summaryrefslogtreecommitdiffstats
path: root/contrib/gcc/fp-test.c
diff options
context:
space:
mode:
authorobrien <obrien@FreeBSD.org>1999-08-26 09:30:50 +0000
committerobrien <obrien@FreeBSD.org>1999-08-26 09:30:50 +0000
commit0bedf4fb30066e5e1d4342a1d3914dae7d37cba7 (patch)
tree68d8110b41afd0ebbf39167b1a4918eea667a7c5 /contrib/gcc/fp-test.c
parentd4db5fb866b7ad5216abd5047774a3973b9901a9 (diff)
downloadFreeBSD-src-0bedf4fb30066e5e1d4342a1d3914dae7d37cba7.zip
FreeBSD-src-0bedf4fb30066e5e1d4342a1d3914dae7d37cba7.tar.gz
Virgin import of gcc from EGCS 1.1.2
Diffstat (limited to 'contrib/gcc/fp-test.c')
-rw-r--r--contrib/gcc/fp-test.c231
1 files changed, 231 insertions, 0 deletions
diff --git a/contrib/gcc/fp-test.c b/contrib/gcc/fp-test.c
new file mode 100644
index 0000000..667059c
--- /dev/null
+++ b/contrib/gcc/fp-test.c
@@ -0,0 +1,231 @@
+/* fp-test.c - Check that all floating-point operations are available.
+ Copyright (C) 1995 Free Software Foundation, Inc.
+ Contributed by Ronald F. Guilmette <rfg@monkeys.com>.
+
+This file is part of GNU CC.
+
+GNU CC 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.
+
+GNU CC 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 GNU CC; see the file COPYING. If not, write to
+the Free Software Foundation, 59 Temple Place - Suite 330,
+Boston, MA 02111-1307, USA. */
+
+/* This is a trivial test program which may be useful to people who are
+ porting the GCC or G++ compilers to a new system. The intent here is
+ merely to check that all floating-point operations have been provided
+ by the port. (Note that I say ``provided'' rather than ``implemented''.)
+
+ To use this file, simply compile it (with GCC or G++) and then try to
+ link it in the normal way (also using GCC or G++ respectively). If
+ all of the floating -point operations (including conversions) have
+ been provided, then this file will link without incident. If however
+ one or more of the primitive floating-point operations have not been
+ properly provided, you will get link-time errors indicating which
+ floating-point operations are unavailable.
+
+ This file will typically be used when porting the GNU compilers to
+ some system which lacks floating-point hardware, and for which
+ software emulation routines (for FP ops) are needed in order to
+ complete the port. */
+
+#if 0
+#include <math.h>
+#endif
+
+extern double acos (double);
+extern double asin (double);
+extern double atan (double);
+extern double atan2 (double, double);
+extern double cos (double);
+extern double sin (double);
+extern double tan (double);
+extern double cosh (double);
+extern double sinh (double);
+extern double tanh (double);
+extern double exp (double);
+extern double frexp (double, int *);
+extern double ldexp (double, int);
+extern double log (double);
+extern double log10 (double);
+extern double modf (double, double *);
+extern double pow (double, double);
+extern double sqrt (double);
+extern double ceil (double);
+extern double fabs (double);
+extern double floor (double);
+extern double fmod (double, double);
+
+int i1, i2 = 2;
+
+volatile signed char sc;
+volatile unsigned char uc;
+
+volatile signed short ss;
+volatile unsigned short us;
+
+volatile signed int si;
+volatile unsigned int ui;
+
+volatile signed long sl;
+volatile unsigned long ul;
+
+volatile float f1 = 1.0, f2 = 1.0, f3 = 1.0;
+volatile double d1 = 1.0, d2 = 1.0, d3 = 1.0;
+volatile long double D1 = 1.0, D2 = 1.0, D3 = 1.0;
+
+int
+main ()
+{
+ /* TYPE: float */
+
+ f1 = -f2;
+ f1 = f2 + f3;
+ f1 = f2 - f3;
+ f1 = f2 * f3;
+ f1 = f2 / f3;
+ f1 += f2;
+ f1 -= f2;
+ f1 *= f2;
+ f1 /= f2;
+
+ si = f1 == f2;
+ si = f1 != f2;
+ si = f1 > f2;
+ si = f1 < f2;
+ si = f1 >= f2;
+ si = f1 <= f2;
+
+ sc = f1;
+ uc = f1;
+ ss = f1;
+ us = f1;
+ si = f1;
+ ui = f1;
+ sl = f1;
+ ul = f1;
+ d1 = f1;
+ D1 = f1;
+
+ f1 = sc;
+ f1 = uc;
+ f1 = ss;
+ f1 = us;
+ f1 = si;
+ f1 = ui;
+ f1 = sl;
+ f1 = ul;
+ f1 = d1;
+ f1 = D1;
+
+ d1 = -d2;
+ d1 = d2 + d3;
+ d1 = d2 - d3;
+ d1 = d2 * d3;
+ d1 = d2 / d3;
+ d1 += d2;
+ d1 -= d2;
+ d1 *= d2;
+ d1 /= d2;
+
+ si = d1 == d2;
+ si = d1 != d2;
+ si = d1 > d2;
+ si = d1 < d2;
+ si = d1 >= d2;
+ si = d1 <= d2;
+
+ sc = d1;
+ uc = d1;
+ ss = d1;
+ us = d1;
+ si = d1;
+ ui = d1;
+ sl = d1;
+ ul = d1;
+ f1 = d1;
+ D1 = d1;
+
+ d1 = sc;
+ d1 = uc;
+ d1 = ss;
+ d1 = us;
+ d1 = si;
+ d1 = ui;
+ d1 = sl;
+ d1 = ul;
+ d1 = f1;
+ d1 = D1;
+
+ D1 = -D2;
+ D1 = D2 + D3;
+ D1 = D2 - D3;
+ D1 = D2 * D3;
+ D1 = D2 / D3;
+ D1 += D2;
+ D1 -= D2;
+ D1 *= D2;
+ D1 /= D2;
+
+ si = D1 == D2;
+ si = D1 != D2;
+ si = D1 > D2;
+ si = D1 < D2;
+ si = D1 >= D2;
+ si = D1 <= D2;
+
+ sc = D1;
+ uc = D1;
+ ss = D1;
+ us = D1;
+ si = D1;
+ ui = D1;
+ sl = D1;
+ ul = D1;
+ f1 = D1;
+ d1 = D1;
+
+ D1 = sc;
+ D1 = uc;
+ D1 = ss;
+ D1 = us;
+ D1 = si;
+ D1 = ui;
+ D1 = sl;
+ D1 = ul;
+ D1 = f1;
+ D1 = d1;
+
+ d1 = acos (d2);
+ d1 = asin (d2);
+ d1 = atan (d2);
+ d1 = atan2 (d2, d3);
+ d1 = cos (d2);
+ d1 = sin (d2);
+ d1 = tan (d2);
+ d1 = cosh (d2);
+ d1 = sinh (d2);
+ d1 = tanh (d2);
+ d1 = exp (d2);
+ d1 = frexp (d2, &i1);
+ d1 = ldexp (d2, i2);
+ d1 = log (d2);
+ d1 = log10 (d2);
+ d1 = modf (d2, &d3);
+ d1 = pow (d2, d3);
+ d1 = sqrt (d2);
+ d1 = ceil (d2);
+ d1 = fabs (d2);
+ d1 = floor (d2);
+ d1 = fmod (d2, d3);
+
+ return 0;
+}
OpenPOWER on IntegriCloud