math: add exception handling functionality

According to standards SVID and SYSV.
Modified lgamma calling in case when 'signgam' variable should not be used.

Signed-off-by: Waldemar Brodkorb <wbx@uclibc-ng.org>

1 files changed, 51 insertions, 0 deletions

diff --git a/libm/w_tgammal.c b/libm/w_tgammal.c
new file mode 100644
index 000000000..64ce302c0
--- /dev/null
+++ b/libm/w_tgammal.c
@@ -0,0 +1,51 @@
+/* w_gammal.c -- long double version of w_gamma.c.
+ * Conversion to long double by Ulrich Drepper,
+ * Cygnus Support, drepper@cygnus.com.
+ */
+
+/*
+ * ====================================================
+ * Copyright (C) 1993 by Sun Microsystems, Inc. All rights reserved.
+ *
+ * Developed at SunPro, a Sun Microsystems, Inc. business.
+ * Permission to use, copy, modify, and distribute this
+ * software is freely granted, provided that this notice
+ * is preserved.
+ * ====================================================
+ */
+
+/* long double gammal(double x)
+ * Return the Gamma function of x.
+ */
+
+#include <math.h>
+#include "math_private.h"
+#if __UCLIBC_HAS_FENV__
+#include <errno.h>
+#endif
+
+#if !defined __NO_LONG_DOUBLE_MATH
+long double
+tgammal(long double x)
+{
+# if __UCLIBC_HAS_FENV__
+	long double y = (long double) __ieee754_tgamma((long double)x);
+
+	if(__builtin_expect (!isfinite (y) || y == 0, 0)
+	   && (isfinite (x) || (isinf (x) && x < 0.0))
+	   && _LIB_VERSION != _IEEE_) {
+	  if(x==0.0)
+	    return __kernel_standard_l(x,x,250); /* tgamma pole */
+	  else if(floorl(x)==x&&x<0.0L)
+	    return __kernel_standard_l(x,x,241); /* tgamma domain */
+	  else if (y == 0)
+	    __set_errno (ERANGE); /* tgamma underflow */
+	  else
+	    return __kernel_standard_l(x,x,240); /* tgamma overflow */
+	}
+	return y;
+# else
+	return (long double) __ieee754_tgamma((long double)x);
+# endif /* __UCLIBC_HAS_FENV__ */
+}
+#endif /* __NO_LONG_DOUBLE_MATH */