summaryrefslogtreecommitdiff
path: root/lib/signbitf.c
diff options
context:
space:
mode:
Diffstat (limited to 'lib/signbitf.c')
-rw-r--r--lib/signbitf.c64
1 files changed, 64 insertions, 0 deletions
diff --git a/lib/signbitf.c b/lib/signbitf.c
new file mode 100644
index 0000000..e185971
--- /dev/null
+++ b/lib/signbitf.c
@@ -0,0 +1,64 @@
+/* signbit() macro: Determine the sign bit of a floating-point number.
+ Copyright (C) 2007, 2009-2016 Free Software Foundation, Inc.
+
+ 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 3 of the License, 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 program. If not, see <http://www.gnu.org/licenses/>. */
+
+#include <config.h>
+
+/* Specification. */
+#include <math.h>
+
+#include <string.h>
+#include "isnanf-nolibm.h"
+#include "float+.h"
+
+#ifdef gl_signbitf_OPTIMIZED_MACRO
+# undef gl_signbitf
+#endif
+
+int
+gl_signbitf (float arg)
+{
+#if defined FLT_SIGNBIT_WORD && defined FLT_SIGNBIT_BIT
+ /* The use of a union to extract the bits of the representation of a
+ 'long double' is safe in practice, despite of the "aliasing rules" of
+ C99, because the GCC docs say
+ "Even with '-fstrict-aliasing', type-punning is allowed, provided the
+ memory is accessed through the union type."
+ and similarly for other compilers. */
+# define NWORDS \
+ ((sizeof (float) + sizeof (unsigned int) - 1) / sizeof (unsigned int))
+ union { float value; unsigned int word[NWORDS]; } m;
+ m.value = arg;
+ return (m.word[FLT_SIGNBIT_WORD] >> FLT_SIGNBIT_BIT) & 1;
+#elif HAVE_COPYSIGNF_IN_LIBC
+ return copysignf (1.0f, arg) < 0;
+#else
+ /* This does not do the right thing for NaN, but this is irrelevant for
+ most use cases. */
+ if (isnanf (arg))
+ return 0;
+ if (arg < 0.0f)
+ return 1;
+ else if (arg == 0.0f)
+ {
+ /* Distinguish 0.0f and -0.0f. */
+ static float plus_zero = 0.0f;
+ float arg_mem = arg;
+ return (memcmp (&plus_zero, &arg_mem, SIZEOF_FLT) != 0);
+ }
+ else
+ return 0;
+#endif
+}