[Bug c++/103852] New: Alias template argument deduction is available in C++17 mode

[Bug c++/103852] New: Alias template argument deduction is available in C++17 mode

[fchelnokov at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

            Bug ID: 103852
           Summary: Alias template argument deduction is available in
                    C++17 mode
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: fchelnokov at gmail dot com
  Target Milestone: ---

This is a valid C++20 code:
```
#include <set>

template<class Key, class Compare = std::less<Key>, class Allocator =
std::allocator<Key>>
using Set = std::set<Key, Compare, Allocator>;

int main() {
    Set set2 = { 1, 2, 3 };
}
```

But in C++17 mode it shall be an error, since alias template argument deduction
is not available there. Still GCC wrongly accepts the code in C++17 mode. Demo:
https://gcc.godbolt.org/z/f5WbWdPqb

Related discussion: https://stackoverflow.com/q/68768937/7325599

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |accepts-invalid
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
      Known to fail|                            |10.1.0
      Known to work|                            |9.1.0
   Last reconfirmed|                            |2021-12-28
   Target Milestone|---                         |10.4
            Summary|Alias template argument     |[10/11/12 Regression] Alias
                   |deduction is available in   |template argument deduction
                   |C++17 mode                  |is available in C++17 mode

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed reduced testcase:
template <class T> struct b{};
template <class T, class T1 = b<T>>
struct s
{
    s(T);
};
s c(100);
template <class T, class T1 = b<T>>
using ss = s<T, T1>;
ss tt(1);

---- CUT ----
tt defintion should fail in C++17 while pass in C++20.


It looks like when the support was added it was added unconditionally; could
there have been a Defect report?

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

Patrick Palka <ppalka at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ppalka at gcc dot gnu.org
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=92060

--- Comment #2 from Patrick Palka <ppalka at gcc dot gnu.org> ---
This happens because GCC implements the proposed resolution of CWG1286, which
makes simple alias templates such as Set equivalent to the underlying class
template.  So the CTAD is not alias CTAD but ordinary CTAD.

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

Jason Merrill <jason at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |jason at gcc dot gnu.org
             Status|NEW                         |ASSIGNED
                 CC|                            |jason at gcc dot gnu.org

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

--- Comment #3 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jason Merrill <jason@gcc.gnu.org>:

https://gcc.gnu.org/g:a11f204e5a30d1dfd2508e3f3014509a0342f604

commit r12-7984-ga11f204e5a30d1dfd2508e3f3014509a0342f604
Author: Jason Merrill <jason@redhat.com>
Date:   Fri Apr 1 15:55:21 2022 -0400

    c++: alias-tmpl equivalence and default targs [PR103852]

    The suggested resolution for CWG1286, which we implemented, ignores default
    template arguments, but this PR is an example of why that doesn't make
    sense: the templates aren't functionally equivalent.

            PR c++/103852
            DR 1286

    gcc/cp/ChangeLog:

            * pt.cc (get_underlying_template): Compare default template args.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/alias-decl-dr1286a.C: Default args now matter.
            * g++.dg/cpp1z/class-deduction-alias1.C: New test.

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

Jason Merrill <jason at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|10.4                        |12.0

--- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> ---
The patch doesn't affect the testcases, because there the templates are still
considered equivalent.

Though now I notice that the suggested resolution of DR1286 only considers
template equivalence in the context of comparing template arguments; it isn't
saying that they're interchangeable in all contexts.  So perhaps we should stop
using that equivalence in this case.

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jason Merrill <jason@gcc.gnu.org>:

https://gcc.gnu.org/g:5d583d24200ac33a1cb281b7ddee9aad351151b5

commit r12-7998-g5d583d24200ac33a1cb281b7ddee9aad351151b5
Author: Jason Merrill <jason@redhat.com>
Date:   Mon Apr 4 11:56:38 2022 -0400

    c++: alias template equivalence and CTAD [PR103852]

    I had been thinking about DR1286 "equivalence" as meaning generally
    interchangeable, but looking back at the proposed resolution in the context
    of this PR, I see that it's just about use as a template argument.  So
let's
    give a pedwarn if we look through a renaming alias.

            PR c++/103852
            DR 1286

    gcc/cp/ChangeLog:

            * pt.cc (do_class_deduction): Pedwarn for renaming alias in C++17.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp1z/class-deduction-alias1.C: Expect warning.

[Bug c++/103852] [10/11/12 Regression] Alias template argument deduction is available in C++17 mode

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103852

Jason Merrill <jason at gcc dot gnu.org> changed:

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

--- Comment #6 from Jason Merrill <jason at gcc dot gnu.org> ---
GCC 12 will give a pedwarn about the testcase.  Not backporting.