Corrections in the implementation of '%' for floats.

The multiplication (m*b) used to test whether 'm' is non-zero and
'm' and 'b' have different signs can underflow for very small numbers,
giving a wrong result. The use of explicit comparisons solves this
problem. This commit also adds several new tests for '%' (both for
floats and for integers) to exercise more corner cases, such as
very large and very small values.

lobject.c

@@ -106,11 +106,7 @@ static lua_Number numarith (lua_State *L, int op, lua_Number v1,
     case LUA_OPPOW: return luai_numpow(L, v1, v2);
     case LUA_OPIDIV: return luai_numidiv(L, v1, v2);
     case LUA_OPUNM: return luai_numunm(L, v1);
-    case LUA_OPMOD: {
-      lua_Number m;
-      luai_nummod(L, v1, v2, m);
-      return m;
-    }
+    case LUA_OPMOD: return luaV_modf(L, v1, v2);
     default: lua_assert(0); return 0;
   }
 }