Merge pull request #2250 from ghaerr/owcmath

[owc libc] Add floating point math library support for OpenWatcom C on ELKS

.github/workflows/cross.yml

@@ -24,7 +24,7 @@ jobs:
 
       - name: cache
         id: cache
-        uses: actions/cache@v1
+        uses: actions/cache@v4
         with:
           path: cross
           key: cross-${{ hashFiles('tools/*') }}-${{ runner.os }}

.github/workflows/main.yml

@@ -28,7 +28,7 @@ jobs:
 
       - name: cache
         id: cache
-        uses: actions/cache@v1
+        uses: actions/cache@v4
         with:
           path: cross
           key: cross-${{ hashFiles('tools/*') }}-${{ runner.os }}

elkscmd/basic/host.h

@@ -3,10 +3,10 @@
 #define DISK_FUNCTIONS			1	/* compile in DELETE/DIR/LIST/SAVE */
 
 #ifdef __WATCOMC__
-#define MATH_FUNCTIONS			0	/* compile in COS, SIN, TAN, etc */
-#define MATH_INCLUDE_POW		0	/* compile in large POW function */
-#define MATH_DBL_PRECISION		0	/* 0=use 4-byte floats vs 8-byte double */
-#define MATH_CORRECT_NEAR_ZERO	0	/* adjust trig routines returning -0, +/-EPSILON */
+#define MATH_FUNCTIONS			1	/* compile in COS, SIN, TAN, etc */
+#define MATH_INCLUDE_POW		1	/* compile in large POW function */
+#define MATH_DBL_PRECISION		1	/* 0=use 4-byte floats vs 8-byte double */
+#define MATH_CORRECT_NEAR_ZERO	1	/* adjust trig routines returning -0, +/-EPSILON */
 #else
 #define MATH_FUNCTIONS			1	/* compile in COS, SIN, TAN, etc */
 #define MATH_INCLUDE_POW		1	/* compile in large POW function */

libc/include/math.h

@@ -4,6 +4,7 @@
  * ELKS libc math header file
  *
  * Apr 2022 Greg Haerr
+ * Mar 2025 Added OpenWatcom math library
  */
 
 #define M_E         2.7182818284590452354   /* e */
@@ -20,6 +21,12 @@
 #define M_SQRT2     1.41421356237309504880  /* sqrt(2) */
 #define M_SQRT1_2   0.70710678118654752440  /* 1/sqrt(2) */
 
+#ifdef __WATCOMC__
+/* OpenWatcom math routines */
+#include <watcom/math.h>
+#else
+
+/* ia16-elf-gcc math routines */
 double floor(double x);		/* round to largest integral value not greater than x */
 float floorf(float x);
 
@@ -68,3 +75,4 @@ float atanf(float x);
 float fminf(float x, float y);
 float fmaxf(float x, float y);
 #endif
+#endif

