[Bug c++/105841] New: Change in behavior of CTAD for alias templates

[Bug c++/105841] New: Change in behavior of CTAD for alias templates

[avemilia at protonmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

            Bug ID: 105841
           Summary: Change in behavior of CTAD for alias templates
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: avemilia at protonmail dot com
  Target Milestone: ---

Created attachment 53080
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53080&action=edit
repro

Since GCC 12, this code does not compile: https://godbolt.org/z/sWvh7P4co
The code used to compile on GCC 10 and 11. If this is incorrect code, please
explain what is wrong.

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=103852
      Known to work|                            |11.3.0
      Known to fail|                            |12.1.0, 13.0
   Last reconfirmed|                            |2022-06-06
                 CC|                            |jason at gcc dot gnu.org,
                   |                            |ppalka at gcc dot gnu.org
   Target Milestone|---                         |12.2
            Summary|Change in behavior of CTAD  |[12/13 Regression] Change
                   |for alias templates         |in behavior of CTAD for
                   |                            |alias templates
             Status|UNCONFIRMED                 |NEW
           Keywords|                            |rejects-valid
     Ever confirmed|0                           |1

--- Comment #1 from Patrick Palka <ppalka at gcc dot gnu.org> ---
We started rejecting the testcase after r12-7984-ga11f204e5a30d1.

Reduced rejects-valid example:

template<class T, int N>
struct A { A(...); };

template<class T, class... Ts>
A(T, Ts...) -> A<T, sizeof...(Ts)>;

template<class T, int N=0>
using B = A<T, N>;

B b(0, 0);

Removing B's default template argument makes us accept again.

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

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

--- Comment #2 from Jason Merrill <jason at gcc dot gnu.org> ---
In alias_ctad_tweaks, I wrote:

  /* This implementation differs from the [standard] in two significant ways:   

     1) We include all template parameters of A, not just some.                 
     2) The added constraint is same_type instead of deducible.                 

     I believe that while it's probably possible to construct a testcase that   
     behaves differently with this simplification, it should have the same      
     effect for real uses.  Including all template parameters means that we     
     deduce all parameters of A when resolving the call, so when we're in the   
     constraint we don't need to deduce them again, we can just check whether   
     the deduction produced the desired result.  */

This testcase is an example of why my assumption above was wrong.  The effect
of the transformations specified by the standard would be roughly

template <class> struct B_deducible;
template <class BT, int BN> struct B_deducible<B<BT,BN>> { };
template <class BT, class... fTs>
auto fp(BT, fTs...) -> A<BT, sizeof...(fTs)>
  requires requires { B_deducible<A<BT,sizeof...(fTs)>>(); };

decltype(fp(0,0)) b(0,0);

which works.

A year ago, Mike Spertus had a partial fix for this; any update on that?

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[mike at spertus dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #3 from Mike Spertus <mike at spertus dot com> ---
Thanks for the nudge, Jason. I will shake the bit rot off the POC and try to
polish it to something mergeable. OK if I ping you if I have questions?

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> ---
(In reply to Mike Spertus from comment #3)
> Thanks for the nudge, Jason. I will shake the bit rot off the POC and try to
> polish it to something mergeable. OK if I ping you if I have questions?

Absolutely!

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=102529

--- Comment #5 from Patrick Palka <ppalka at gcc dot gnu.org> ---
PR102529 seems related, though not a regression.  There it's a constraint (on
the alias template) rather than a default template argument that messes us up.

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

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

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #7 from Jason Merrill <jason at gcc dot gnu.org> ---
Created attachment 54444
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54444&action=edit
fix

Here's a patchset to implement the standard behavior plus the CWG2664
clarification.  Mike, does this look good to you?  Any additional testcases?

Also pushed to refs/users/jason/heads/alias-ctad in the git repository.
(https://gcc.gnu.org/git/?p=gcc.git;a=shortlog;h=refs/users/jason/heads/alias-ctad)

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[mike at spertus dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #8 from Mike Spertus <mike at spertus dot com> ---
Thanks, Jason! My course starts in 6 minutes, so I can't look at it now but
will give you feedback by 8:30AM tomorrow.

