[Bug c++/105191] New: '<expression>' "is not a constant expression" regression in GCC 12

[Bug c++/105191] New: '<expression>' "is not a constant expression" regression in GCC 12

[strager.nds at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

            Bug ID: 105191
           Summary: '<expression>' "is not a constant expression"
                    regression in GCC 12
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: strager.nds at gmail dot com
  Target Milestone: ---

Created attachment 52763
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52763&action=edit
C++ source file which demonstrates the issue

Compared to GCC 10 and GCC 11.2, GCC 12 reports an error when compiling my
program. A reduced test case is attached (test.cpp). I think the error is a
false positive; I think my code is valid C++.

g++
(Compiler-Explorer-Build-gcc-d8ac63e4c638a814c2cb7a9f4cebb6606caa4985-binutils-2.36.1)
12.0.1 20220404 (experimental):
g++ (GCC) 12.0.1 20220401 (Red Hat 12.0.1-0):
$ /usr/bin/c++ -std=gnu++11 test.cpp -c
test.cpp:17:5: error: non-constant array initialization
   17 |   D{},
      |     ^
test.cpp:18:1: error: call to non-‘constexpr’ function ‘C::C()’
   18 | };
      | ^
test.cpp:7:8: note: ‘C::C()’ is not usable as a ‘constexpr’ function because:
    7 | struct C {
      |        ^
test.cpp:9:5: note: defaulted default constructor does not initialize ‘B C::b’
    9 |   B b;
      |     ^
$ /usr/bin/c++  -std=gnu++20 test.cpp -c
test.cpp:18:1: error: ‘D [1]{D{C [1]{C{A{((const char*)"")}}}}}’ is not a
constant expression
   18 | };
      | ^

g++-10 (Ubuntu 10.3.0-1ubuntu1~20.04) 10.3.0:
g++ (Compiler-Explorer-Build-gcc--binutils-2.36.1) 11.2.0:
$ g++-10 -std=gnu++11 test.cpp -c
(no error)
$ g++-10 -std=gnu++20 test.cpp -c
(no error)

[Bug c++/105191] [12 Regression] '<expression>' "is not a constant expression" regression in GCC 12

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
                 CC|                            |jason at gcc dot gnu.org,
                   |                            |ppalka at gcc dot gnu.org
      Known to work|                            |10.3.0, 11.2.0
      Known to fail|                            |12.0
            Summary|'<expression>' "is not a    |[12 Regression]
                   |constant expression"        |'<expression>' "is not a
                   |regression in GCC 12        |constant expression"
                   |                            |regression in GCC 12
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=92385
   Last reconfirmed|                            |2022-04-07
   Target Milestone|---                         |12.0
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Confirmed, started with r12-6326-ge948436eab818c

[Bug c++/105191] [12 Regression] '<expression>' "is not a constant expression" regression in GCC 12

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jakub at gcc dot gnu.org
           Priority|P3                          |P1

[Bug c++/105191] [12 Regression] '<expression>' "is not a constant expression" regression in GCC 12

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

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

[Bug c++/105191] [12 Regression] '<expression>' "is not a constant expression" regression in GCC 12

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

--- Comment #2 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:4822108e61ab879067482704f2f7d1670813d61a

commit r12-8066-g4822108e61ab879067482704f2f7d1670813d61a
Author: Jason Merrill <jason@redhat.com>
Date:   Fri Apr 8 15:33:41 2022 -0400

    c++: constexpr non-trivial aggregate init [PR105191]

    My patch for PR92385 made us use VEC_INIT_EXPR for aggregate initialization
    of an array where some elements are not explicitly initialized.  Constexpr
    handling of that was treating initialization from {} as equivalent to
    value-initialization, which is problematic for classes with default member
    initializers that make the default constructor non-trivial; in older
    standard modes, not initializing all members makes a constructor
    non-constexpr, but aggregate initialization is fine.

            PR c++/105191
            PR c++/92385

    gcc/cp/ChangeLog:

            * tree.cc (build_vec_init_elt): Do {}-init for aggregates.
            * constexpr.cc (cxx_eval_vec_init): Only treat {} as value-init
            for non-aggregate types.
            (build_vec_init_expr): Also check constancy of explicit
            initializer elements.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/constexpr-array28.C: New test.

[Bug c++/105191] [12 Regression] '<expression>' "is not a constant expression" regression in GCC 12

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

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

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

--- Comment #3 from Jason Merrill <jason at gcc dot gnu.org> ---
Fixed.

[Bug c++/105191] [12 Regression] '<expression>' "is not a constant expression" regression in GCC 12

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

--- Comment #4 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:ce46d6041358052dfa26f3720732f0357c5d72e7

commit r13-463-gce46d6041358052dfa26f3720732f0357c5d72e7
Author: Jason Merrill <jason@redhat.com>
Date:   Fri May 13 16:07:10 2022 -0400

    c++: array {}-init [PR105589]

    My patch for 105191 made us use build_value_init more frequently from
    build_vec_init_expr, but build_value_init doesn't like to be called to
    initialize a class in a template.  That's caused trouble in the past, and
    seems like a strange restriction, so let's fix it.

            PR c++/105589
            PR c++/105191
            PR c++/92385

    gcc/cp/ChangeLog:

            * init.cc (build_value_init): Handle class in template.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/initlist-array16.C: New test.

[Bug c++/105191] [12 Regression] '<expression>' "is not a constant expression" regression in GCC 12

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105191

--- Comment #5 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Jason Merrill
<jason@gcc.gnu.org>:

https://gcc.gnu.org/g:40f749b364b740f41ea6b211f81c21919a2e8bee

commit r12-8380-g40f749b364b740f41ea6b211f81c21919a2e8bee
Author: Jason Merrill <jason@redhat.com>
Date:   Fri May 13 16:07:10 2022 -0400

    c++: array {}-init [PR105589]

    My patch for 105191 made us use build_value_init more frequently from
    build_vec_init_expr, but build_value_init doesn't like to be called to
    initialize a class in a template.  That's caused trouble in the past, and
    seems like a strange restriction, so let's fix it.

            PR c++/105589
            PR c++/105191
            PR c++/92385

    gcc/cp/ChangeLog:

            * init.cc (build_value_init): Handle class in template.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/initlist-array16.C: New test.