[Bug c++/96999] New: regression: std::variant requires default constructor with gcc 8-9, gcc 7.x and 10.x works

[Bug c++/96999] New: regression: std::variant requires default constructor with gcc 8-9, gcc 7.x and 10.x works

[dushistov at mail dot ru]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96999

            Bug ID: 96999
           Summary: regression: std::variant requires default constructor
                    with gcc 8-9, gcc 7.x and 10.x works
           Product: gcc
           Version: 9.3.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: dushistov at mail dot ru
  Target Milestone: ---

For code bellow, gcc 9.3 from Ubuntu reports error:
```
In file included from /usr/include/c++/9/variant:36,
                 from test.cpp:1:
...
/usr/include/c++/9/type_traits:931:47: error: ‘Foo::Foo()’ is private within
this context                                                                    
  931 |     : public integral_constant<bool, noexcept(_Tp())>                   
      |                                               ^~~~~                     
test.cpp:10:3: note: declared private here                                      
   10 |   Foo() noexcept {}                                                     
      |   ^~~                       
```
gcc 8.x also reports error. gcc 7.x and 10.x works just fine:

```
#include <variant>
#include <string>

struct Foo {
public:
  explicit Foo(int) noexcept {}
  Foo(Foo &&) noexcept = default;
  Foo &operator=(Foo &&) = default;
private:
  Foo() noexcept {}
};

struct Boo {
public:
  explicit Boo(int) noexcept {}
  Boo(Boo &&) noexcept = default;
  Boo &operator=(Boo &&) = default;
private:
  Boo() noexcept {}
};


template<bool X>
std::variant<Foo, Boo> g(int v, int x) {
 return  v == 0 ? std::variant<Foo, Boo>{Foo{x}} :
                                 std::variant<Foo, Boo>{Boo{x}};
}

int main()
{
  std::variant<std::variant<Foo, Boo>, std::string> err{std::string("aaa")};
}
```

[Bug libstdc++/96999] [8/9 Regression] std::variant requires default constructor with gcc 8-9, gcc 7.x and 10.x works

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96999

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |8.4.0, 9.3.0
   Last reconfirmed|                            |2020-09-09
             Status|UNCONFIRMED                 |ASSIGNED
           Keywords|                            |rejects-valid
      Known to work|                            |10.1.0, 11.0, 7.4.0
     Ever confirmed|0                           |1
   Target Milestone|---                         |8.5
          Component|c++                         |libstdc++
           Assignee|unassigned at gcc dot gnu.org      |redi at gcc dot gnu.org
            Summary|regression: std::variant    |[8/9 Regression]
                   |requires default            |std::variant requires
                   |constructor with gcc 8-9,   |default constructor with
                   |gcc 7.x and 10.x works      |gcc 8-9, gcc 7.x and 10.x
                   |                            |works

--- Comment #1 from Jonathan Wakely <redi at gcc dot gnu.org> ---
This seems to be a similar problem to PR 94033, and is fixed by its r10-7270

[Bug libstdc++/96999] [8/9 Regression] std::variant requires default constructor with gcc 8-9, gcc 7.x and 10.x works

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96999

--- Comment #2 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Jonathan Wakely
<redi@gcc.gnu.org>:

https://gcc.gnu.org/g:6d06c17e9f470f4a2d7e41ca6f3a8b36c729dc6a

commit r9-8860-g6d06c17e9f470f4a2d7e41ca6f3a8b36c729dc6a
Author: Jonathan Wakely <jwakely@redhat.com>
Date:   Wed Mar 18 23:19:12 2020 +0000

    libstdc++: Fix is_trivially_constructible (PR 94033)

    This attempts to make is_nothrow_constructible more robust (and
    efficient to compile) by not depending on is_constructible. Instead the
    __is_constructible intrinsic is used directly. The helper class
    __is_nt_constructible_impl which checks whether the construction is
    non-throwing now takes a bool template parameter that is substituted by
    the result of the instrinsic. This fixes the reported bug by not using
    the already-instantiated (and incorrect) value of std::is_constructible.
    I don't think it really fixes the problem in general, because
    std::is_nothrow_constructible itself could already have been
    instantiated in a context where it gives the wrong result. A proper fix
    needs to be done in the compiler.

    Backported to the gcc-8 and gcc-9 branches to fix PR 96999.

            PR libstdc++/94033
            * include/std/type_traits (__is_nt_default_constructible_atom):
Remove.
            (__is_nt_default_constructible_impl): Remove.
            (__is_nothrow_default_constructible_impl): Remove.
            (__is_nt_constructible_impl): Add bool template parameter. Adjust
            partial specializations.
            (__is_nothrow_constructible_impl): Replace class template with
alias
            template.
            (is_nothrow_default_constructible): Derive from alias template
            __is_nothrow_constructible_impl instead of
            __is_nothrow_default_constructible_impl.
            * testsuite/20_util/is_nothrow_constructible/94003.cc: New test.

    (cherry picked from commit b3341826531e80e02f194460b4fbe1b0541c0463)

[Bug libstdc++/96999] [8/9 Regression] std::variant requires default constructor with gcc 8-9, gcc 7.x and 10.x works

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96999

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-8 branch has been updated by Jonathan Wakely
<redi@gcc.gnu.org>:

https://gcc.gnu.org/g:1521433e2b6bff583b227fe391fe9dcf02f6000e

commit r8-10445-g1521433e2b6bff583b227fe391fe9dcf02f6000e
Author: Jonathan Wakely <jwakely@redhat.com>
Date:   Wed Mar 18 23:19:12 2020 +0000

    libstdc++: Fix is_trivially_constructible (PR 94033)

    This attempts to make is_nothrow_constructible more robust (and
    efficient to compile) by not depending on is_constructible. Instead the
    __is_constructible intrinsic is used directly. The helper class
    __is_nt_constructible_impl which checks whether the construction is
    non-throwing now takes a bool template parameter that is substituted by
    the result of the instrinsic. This fixes the reported bug by not using
    the already-instantiated (and incorrect) value of std::is_constructible.
    I don't think it really fixes the problem in general, because
    std::is_nothrow_constructible itself could already have been
    instantiated in a context where it gives the wrong result. A proper fix
    needs to be done in the compiler.

    Backported to the gcc-8 and gcc-9 branches to fix PR 96999.

            PR libstdc++/94033
            * include/std/type_traits (__is_nt_default_constructible_atom):
Remove.
            (__is_nt_default_constructible_impl): Remove.
            (__is_nothrow_default_constructible_impl): Remove.
            (__is_nt_constructible_impl): Add bool template parameter. Adjust
            partial specializations.
            (__is_nothrow_constructible_impl): Replace class template with
alias
            template.
            (is_nothrow_default_constructible): Derive from alias template
            __is_nothrow_constructible_impl instead of
            __is_nothrow_default_constructible_impl.
            * testsuite/20_util/is_nothrow_constructible/94003.cc: New test.

    (cherry picked from commit b3341826531e80e02f194460b4fbe1b0541c0463)

[Bug libstdc++/96999] [8/9 Regression] std::variant requires default constructor with gcc 8-9, gcc 7.x and 10.x works

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96999

Jonathan Wakely <redi at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED

--- Comment #4 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Should be fixed for 8.5 and 9.4, thanks for the report.