Mike

On Thu, Feb 9, 2023 at 3:07 PM jason at gcc dot gnu.org <
gcc-bugzilla@gcc.gnu.org> wrote:

> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841
>
> --- Comment #7 from Jason Merrill <jason at gcc dot gnu.org> ---
> Created attachment 54444
>   --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54444&action=edit
> fix
>
> Here's a patchset to implement the standard behavior plus the CWG2664
> clarification.  Mike, does this look good to you?  Any additional
> testcases?
>
> Also pushed to refs/users/jason/heads/alias-ctad in the git repository.
> (
> https://gcc.gnu.org/git/?p=gcc.git;a=shortlog;h=refs/users/jason/heads/alias-ctad
> )
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[mike at spertus dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #9 from Mike Spertus <mike at spertus dot com> ---
Hi Jason,
Very exciting. Some additional tests: Both versions of
https://godbolt.org/z/aM93PEWcz should be included in the tests. There are
two versions of the deduction guides in the godbolt. Either guide should
work. (They may already work, but it remains a good test). See
https://mspertus.github.io/MSVC_CTAD_1428672/ for more details.

In addition, the examples at https://godbolt.org/z/PjqMa8T35 should work.
Do you want versions of these that do not include standard library headers?

Thanks,
Mike

On Thu, Feb 9, 2023 at 3:26 PM <gcc-bugzilla@gcc.gnu.org> wrote:

> Attachments with a MIME type of "text/html" are not allowed on this
> installation.
>
> Michael Spertus wrote:
> > Thanks, Jason! My course starts in 6 minutes, so I can't look at it now
> but
> > will give you feedback by 8:30AM tomorrow.
> >
> > Mike
> >
> > On Thu, Feb 9, 2023 at 3:07 PM jason at gcc dot gnu.org <
> > gcc-bugzilla@gcc.gnu.org> wrote:
> >
> > > https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841
> > >
> > > --- Comment #7 from Jason Merrill <jason at gcc dot gnu.org> ---
> > > Created attachment 54444
> > >   --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54444&action=edit
> > > fix
> > >
> > > Here's a patchset to implement the standard behavior plus the CWG2664
> > > clarification.  Mike, does this look good to you?  Any additional
> > > testcases?
> > >
> > > Also pushed to refs/users/jason/heads/alias-ctad in the git repository.
> > > (
> > >
> https://gcc.gnu.org/git/?p=gcc.git;a=shortlog;h=refs/users/jason/heads/alias-ctad
> > > )
> > >
> > > --
> > > You are receiving this mail because:
> > > You are on the CC list for the bug.
>

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Jason Merrill <jason@gcc.gnu.org>:

https://gcc.gnu.org/g:148cbb153dafd6b21d83c00787acd430aec68a3d

commit r13-6553-g148cbb153dafd6b21d83c00787acd430aec68a3d
Author: Jason Merrill <jason@redhat.com>
Date:   Thu Feb 9 12:51:51 2023 -0800

    c++: add __is_deducible trait [PR105841]

    C++20 class template argument deduction for an alias template involves
    adding a constraint that the template arguments for the alias template can
    be deduced from the return type of the deduction guide for the underlying
    class template.  In the standard, this is modeled as defining a class
    template with a partial specialization, but it's much more efficient to
    implement with a trait that directly tries to perform the deduction.

    The first argument to the trait is a template rather than a type, so
various
    places needed to be adjusted to accommodate that.

            PR c++/105841

    gcc/ChangeLog:

            * doc/extend.texi (Type Traits):: Document __is_deducible.

    gcc/cp/ChangeLog:

            * cp-trait.def (IS_DEDUCIBLE): New.
            * cxx-pretty-print.cc (pp_cxx_trait): Handle non-type.
            * parser.cc (cp_parser_trait): Likewise.
            * tree.cc (cp_tree_equal): Likewise.
            * pt.cc (tsubst_copy_and_build): Likewise.
            (type_targs_deducible_from): New.
            (alias_ctad_tweaks): Use it.
            * semantics.cc (trait_expr_value): Handle CPTK_IS_DEDUCIBLE.
            (finish_trait_expr): Likewise.
            * constraint.cc (diagnose_trait_expr): Likewise.
            * cp-tree.h (type_targs_deducible_from): Declare.

    gcc/testsuite/ChangeLog:

            * g++.dg/ext/is_deducible1.C: New test.

