Infinity is now by default an acceptable value for type 'float'.

This fix makes the handling of type 'float' and 'double' the same. The implementation requires the
C99 isfinite() macro, or otherwise some platform dependent equivalents, to be available.

Examples/test-suite/python/overload_numeric_runme.py

@@ -0,0 +1,43 @@
+
+from overload_numeric import *
+import math
+
+nums = Nums()
+limits = Limits()
+
+def check(got, expected):
+  if got != expected:
+    raise RuntimeError("got: " + got + " expected: " + expected)
+
+check(nums.over(0), "signed char")
+check(nums.over(0.0), "float")
+
+check(nums.over(limits.schar_min()), "signed char")
+check(nums.over(limits.schar_max()), "signed char")
+
+check(nums.over(limits.schar_min()-1), "short")
+check(nums.over(limits.schar_max()+1), "short")
+check(nums.over(limits.shrt_min()), "short")
+check(nums.over(limits.shrt_max()), "short")
+
+check(nums.over(limits.shrt_min()-1), "int")
+check(nums.over(limits.shrt_max()+1), "int")
+check(nums.over(limits.int_min()), "int")
+check(nums.over(limits.int_max()), "int")
+
+check(nums.over(limits.flt_min()), "float")
+check(nums.over(limits.flt_max()), "float")
+
+check(nums.over(limits.flt_max()*10), "double")
+check(nums.over(-limits.flt_max()*10), "double")
+check(nums.over(limits.dbl_max()), "double")
+check(nums.over(-limits.dbl_max()), "double")
+
+check(nums.over(float("inf")), "float")
+check(nums.over(float("-inf")), "float")
+check(nums.over(float("nan")), "float")
+
+# Just check if the following are accepted without exceptions being thrown
+nums.doublebounce(float("inf"))
+nums.doublebounce(float("-inf"))
+nums.doublebounce(float("nan"))