libc/include/watcom/math.h

@@ -0,0 +1,322 @@
+/***************************************************************************
+ * FILE: math.h/cmath (Math functions)
+ *
+ * =========================================================================
+ *
+ *                          Open Watcom Project
+ *
+ * Copyright (c) 2004-2024 The Open Watcom Contributors. All Rights Reserved.
+ * Portions Copyright (c) 1983-2002 Sybase, Inc. All Rights Reserved.
+ *
+ *    This file is automatically generated. Do not edit directly.
+ *
+ * =========================================================================
+ *
+ * Description: This header is part of the C/C++ standard library. It
+ *              describes the math functions provided by those standards.
+ ***************************************************************************/
+#ifndef _MATH_H_INCLUDED
+#define _MATH_H_INCLUDED
+
+#ifndef _ENABLE_AUTODEPEND
+ #pragma read_only_file
+#endif
+
+#ifdef __cplusplus
+
+#define _INCLUDED_FROM_MATH_H
+#include <cmath>
+#undef  _INCLUDED_FROM_MATH_H
+
+// functions in math.h
+using std::acos;
+using std::asin;
+using std::atan;
+using std::atan2;
+using std::ceil;
+using std::cos;
+using std::cosh;
+using std::exp;
+using std::fabs;
+using std::floor;
+using std::fmod;
+using std::frexp;
+using std::ldexp;
+using std::log;
+using std::log10;
+using std::modf;
+using std::pow;
+using std::sin;
+using std::sinh;
+using std::sqrt;
+using std::tan;
+using std::tanh;
+
+#if !defined( _NO_EXT_KEYS ) || __STDC_VERSION__ >= 199901L /* extensions enabled or C99 */
+
+using std::acosh;
+using std::asinh;
+using std::atanh;
+using std::cbrt;
+using std::copysign;
+using std::erf;
+using std::erfc;
+using std::exp2;
+using std::expm1;
+using std::fdim;
+using std::fmax;
+using std::fma;
+using std::fmin;
+using std::hypot;
+using std::ilogb;
+using std::lgamma;
+using std::lgamma_r;
+using std::log1p;
+using std::log2;
+using std::logb;
+using std::nearbyint;
+using std::nextafter;
+using std::remainder;
+using std::remquo;
+using std::rint;
+using std::round;
+using std::scalbn;
+using std::tgamma;
+using std::trunc;
+
+using std::nan;
+using std::nanf;
+using std::nanl;
+
+#endif /* extensions enabled */
+
+#else /* __cplusplus not defined */
+
+#ifndef __COMDEF_H_INCLUDED
+ #include <_comdef.h>
+#endif
+
+#ifdef _M_IX86
+ #pragma pack( __push, 1 )
+#else
+ #pragma pack( __push, 8 )
+#endif
+
+_WCRTLINK extern const double   *__get_HugeValue_ptr( void );
+#ifdef __FUNCTION_DATA_ACCESS
+ #define _HugeValue             (*__get_HugeValue_ptr())
+#elif defined(__SW_BR) || defined(_RTDLL)
+ #define _HugeValue             _HugeValue_br
+#endif
+_WCRTDATA extern const double _WCDATA   _HugeValue;
+#define HUGE_VAL                _HugeValue
+
+
+/* Internal library helper routines */
+_WMRTLINK extern int _FSClass( float __x );
+_WMRTLINK extern int _FDClass( double __x );
+_WMRTLINK extern int _FLClass( long double __x );
+_WMRTLINK extern int _FSSign( float __x );
+_WMRTLINK extern int _FDSign( double __x );
+_WMRTLINK extern int _FLSign( long double __x );
+
+/* Floating-point classification macros */
+#define FP_ZERO         0
+#define FP_SUBNORMAL    1
+#define FP_NORMAL       2
+#define FP_NAN          3
+#define FP_INFINITE     4
+
+#define fpclassify(__x)                                 \
+    ((sizeof(__x) == sizeof(float)) ? _FSClass(__x) :   \
+    (sizeof(__x) == sizeof(double)) ? _FDClass(__x) :   \
+                                      _FLClass(__x))
+
+#define isfinite(__x)   (fpclassify(__x) <= FP_NORMAL)
+#define isinf(__x)      (fpclassify(__x) == FP_INFINITE)
+#define isnan(__x)      (fpclassify(__x) == FP_NAN)
+#define isnormal(__x)   (fpclassify(__x) == FP_NORMAL)
+
+#define signbit(__x)                                    \
+    ((sizeof(__x) == sizeof(float)) ? _FSSign(__x) :    \
+    (sizeof(__x) == sizeof(double)) ? _FDSign(__x) :    \
+                                      _FLSign(__x))
+
+/* Floating-point Infinity and NaN constants */
+_WMRTDATA extern const float    __f_infinity;
+_WMRTDATA extern const float    __f_posqnan;
+
+#define INFINITY   __f_infinity
+#define NAN        __f_posqnan
+
+/* The sign of the gamma function as returned by
+ * lgamma
+ */
+extern int signgam;
+
+_WMRTLINK extern double ceil( double __x );
+_WMRTLINK extern double floor( double __x );
+#if defined(_M_IX86) && 0
+#pragma aux ceil  __parm __nomemory __modify __nomemory
+#pragma aux floor __parm __nomemory __modify __nomemory
+#endif
+
+_WMRTLINK extern double frexp( double __value, int *__exp );
+_WMRTLINK extern double ldexp( double __x, int __exp );
+_WMRTLINK extern double modf( double __value, double *__iptr );
+
+_WMIRTLINK extern double acos( double __x );
+_WMIRTLINK extern double asin( double __x );
+_WMIRTLINK extern double atan( double __x );
+_WMIRTLINK extern double atan2( double __y, double __x );
+_WMIRTLINK extern double cos( double __x );
+_WMIRTLINK extern double cosh( double __x );
+_WMIRTLINK extern double exp( double __x );
+_WMIRTLINK extern double fabs( double __x );
+_WMIRTLINK extern double fmod( double __x, double __y );
+_WMIRTLINK extern double log( double __x );
+_WMIRTLINK extern double log10( double __x );
+_WMIRTLINK extern double pow( double __x, double __y );
+_WMIRTLINK extern double sin( double __x );
+_WMIRTLINK extern double sinh( double __x );
+_WMIRTLINK extern double sqrt( double __x );
