about summary refs log tree commit diff
path: root/math
diff options
context:
space:
mode:
Diffstat (limited to 'math')
-rw-r--r--math/Makefile5
-rw-r--r--math/test-tgmath.c666
2 files changed, 669 insertions, 2 deletions
diff --git a/math/Makefile b/math/Makefile
index 2e240197a5..f02517c059 100644
--- a/math/Makefile
+++ b/math/Makefile
@@ -1,4 +1,4 @@
-# Copyright (C) 1996, 1997, 1998, 1999, 2000 Free Software Foundation, Inc.
+# Copyright (C) 1996,1997,1998,1999,2000,2001 Free Software Foundation, Inc.
 # This file is part of the GNU C Library.
 
 # The GNU C Library is free software; you can redistribute it and/or
@@ -79,7 +79,7 @@ include ../Makeconfig
 
 # Rules for the test suite.
 tests = test-matherr test-fenv atest-exp atest-sincos atest-exp2 basic-test \
-	test-misc test-fpucw tst-definitions
+	test-misc test-fpucw tst-definitions test-tgmath
 # We do the `long double' tests only if this data type is available and
 # distinct from `double'.
 test-longdouble-yes = test-ldouble test-ildoubl
@@ -120,6 +120,7 @@ CFLAGS-test-ldouble.c = -fno-inline -ffloat-store
 CFLAGS-test-ifloat.c = -U__LIBC_INTERNAL_MATH_INLINES -D__FAST_MATH__
 CFLAGS-test-idouble.c = -U__LIBC_INTERNAL_MATH_INLINES -D__FAST_MATH__
 CFLAGS-test-ildoubl.c = -U__LIBC_INTERNAL_MATH_INLINES -D__FAST_MATH__
+CFLAGS-test-tgmath.c = -fno-builtin
 
