
#endif /* HB_FLOAT_H_ */
|