[Bug c++/101988] New: Accepts invalid new-expression of array of deduced class template

[Bug c++/101988] New: Accepts invalid new-expression of array of deduced class template

[ed at catmur dot uk]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

            Bug ID: 101988
           Summary: Accepts invalid new-expression of array of deduced
                    class template
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: ed at catmur dot uk
  Target Milestone: ---

template<class T> struct A { A(T); };
auto p = new A[]{1};

g++
(Compiler-Explorer-Build-gcc-f0fca213bc52644ba896da622b35842a6157bd98-binutils-2.36.1)
12.0.0 20210818 (experimental) accepts (it allocates and constructs a single
`A<int>`.

gcc 11.2 and below rejects, correctly: error: creating array of 'A<...auto...>'

Per [expr.new]/2 the invented declaration should be:
A[] x{1};
which is ill-formed.

nb. clang ICEs on this (https://bugs.llvm.org/show_bug.cgi?id=51547)

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2021-08-19
     Ever confirmed|0                           |1
            Summary|Accepts invalid             |[12 Regression] Accepts
                   |new-expression of array of  |invalid new-expression of
                   |deduced class template      |array of deduced class
                   |                            |template
   Target Milestone|---                         |12.0
           Keywords|                            |accepts-invalid

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed, note doing [1] instead of [] causes GCC to reject (and clang not to
ICE either).

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
I also noticed, GCC accepts while clang rejects:
template<class T> struct A { A(T); };
auto p = new A<int>[]{A(1),A(1)};

While changing it to [2], both GCC and clang accepts it.

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #2)
> I also noticed, GCC accepts while clang rejects:
> template<class T> struct A { A(T); };
> auto p = new A<int>[]{A(1),A(1)};
> 
> While changing it to [2], both GCC and clang accepts it.

Note ICC and MSVC accepts the above so I think it is a clang bug.


Oh and GCC rejects this:
template<class T> struct A { A(T); };
auto p = new A[]{1,2};

But thinks A[] is really A:
<source>:2:21: error: class template argument deduction failed:
    2 | auto p = new A[]{1,2};
      |                     ^
<source>:2:21: error: no matching function for call to 'A(int, int)'
<source>:1:30: note: candidate: 'template<class T> A(T)-> A<T>'
    1 | template<class T> struct A { A(T); };
      |                              ^
<source>:1:30: note:   template argument deduction/substitution failed:
<source>:2:21: note:   candidate expects 1 argument, 2 provided
    2 | auto p = new A[]{1,2};
      |                     ^
<source>:1:26: note: candidate: 'template<class T> A(A<T>)-> A<T>'
    1 | template<class T> struct A { A(T); };
      |                          ^
<source>:1:26: note:   template argument deduction/substitution failed:
<source>:2:21: note:   mismatched types 'A<T>' and 'int'
    2 | auto p = new A[]{1,2};
      |                     ^

So that might get the hint of the problem.

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Maybe related to PR 82110 but there has been other CTAD changes in GCC 12 too.

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

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

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

--- Comment #5 from Jonathan Wakely <redi at gcc dot gnu.org> ---
Started to be accepted with r12-1933:

c++: DR2397 - auto specifier for * and & to arrays [PR100975]

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

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

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

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P1

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Marek Polacek <mpolacek@gcc.gnu.org>:

https://gcc.gnu.org/g:0c0f453c4af4880c522c8472c33eef42bee9eda1

commit r12-6894-g0c0f453c4af4880c522c8472c33eef42bee9eda1
Author: Marek Polacek <polacek@redhat.com>
Date:   Wed Jan 26 17:29:19 2022 -0500

    c++: new-expr of array of deduced class tmpl [PR101988]

    In r12-1933 I attempted to implement DR2397 aka allowing

      int a[3];
      auto (&r)[3] = a;

    by removing the type_uses_auto check in create_array_type_for_decl.
    That may have gone too far, because it also allows arrays of
    CLASS_PLACEHOLDER_TEMPLATE and it looks like [dcl.type.class.deduct]
    prohibits that: "...the declared type of the variable shall be cv T,
    where T is the placeholder."  However, in /2 it explicitly states that
    "A placeholder for a deduced class type can also be used in the
    type-specifier-seq in the new-type-id or type-id of a new-expression."

    In this PR, it manifested by making us accept invalid

      template<class T> struct A { A(T); };
      auto p = new A[]{1};

    [expr.new]/2 says that such a construct is treated as an invented
    declaration of the form

      A x[]{1};

    but, I think, that ought to be ill-formed as per above.  So this patch
    sort of restores the create_array_type_for_decl check.  I should mention
    that the difference between [] and [1] is due to cp_parser_new_type_id:

          if (*nelts == NULL_TREE)
            /* Leave [] in the declarator.  */;

    and groktypename returning different types based on that.

            PR c++/101988

    gcc/cp/ChangeLog:

            * decl.cc (create_array_type_for_decl): Reject forming an array of
            placeholder for a deduced class type.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp1z/class-deduction-new1.C: New test.
            * g++.dg/cpp23/auto-array2.C: New test.

[Bug c++/101988] [12 Regression] Accepts invalid new-expression of array of deduced class template

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=101988

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

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

--- Comment #7 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Should be fixed.  I've also added pretty much all of the other testcases in
this PR.

> Note ICC and MSVC accepts the above so I think it is a clang bug.

I think so too.