 distribute += libm-test.inc gen-libm-test.pl README.libm-test
 
diff --git a/math/test-tgmath.c b/math/test-tgmath.c
new file mode 100644
index 0000000000..0040738244
--- /dev/null
+++ b/math/test-tgmath.c
@@ -0,0 +1,666 @@
+/* Test compilation of tgmath macros.
+   Copyright (C) 2001 Free Software Foundation, Inc.
+   This file is part of the GNU C Library.
+   Contributed by Jakub Jelinek <jakub@redhat.com> and
+   Ulrich Drepper <drepper@redhat.com>, 2001.
+
+   The GNU C Library is free software; you can redistribute it and/or
+   modify it under the terms of the GNU Library General Public License as
+   published by the Free Software Foundation; either version 2 of the
+   License, or (at your option) any later version.
+
+   The GNU C Library 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
+   Library General Public License for more details.
+
+   You should have received a copy of the GNU Library General Public
+   License along with the GNU C Library; see the file COPYING.LIB.  If not,
+   write to the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
+   Boston, MA 02111-1307, USA.  */
+
+#ifndef HAVE_MAIN
+#undef __NO_MATH_INLINES
+#define __NO_MATH_INLINES 1
+#include <math.h>
+#include <stdio.h>
+#include <tgmath.h>
+
+#define DEBUG
+
+static void compile_test (void);
+static void compile_testf (void);
+static void compile_testl (void);
+
+float fx;
+double dx;
+long double lx;
+
+int count_double;
+int count_float;
+int count_ldouble;
+
+#define NCALLS 115
+
+int
+main (void)
+{
+  int result = 0;
+
+  count_float = count_double = count_ldouble = 0;
+  compile_test ();
+  if (count_float != 0)
+    {
+      puts ("float function called for double test");
+      result = 1;
+    }
+  if (count_ldouble != 0)
+    {
+      puts ("long double function called for double test");
+      result = 1;
+    }
+  if (count_double < NCALLS)
+    {
+      printf ("double functions not called often enough (%d)\n",
+	      count_double);
+      result = 1;
+    }
+  else if (count_double > NCALLS)
+    {
+      printf ("double functions called too often (%d)\n",
+	      count_double);
+      result = 1;
+    }
+
+  count_float = count_double = count_ldouble = 0;
+  compile_testf ();
+  if (count_double != 0)
+    {
+      puts ("double function called for float test");
+      result = 1;
+    }
+  if (count_ldouble != 0)
+    {
+      puts ("long double function called for float test");
+      result = 1;
+    }
+  if (count_float < NCALLS)
+    {
+      printf ("float functions not called often enough (%d)\n", count_float);
+      result = 1;
+    }
+  else if (count_float > NCALLS)
+    {
+      printf ("float functions called too often (%d)\n",
+	      count_double);
+      result = 1;
+    }
+
+#ifndef NO_LONG_DOUBLE
+  count_float = count_double = count_ldouble = 0;
+  compile_testl ();
+  if (count_float != 0)
+    {
+      puts ("float function called for long double test");
+      result = 1;
+    }
+  if (count_double != 0)
+    {
+      puts ("double function called for long double test");
+      result = 1;
+    }
+  if (count_ldouble < NCALLS)
+    {
+      printf ("long double functions not called often enough (%d)\n",
+	      count_ldouble);
+      result = 1;
+    }
+  else if (count_ldouble > NCALLS)
+    {
+      printf ("long double functions called too often (%d)\n",
+	      count_double);
+      result = 1;
+    }
+#endif
+
+  return result;
+}
+
+/* Now generate the three functions.  */
+#define HAVE_MAIN
+
+#define F(name) name
+#define TYPE double
+#define x dx
+#define count count_double
+#include "test-tgmath.c"
+
+#define F(name) name##f
+#define TYPE float
+#define x fx
+#define count count_float
+#include "test-tgmath.c"
+
+#ifndef NO_LONG_DOUBLE
+#define F(name) name##l
+#define TYPE long double
+#define x lx
+#define count count_ldouble
+#include "test-tgmath.c"
+#endif
+
+#else
+
+#ifdef DEBUG
+#define P() puts (__FUNCTION__)
+#else
+#define P()
+#endif
+
+static void
+F(compile_test) (void)
+{
+  TYPE a, b, c;
+  int i;
+  long int j;
+  long long int k;
+
+  a = cos (cos (x));
+  b = acos (acos (a));
+  a = sin (sin (x));
+  b = asin (asin (a));
+  a = tan (tan (x));
+  b = atan (atan (a));
+  c = atan2 (atan2 (a, c), atan2 (b, x));
+  a = cosh (cosh (x));
+  b = acosh (acosh (a));
+  a = sinh (sinh (x));
+  b = asinh (asinh (a));
+  a = tanh (tanh (x));
+  b = atanh (atanh (a));
+  a = exp (exp (x));
+  b = log (log (a));
+  a = log10 (log10 (x));
+  b = ldexp (ldexp (a, 1), 5);
+  a = frexp (frexp (x, &i), &i);
+  b = expm1 (expm1 (a));
+  a = log1p (log1p (x));
+  b = logb (logb (a));
+  a = exp2 (exp2 (x));
+  b = log2 (log2 (a));
+  a = pow (pow (x, a), pow (c, b));
+  b = sqrt (sqrt (a));
+  a = hypot (hypot (x, b), hypot (c, a));
+  b = cbrt (cbrt (a));
+  a = ceil (ceil (x));
+  b = fabs (fabs (a));
+  a = floor (floor (x));
+  b = fmod (fmod (a, b), fmod (c, x));
+  a = nearbyint (nearbyint (x));
+  b = round (round (a));
+  a = trunc (trunc (x));
+  b = remquo (remquo (a, b, &i), remquo (c, x, &i), &i);
+  j = lrint (x) + lround (a);
+  k = llrint (b) + llround (c);
+  a = erf (erf (x));
+  b = erfc (erfc (a));
+  a = tgamma (tgamma (x));
+  b = lgamma (lgamma (a));
+  a = rint (rint (x));
+  b = nextafter (nextafter (a, b), nextafter (c, x));
+  a = nexttoward (nexttoward (x, a), c);
+  b = remainder (remainder (a, b), remainder (c, x));
+  a = scalb (scalb (x, a), (TYPE) (6));
+  k = scalbn (a, 7) + scalbln (c, 10l);
+  i = ilogb (x);
+  a = fdim (fdim (x, a), fdim (c, b));
+  b = fmax (fmax (a, x), fmax (c, b));
+  a = fmin (fmin (x, a), fmin (c, b));
+  b = fma (sin (a), sin (x), sin (c));
+}
+#undef x
+
+
+TYPE
+(F(cos)) (TYPE x)
+{
+  ++count;
+  return x;
+}
+
+TYPE
+(F(acos)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(sin)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(asin)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(tan)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(atan)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(atan2)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(cosh)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(acosh)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(sinh)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(asinh)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(tanh)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(atanh)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(exp)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(log)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(log10)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(ldexp)) (TYPE x, int y)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(frexp)) (TYPE x, int *y)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(expm1)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(log1p)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(logb)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(exp2)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(log2)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(pow)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(sqrt)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(hypot)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(cbrt)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(ceil)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(fabs)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(floor)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(fmod)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(nearbyint)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(round)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(trunc)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(remquo)) (TYPE x, TYPE y, int *i)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+long int
+(F(lrint)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+long int
+(F(lround)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+long long int
+(F(llrint)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+long long int
+(F(llround)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(erf)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(erfc)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(tgamma)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(lgamma)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(rint)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(nextafter)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(nexttoward)) (TYPE x, long double y)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(remainder)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(scalb)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(scalbn)) (TYPE x, int y)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(scalbln)) (TYPE x, long int y)
+{
+  ++count;
+  P();
+  return x;
+}
+
+int
+(F(ilogb)) (TYPE x)
+{
+  ++count;
+  P();
+  return x;
+}
+
+TYPE
+(F(fdim)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(fmin)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(fmax)) (TYPE x, TYPE y)
+{
+  ++count;
+  P();
+  return x + y;
+}
+
+TYPE
+(F(fma)) (TYPE x, TYPE y, TYPE z)
+{
+  ++count;
+  P();
+  return x + y + z;
+}
+
+#undef F
+#undef TYPE
+#undef count
+#endif