[Bug c++/96840] New: Recursive substitution in constrained commutative operator

[Bug c++/96840] New: Recursive substitution in constrained commutative operator

[johelegp at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

            Bug ID: 96840
           Summary: Recursive substitution in constrained commutative
                    operator
           Product: gcc
           Version: 11.0
               URL: https://godbolt.org/z/cedacs
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: johelegp at gmail dot com
  Target Milestone: ---

I may be wrong, but I think this should compile. Clang accepts it as does GCC
10.2 and did the GCC 11 build from a little while ago.
See the URL:
```C++
template <class T, class U> concept C = requires(T t, U u) { t * u; };
template <class Rep> struct Int {
     template <class T> requires C<T, Rep> friend void operator*(T, Int) { }
     template <class T> requires C<T, Rep> friend void operator*(Int, T) { }
};
void f() { 0 * Int<int>{}; }
```

[Bug c++/96840] [11 Regression] Recursive substitution in constrained commutative operator

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |rejects-valid
   Target Milestone|---                         |11.0
            Summary|Recursive substitution in   |[11 Regression] Recursive
                   |constrained commutative     |substitution in constrained
                   |operator                    |commutative operator

[Bug c++/96840] [11 Regression] Recursive substitution in constrained commutative operator

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jason at gcc dot gnu.org,
                   |                            |ppalka at gcc dot gnu.org

--- Comment #1 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Started with r11-2774:

    c++: Check satisfaction before non-dep convs. [CWG2369]

    It's very hard to use concepts to protect a template from hard errors due
to
    unwanted instantiation if constraints aren't checked until after doing all
    substitution and checking of non-dependent conversions.

We fall into a loop when checking the constraints of the second overload
(instantiated with Rep=int and T=Int<int>).

It looks like we're correct to reject the testcase as of this DR?

[Bug c++/96840] [11 Regression] Recursive substitution in constrained commutative operator

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
     Ever confirmed|0                           |1
   Last reconfirmed|                            |2020-10-16

[Bug c++/96840] [11 Regression] Recursive substitution in constrained commutative operator

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

--- Comment #2 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Patrick Palka <ppalka@gcc.gnu.org>:

https://gcc.gnu.org/g:79f57d5cb070bb02ea0a34b5f42658d6659b19a8

commit r11-6245-g79f57d5cb070bb02ea0a34b5f42658d6659b19a8
Author: Patrick Palka <ppalka@redhat.com>
Date:   Thu Dec 17 22:18:07 2020 -0500

    c++: Diagnose self-recursive satisfaction

    This patch further extends the satisfaction_cache class to diagnose
    self-recursive satisfaction.

    gcc/cp/ChangeLog:

            * constraint.cc (sat_entry::evaluating): New member.
            (satisfaction_cache::get): If entry->evaluating, diagnose
            self-recursive satisfaction.  Otherwise, set entry->evaluating
            if we're not reusing a cached satisfaction result.
            (satisfaction_cache::save): Clear entry->evaluating.
            (satisfy_atom): Set up diagnosing_failed_constraint before the
            first call to get().

    gcc/testsuite/ChangeLog:

            PR c++/96840
            * g++.dg/cpp2a/concepts-pr88395.C: Adjust to expect the
            self-recursive satisfaction to get directly diagnosed.
            * g++.dg/cpp2a/concepts-recursive-sat2.C: Likewise.
            * g++.dg/cpp2a/concepts-recursive-sat4.C: New test.

[Bug c++/96840] [11 Regression] Recursive substitution in constrained commutative operator

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

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

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

--- Comment #3 from Patrick Palka <ppalka at gcc dot gnu.org> ---
With the above patch we now detect and diagnose the self-recursive satisfaction
in the testcase:

96840.C: In substitution of ‘template<class T>  requires  C<T, Rep> void
operator*(T, Int<int>) [with T = int]’:
96840.C:1:64:   required by substitution of ‘template<class T>  requires  C<T,
Rep> void operator*(Int<int>, T) [with T = int]’
96840.C:6:28:   required from here
96840.C:1:37:   required for the satisfaction of ‘C<T, Rep>’ [with T =
Int<int>; Rep = int]
96840.C:1:41:   in requirements with ‘T t’, ‘U u’ [with U = int; T = Int<int>]
96840.C:1:41: error: satisfaction of atomic constraint ‘requires(T t, U u) {t *
 u;} [with U = Rep; T = T]’ depends on itself
    1 | template <class T, class U> concept C = requires(T t, U u) { t * u; };
      |                                         ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~

So I suppose we can mark this PR as resolved.

[Bug c++/96840] [11 Regression] Recursive substitution in constrained commutative operator

[johelegp at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

--- Comment #4 from Johel Ernesto Guerrero Peña <johelegp at gmail dot com> ---
Seems like this regressed. Do I open a new bug report?

[Bug c++/96840] [11 Regression] Recursive substitution in constrained commutative operator

[johelegp at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=96840

--- Comment #5 from Johel Ernesto Guerrero Peña <johelegp at gmail dot com> ---
I think I was wrong and this has not regressed.