[Bug c++/109065] New: [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[Bug c++/109065] New: [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[marcel.jacobse at ewetel dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

            Bug ID: 109065
           Summary: [11/12/13 Regression] Type alias combination
                    erroneously fails with "incomplete type"
           Product: gcc
           Version: 11.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: marcel.jacobse at ewetel dot net
  Target Milestone: ---

Starting with gcc 11.1, the code

template <class T> using DataAlias = int;
template <class T> struct MyUniquePtr {};

template <class T> struct Test {
    using Data = DataAlias<T>;
    MyUniquePtr<Data[]> d;
};
Test<int> test;

fails to compile with

<source>: In instantiation of 'struct Test<int>':
<source>:8:11:   required from here
<source>:6:25: error: 'Test<T>::d' has incomplete type
    6 |     MyUniquePtr<Data[]> d;
      |                         ^
<source>:2:27: note: declaration of 'struct MyUniquePtr<int []>'
    2 | template <class T> struct MyUniquePtr {};
      |                           ^~~~~~~~~~~

See for example: https://godbolt.org/z/oq9TjqKrW
I believe this to be valid C++11 through C++20 and it seems to compile just
fine with clang, MSVC and any gcc <= 10.4 that supports C++11.

A range of minor changes make it compile again, for example:
* Make DataAlias actually use T in any way, e.g.
https://godbolt.org/z/cb9dW1E6s
* Don't use an array type for d, i.e. https://godbolt.org/z/MGKPcjEz3
* Don't use the second alias level for d, i.e. https://godbolt.org/z/KaWq9Pfq8

[Bug c++/109065] [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |needs-bisection,
                   |                            |rejects-valid
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2023-03-08
   Target Milestone|---                         |11.4
     Ever confirmed|0                           |1
      Known to fail|                            |11.1.0
      Known to work|                            |10.4.0

--- Comment #1 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.

[Bug c++/109065] [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
>* Don't use the second alias level for d, i.e. https://godbolt.org/z/KaWq9Pfq8

Using a typedef instead of alias still fails.
That is:
```
template <class T> using DataAlias = int;
template <class T> struct MyUniquePtr {};

template <class T> struct Test {
    typedef DataAlias<T> Data;
    MyUniquePtr<Data[1]> d;
};
Test<int> test;
```
Still fails.

[Bug c++/109065] [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

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

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

[Bug c++/109065] [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nathan at gcc dot gnu.org,
                   |                            |ppalka at gcc dot gnu.org
           Keywords|needs-bisection             |

--- Comment #3 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Started with r11-5825-gffb268ffcf9f21

[Bug c++/109065] [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

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

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

[Bug c++/109065] [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

--- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> ---
The problem is that r11-5825 assumes that if a type is dependent, its
TYPE_MAIN_VARIANT and TYPE_CANONICAL are also dependent, which is not
universally true: DataAlias<T> is dependent, so that any errors from
substituting into the template-id are caught; but its TYPE_MAIN_VARIANT and
TYPE_CANONICAL are int, which is not.

[Bug c++/109065] [11/12/13 Regression] Type alias combination erroneously fails with "incomplete type"

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

--- Comment #5 from Jason Merrill <jason at gcc dot gnu.org> ---
Created attachment 54665
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54665&action=edit
first try

This fixes the PR, but breaks the modules tests that the earlier patch fixed. 
I guess we need a different fix for the modules issue.

[Bug c++/109065] [11/12/13/14 Regression] Type alias combination erroneously fails with "incomplete type"

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=109065

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|11.4                        |11.5

--- Comment #6 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 11.4 is being released, retargeting bugs to GCC 11.5.