[Bug c++/105841] [12/13 Regression] Change in behavior of CTAD for alias templates

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The trunk branch has been updated by Jason Merrill <jason@gcc.gnu.org>:

https://gcc.gnu.org/g:9e6170098d5e7756e85e880f8f4cb18e885a64fd

commit r13-6555-g9e6170098d5e7756e85e880f8f4cb18e885a64fd
Author: Jason Merrill <jason@redhat.com>
Date:   Wed Feb 8 22:06:22 2023 -0800

    c++: fix alias CTAD [PR105841]

    In my initial implementation of alias CTAD, I described a couple of
    differences from the specification that I thought would not have a
practical
    effect; this testcase demonstrates that I was wrong.  One difference is
    resolved by the CPTK_IS_DEDUCIBLE commit; the other (adding too many of the
    alias template parameters to the new deduction guide) is fixed by this
    patch.

            PR c++/105841

    gcc/cp/ChangeLog:

            * pt.cc (corresponding_template_parameter_list): Split out...
            (corresponding_template_parameter): ...from here.
            (find_template_parameters): Factor out...
            (find_template_parameter_info::find_in): ...this function.
            (find_template_parameter_info::find_in_recursive): New.
            (find_template_parameter_info::found): New.
            (alias_ctad_tweaks): Only add parms used in the deduced args.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp2a/class-deduction-alias14.C: New test.

    Co-authored-by: Michael Spertus <mike@spertus.com>

[Bug c++/105841] [12 Regression] Change in behavior of CTAD for alias templates

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |13.0
            Summary|[12/13 Regression] Change   |[12 Regression] Change in
                   |in behavior of CTAD for     |behavior of CTAD for alias
                   |alias templates             |templates
      Known to fail|13.0                        |

--- Comment #12 from Jason Merrill <jason at gcc dot gnu.org> ---
Fixed for GCC 13 so far.

[Bug c++/105841] [12 Regression] Change in behavior of CTAD for alias templates

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|12.3                        |12.4

--- Comment #13 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 12.3 is being released, retargeting bugs to GCC 12.4.

[Bug c++/105841] [12 Regression] Change in behavior of CTAD for alias templates

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|12.4                        |13.0
         Resolution|---                         |FIXED
             Status|ASSIGNED                    |RESOLVED

--- Comment #14 from Jason Merrill <jason at gcc dot gnu.org> ---
This was a big enough change that I lean toward not backporting.

[Bug c++/105841] [12 Regression] Change in behavior of CTAD for alias templates

[hokein.wu at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

Haojian Wu <hokein.wu at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |hokein.wu at gmail dot com

--- Comment #15 from Haojian Wu <hokein.wu at gmail dot com> ---
Hi, I notice that the __is_deducible was hidden in the commit
https://github.com/gcc-mirror/gcc/commit/30556bf81f4385c2a9c449948865dbcf35664764.

Is there any reason behind the change?

(Context: I'm implementing the is_deducible part for alias CTAD in clang, and
I'm considering to add a similar builtin in clang).

[Bug c++/105841] [12 Regression] Change in behavior of CTAD for alias templates

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #16 from Patrick Palka <ppalka at gcc dot gnu.org> ---
(In reply to Haojian Wu from comment #15)
> Hi, I notice that the __is_deducible was hidden in the commit
> https://github.com/gcc-mirror/gcc/commit/
> 30556bf81f4385c2a9c449948865dbcf35664764.
> 
> Is there any reason behind the change?
> 
> (Context: I'm implementing the is_deducible part for alias CTAD in clang,
> and I'm considering to add a similar builtin in clang).
IIRC we didn't want to commit to an API for the built-in, and we also didn't
have any motivating use cases for the it within libstdc++.

[Bug c++/105841] [12 Regression] Change in behavior of CTAD for alias templates

[hokein.wu at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=105841

--- Comment #17 from Haojian Wu <hokein.wu at gmail dot com> ---
> IIRC we didn't want to commit to an API for the built-in, and we also didn't have any motivating use cases for the it within libstdc++.

Thanks for the reply. Fair enough.