math: tgammal.c fixes

this is not a full rewrite just fixes to the special case logic:
+-0 and non-integer x<INT_MIN inputs incorrectly raised invalid
exception and for +-0 the return value was wrong

so integer test and odd/even test for negative inputs are changed
and a useless overflow test was removed

diff --git a/src/math/tgammal.c b/src/math/tgammal.c
index 86782a9..5c1a02a 100644 (file)
--- a/src/math/tgammal.c
+++ b/src/math/tgammal.c
@@ -205,42 +205,36 @@ static long double stirf(long double x)
 long double tgammal(long double x)
 {
        long double p, q, z;
-       int i, sign;
 
-       if (isnan(x))
-               return NAN;
-       if (x == INFINITY)
-               return INFINITY;
-       if (x == -INFINITY)
-               return x - x;
+       if (!isfinite(x))
+               return x + INFINITY;
+
        q = fabsl(x);
        if (q > 13.0) {
-               sign = 1;
-               if (q > MAXGAML)
-                       goto goverf;
                if (x < 0.0) {
                        p = floorl(q);
-                       if (p == q)
-                               return (x - x) / (x - x);
-                       i = p;
-                       if ((i & 1) == 0)
-                               sign = -1;
                        z = q - p;
-                       if (z > 0.5L) {
-                               p += 1.0;
-                               z = q - p;
-                       }
-                       z = q * sinl(PIL * z);
-                       z = fabsl(z) * stirf(q);
-                       if (z <= PIL/LDBL_MAX) {
-goverf:
-                               return sign * INFINITY;
+                       if (z == 0)
+                               return 0 / z;
+                       if (q > MAXGAML) {
+                               z = 0;
+                       } else {
+                               if (z > 0.5) {
+                                       p += 1.0;
+                                       z = q - p;
+                               }
+                               z = q * sinl(PIL * z);
+                               z = fabsl(z) * stirf(q);
+                               z = PIL/z;
                        }
-                       z = PIL/z;
+                       if (0.5 * p == floorl(q * 0.5))
+                               z = -z;
+               } else if (x > MAXGAML) {
+                       z = x * 0x1p16383L;
                } else {
                        z = stirf(x);
                }
-               return sign * z;
+               return z;
        }
 
        z = 1.0;
@@ -268,8 +262,9 @@ goverf:
        return z;
 
 small:
-       if (x == 0.0)
-               return (x - x) / (x - x);
+       /* z==1 if x was originally +-0 */
+       if (x == 0 && z != 1)
+               return x / x;
        if (x < 0.0) {
                x = -x;
                q = z / (x * __polevll(x, SN, 8));