[Please CC my personal email address in the replies otherwise I'll can't properly follow-up]

So let's take a better example:

template <typename T>

    {

        // Let the compiler generate the most efficient code

        template <qualifier Q>

            inline T Q operator () (node_proxy & __y, T Q po)

            {

            }

struct ConstClass

{

    constexpr ConstClass() {}

};

struct NonConstClass

{

    NonConstClass() {}

};

// This will result in:

construct<ConstClass>()(ConstClass()); // constexpr expression

construct<NonConstClass>()(NonConstClass()); // not a constexpr expression

Thus qualifiers would be:

<nothing> (copy)

const

volatile

const volatile

const &

volatile &

const volatile &

const constexpr

volatile constexpr

const volatile constexpr

const constexpr &

volatile constexpr &

const volatile constexpr &

&&

Why the need? Because otherwise we can't have unique and syntactic sugar functions to handle all possibilities efficiently.

--

Phil Bouchard
Founder
C.: (819) 328-4743

Is there some reason to use assignment as your example? It seems a strange example, since we don't normally assign to const objects, nor build assignments that don't assign. 

Maybe a more motivating and less distracting example would be better.

Sent from my BlackBerry portable Babbage Device

From: Phil Bouchard via Std-Proposals Sent: Sunday, September 29, 2019 8:52 AM To: std-proposals@lists.isocpp.org Reply To: std-proposals@lists.isocpp.org Cc: Phil Bouchard Subject: [std-proposals] Template qualifiers

Greetings,

How many time did we have to redundantly write the same member functions just to add a const version? Ex.:

struct A

{

    A & operator = (A const &) { return * this; }

    A const & operator = (A const &) const { return * this; }

};

What I would like to suggest would be to templatize those instances as such:

struct A

{

    template <qualifier Q>

        A Q & operator = (A const &) Q { return * this; }

};

This would in instantiate on-the-fly the 2 (or 4) versions of the same code.

Thank you,

--

Phil Bouchard
Founder
C.: (819) 328-4743