[Bug c++/114395] New: std::is_constructible_v result of const reference incorrect

[Bug c++/114395] New: std::is_constructible_v result of const reference incorrect

[yronglin777 at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

            Bug ID: 114395
           Summary: std::is_constructible_v result of const reference
                    incorrect
           Product: gcc
           Version: 14.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: yronglin777 at gmail dot com
  Target Milestone: ---

https://godbolt.org/z/h5af6f5x9
```
#include <type_traits>

struct A {
    int a;
};
struct B : public A {};

static_assert(!std::is_constructible_v<const B &, const A &>);
static_assert(!std::is_constructible_v<const B *, const A *>);

```
Clang and MSVC accepted this code, but gcc rejected.

```
<source>:8:15: error: static assertion failed
    8 | static_assert(!std::is_constructible_v<const B &, const A &>);
      |               ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Compiler returned: 1
```

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|std::is_constructible_v     |[c++20+]
                   |result of const reference   |std::is_constructible_v
                   |incorrect                   |result of const reference
                   |                            |incorrect

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Hmm, this fails only for C++20+.

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Except this works with C++20+ for GCC:
```
struct A {
    int a;
};
struct B : public A {};

void f(const A &a)
{
        typedef const B & Btype;
        Btype b(a);
}

```

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mpolacek at gcc dot gnu.org

--- Comment #3 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Changed with r10-5137-g43aae289866f5e:

commit 43aae289866f5ea55d187444520412554aa2e171
Author: Marek Polacek <polacek@redhat.com>
Date:   Tue Dec 3 15:59:40 2019 +0000

    PR c++/91363 - P0960R3: Parenthesized initialization of aggregates.

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Marek Polacek from comment #3)
> Changed with r10-5137-g43aae289866f5e:
> 
> commit 43aae289866f5ea55d187444520412554aa2e171
> Author: Marek Polacek <polacek@redhat.com>
> Date:   Tue Dec 3 15:59:40 2019 +0000
> 
>     PR c++/91363 - P0960R3: Parenthesized initialization of aggregates.

Hmm, reading the paper, I am not 100% sure if GCC implemented this correctly
due to the wording of the design principles:
```
Any existing meaning of A(b) should not change.
```

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
I disagree, A(b) definitely changed meaning. It now initializes A by copying
the base part of b

This is a combination of aggregates being allowed to have base classes and
paren-init doing aggregate-init.

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

--- Comment #6 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Right, another design principle is that () should work where {} works, and
const B &b{a}; works.  A(b) previously didn't work so it's not really changing
meaning.  So not a bug IMHO.

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> ---
All of Clang, MSVC and EDG disagree with GCC though, so maybe there's a DR that
we're missing.

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

--- Comment #8 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Jonathan Wakely from comment #7)
> All of Clang, MSVC and EDG disagree with GCC though, so maybe there's a DR
> that we're missing.

Maybe https://cplusplus.github.io/CWG/issues/1467.html (again).
See clang bug https://github.com/llvm/llvm-project/pull/77768 .

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

--- Comment #9 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
https://eel.is/c++draft/dcl.init.general#16.2 applies here I think. Which comes
before the `Parenthesized aggregates initialization` clause. But I could be
wrong.

https://github.com/cplusplus/papers/issues/1494 is kinda of related.

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[de34 at live dot cn]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

Jiang An <de34 at live dot cn> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |de34 at live dot cn

--- Comment #10 from Jiang An <de34 at live dot cn> ---
I think this is basically CWG2709
(https://cplusplus.github.io/CWG/issues/2709.html).
Given CWG2709 is closed as NAD, we should reject such initializations.

[Bug c++/114395] [c++20+] std::is_constructible_v result of const reference incorrect

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114395

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
           Assignee|unassigned at gcc dot gnu.org      |mpolacek at gcc dot gnu.org
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2024-03-20