Fix logb on zero, infinite, NaN args

Change logb to return -infinity, +infinity, and NaN respectively.
Formerly logb returned an extreme fixnum to represent
infinity, but this is no longer the right thing to do now that
we have bignums and there is no extreme integer.
* doc/lispref/numbers.texi (Float Basics), etc/NEWS: Document.
* src/floatfns.c (Flogb): Implement this.

1 files changed, 11 insertions, 16 deletions

diff --git a/src/floatfns.c b/src/floatfns.c
index 2d76b97eec7..a913aad5aac 100644
--- a/src/floatfns.c
+++ b/src/floatfns.c
@@ -306,27 +306,22 @@ This is the same as the exponent of a float.  */)
   if (FLOATP (arg))
     {
       double f = XFLOAT_DATA (arg);
-
       if (f == 0)
-        value = MOST_NEGATIVE_FIXNUM;
-      else if (isfinite (f))
-        {
-          int ivalue;
-          frexp (f, &ivalue);
-          value = ivalue - 1;
-        }
-      else
-        value = MOST_POSITIVE_FIXNUM;
+        return make_float (-HUGE_VAL);
+      if (!isfinite (f))
+        return f < 0 ? make_float (-f) : arg;
+      int ivalue;
+      frexp (f, &ivalue);
+      value = ivalue - 1;
     }
-  else if (BIGNUMP (arg))
+  else if (!FIXNUMP (arg))
     value = mpz_sizeinbase (XBIGNUM (arg)->value, 2) - 1;
   else
     {
-      eassert (FIXNUMP (arg));
-      EMACS_INT i = eabs (XFIXNUM (arg));
-      value = (i == 0
-               ? MOST_NEGATIVE_FIXNUM
-               : EMACS_UINT_WIDTH - 1 - ecount_leading_zeros (i));
+      EMACS_INT i = XFIXNUM (arg);
+      if (i == 0)
+        return make_float (-HUGE_VAL);
+      value = EMACS_UINT_WIDTH - 1 - ecount_leading_zeros (eabs (i));
     }
 
   return make_fixnum (value);