+_WMIRTLINK extern double tan( double __x );
+_WMIRTLINK extern double tanh( double __x );
+
+#if !defined( _NO_EXT_KEYS ) || __STDC_VERSION__ >= 199901L /* extensions enabled or C99 */
+
+_WMRTLINK extern double acosh( double __x );
+_WMRTLINK extern double asinh( double __x );
+_WMRTLINK extern double atanh( double __x );
+_WMRTLINK extern double cbrt( double __x );
+_WMRTLINK extern double copysign( double __x, double __y );
+_WMRTLINK extern double erf( double __x );
+_WMRTLINK extern double erfc( double __x );
+_WMRTLINK extern double exp2( double __x );
+_WMRTLINK extern double expm1( double __x );
+_WMRTLINK extern double fdim( double __x, double __y );
+_WMRTLINK extern double fmax( double __x, double __y );
+_WMRTLINK extern double fma( double __a, double __b, double __c );
+_WMRTLINK extern double fmin( double __x, double __y );
+_WMRTLINK extern double hypot( double __x, double __y );
+_WMRTLINK extern int ilogb( double __x );
+_WMRTLINK extern double lgamma( double __x );
+_WMRTLINK extern double lgamma_r( double __x, int *__signgamma );
+_WMRTLINK extern double log1p( double __x );
+_WMRTLINK extern double log2( double __x );
+_WMRTLINK extern double logb( double __x );
+_WMRTLINK extern long lrint( double __x );
+_WMRTLINK extern long long llrint( double __x );
+_WMRTLINK extern double nearbyint( double __x );
+_WMRTLINK extern double nextafter( double __x, double __y );
+_WMRTLINK extern double remainder( double __x, double __y );
+_WMRTLINK extern double remquo( double __x, double __y, int *__q );
+_WMRTLINK extern double rint( double __x );
+_WMRTLINK extern double round( double __x );
+_WMRTLINK extern double scalbn( double __x, int __n );
+_WMRTLINK extern double tgamma( double __x );
+_WMRTLINK extern double trunc( double __x );
+
+_WMRTLINK extern float nanf(const char *__ignored);
+_WMRTLINK extern double nan(const char *__ignored);
+#ifdef _LONG_DOUBLE_
+_WMRTLINK extern long double nanl(const char *__ignored);
+#else
+_WMRTLINK extern double nanl(const char *__ignored);
+#endif
+
+#ifdef _M_I86
+ #define FP_ILOGBNAN  (-32767-1)
+ #define FP_ILOGB0    32767
+#else
+ #define FP_ILOGBNAN  (-2147483647-1)
+ #define FP_ILOGB0    2147483647
+#endif
+
+/* Math error handling mode ala C99 */
+extern int __math_errhandling_flag;
+#define MATH_ERRNO      1
+#define MATH_ERREXCEPT  2
+/* As an extension, support Watcom errors as well */
+#define MATH_ERRWATCOM  4
+
+#define math_errhandling __math_errhandling_flag
+
+#endif /* extensions enabled */
+
+#if !defined( _NO_EXT_KEYS ) /* extensions enabled */
+
+/* non-ANSI */
+
+struct _complex {
+    double  x;
+    double  y;
+};
+
+struct complex {
+    double  x;
+    double  y;
+};
+
+_WMRTLINK extern double cabs( struct _complex );
+_WMRTLINK extern double j0( double __x );
+_WMRTLINK extern double j1( double __x );
+_WMRTLINK extern double jn( int __n, double __x );
+_WMRTLINK extern double y0( double __x );
+_WMRTLINK extern double y1( double __x );
+_WMRTLINK extern double yn( int __n, double __x );
+
+/* The following struct is used to record errors detected in the math library.
+ * matherr is called with a pointer to this struct for possible error recovery.
+ */
+
+struct _exception {
+    int     type;           /* type of error, see below */
+    const char *name;       /* name of math function */
+    double  arg1;           /* value of first argument to function */
+    double  arg2;           /* second argument (if indicated) */
+    double  retval;         /* default return value */
+};
+
+struct  exception {
+    int     type;           /* type of error, see below */
+    const char *name;       /* name of math function */
+    double  arg1;           /* value of first argument to function */
+    double  arg2;           /* second argument (if indicated) */
+    double  retval;         /* default return value */
+};
+
+#define DOMAIN          1       /* argument domain error */
+#define SING            2       /* argument singularity  */
+#define OVERFLOW        3       /* overflow range error  */
+#define UNDERFLOW       4       /* underflow range error */
+#define TLOSS           5       /* total loss of significance */
+#define PLOSS           6       /* partial loss of significance */
+
+_WCHANDLER extern int   matherr( struct _exception * );
+_WMRTLINK extern double _matherr( struct _exception * );
+_WMRTLINK extern void   _set_matherr( int (_WCCALLBACK *rtn)( struct _exception * ) );
+
+_WCRTLINK extern int _dieeetomsbin( double *__x, double *__y );
+_WCRTLINK extern int _dmsbintoieee( double *__x, double *__y );
+_WCRTLINK extern int _fieeetomsbin( float *__x, float *__y );
+_WCRTLINK extern int _fmsbintoieee( float *__x, float *__y );
+
+#endif /* extensions enabled */
+
+#if !defined(__NO_MATH_OPS) && defined(_M_IX86)
+/*
+    Defining the __NO_MATH_OPS macro will stop the compiler from
+    recognizing the following functions as intrinsic operators.
+*/
+
+#pragma intrinsic(log,cos,sin,tan,sqrt,fabs,pow,atan2,fmod)
+#pragma intrinsic(acos,asin,atan,cosh,exp,log10,sinh,tanh)
+
+#endif
+
+#pragma pack( __pop )
+
+#endif /* __cplusplus not defined */
+
+#endif /* _CMATH_INCLUDED */

libc/watcom/Makefile

@@ -8,7 +8,9 @@ include $(TOPDIR)/libc/$(COMPILER).inc
 all:
 	$(MAKE) -C syscall
 	$(MAKE) -C asm
+	$(MAKE) -C math
 
 clean:
 	$(MAKE) -C syscall clean
 	$(MAKE) -C asm clean
+	$(MAKE) -C math clean