Reapply "[libc++] Simplify the implementation of std::sort a bit (#104902)" (#114023)

This reverts commit ef44e46. The patch was originally reverted
because it was
deemed to introduce a performance regression for small inputs, however
it also fixed
a previous performance regression for larger inputs. So overall, this
patch is desirable.

Author: philnik777

libcxx/include/__algorithm/comp.h

@@ -42,6 +42,9 @@ struct __less<void, void> {
   }
 };
 
+template <class _Tp>
+inline const bool __desugars_to_v<__less_tag, __less<>, _Tp, _Tp> = true;
+
 template <class _Tp>
 inline const bool __desugars_to_v<__totally_ordered_less_tag, __less<>, _Tp, _Tp> = is_integral<_Tp>::value;
 

libcxx/include/__algorithm/ranges_minmax.h

@@ -89,7 +89,7 @@ struct __minmax {
     // vectorize the code.
     if constexpr (contiguous_range<_Range> && is_integral_v<_ValueT> &&
                   __is_cheap_to_copy<_ValueT> & __is_identity<_Proj>::value &&
-                  __desugars_to_v<__totally_ordered_less_tag, _Comp, _ValueT, _ValueT>) {
+                  __desugars_to_v<__less_tag, _Comp, _ValueT, _ValueT>) {
       minmax_result<_ValueT> __result = {__r[0], __r[0]};
       for (auto __e : __r) {
         if (__e < __result.min)