Derek Parnell wrote: > I'm struggling to see why the compiler cannot just disallow any > signed<->unsigned implicit conversion? Is it a matter of backward > compatibility again? What's the signed-ness of 5? When you index a pointer, is the index signed or unsigned?