Use casts for most common atomic cases (#227)

Resolves #85 / DevCom-706195

Casey applying the new atomic implementation fixed us breaking the
strict aliasing rules, but the memcpy is causing a code size regression
for non-`/Oi` customers. This change should restore code size for the
most common uses of atomic, which are `atomic<integral>` and
`atomic<pointer>`.

stl/inc/atomic

@@ -238,10 +238,19 @@ template <class _Integral, class _Ty>
 _NODISCARD _Integral _Atomic_reinterpret_as(const _Ty& _Source) noexcept {
     // interprets _Source as the supplied integral type
     static_assert(is_integral_v<_Integral>, "Tried to reinterpret memory as non-integral");
+#if _HAS_IF_CONSTEXPR
+    if constexpr (is_integral_v<_Ty> && sizeof(_Integral) == sizeof(_Ty)) {
+        return static_cast<_Integral>(_Source);
+    } else if constexpr (is_pointer_v<_Ty> && sizeof(_Integral) == sizeof(_Ty)) {
+        return reinterpret_cast<_Integral>(_Source);
+    } else
+#endif // _HAS_IF_CONSTEXPR
+    {
         _Integral _Result{}; // zero padding bits
         _CSTD memcpy(&_Result, _STD addressof(_Source), sizeof(_Source));
         return _Result;
     }
+}
 
 // FUNCTION _Load_barrier
 inline void _Load_barrier(const memory_order _Order) noexcept { // implement memory barrier for atomic load functions