[Bug debug/106955] New: [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[Bug debug/106955] New: [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

            Bug ID: 106955
           Summary: [13 Regression] '-fcompare-debug' failure w/
                    -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra
                    --param ggc-min-expand=55
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Keywords: compare-debug-failure
          Severity: normal
          Priority: P3
         Component: debug
          Assignee: unassigned at gcc dot gnu.org
          Reporter: asolokha at gmx dot com
  Target Milestone: ---

Created attachment 53582
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53582&action=edit
gkd diff

g++ 13.0.0 20220911 snapshot (g:0ea5e3f4542832b8da016b152695e64a2a386309) fails
-fcompare-debug check when compiling the following testcase, reduced from
test/std/containers/sequences/forwardlist/forwardlist.ops/sort.pass.cpp from
the libc++ 14.0.6 test suite, w/ -std=c++20 -O1 -ftree-parallelize-loops=2
-fno-ipa-sra --param ggc-min-expand=55:

#include <forward_list>
#include <iterator>
#include <algorithm>
#include <vector>

void
test_stable ()
{
  std::vector<int> v;

  for (int i = 0; i < 1000; ++i)
    v.push_back (i);

  for (int i = 0; i < 1000; ++i)
    v[i] = i;
}

% g++-13.0.0 -std=c++20 -O1 -fcompare-debug -ftree-parallelize-loops=2
-fno-ipa-sra --param ggc-min-expand=55 -c dfptqjrz.cpp
g++-13.0.0: error: dfptqjrz.cpp: '-fcompare-debug' failure

gkd diff attached.

Preprocessed testcase is still being reduced, currently at
06:43:08 INFO (66.7%, 468334 bytes, 7781 lines)
I'll post it here once the reducer yields something meaningful under several
kilobytes in size.

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |13.0
           Keywords|                            |needs-bisection

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

--- Comment #1 from Arseny Solokha <asolokha at gmx dot com> ---
Created attachment 53589
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53589&action=edit
Reduced preprocessed testcase

Unfortunately a 173K file is the best I could come up with after three days of
reduction, though it fails the check w/ just -std=c++20 -O1 -fcompare-debug
-ftree-parallelize-loops=2.

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
                 CC|                            |marxin at gcc dot gnu.org
   Last reconfirmed|                            |2022-09-20
     Ever confirmed|0                           |1

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
(In reply to Arseny Solokha from comment #1)
> Created attachment 53589 [details]
> Reduced preprocessed testcase
> 
> Unfortunately a 173K file is the best I could come up with after three days
> of reduction, though it fails the check w/ just -std=c++20 -O1
> -fcompare-debug -ftree-parallelize-loops=2.

Can't compile this due to:

g++ pr106955.c -std=c++20 -O1 -fcompare-debug -ftree-parallelize-loops=2 -c
: In instantiation of ‘struct __gnu_cxx::__normal_iterator<int, int>’:
:822:5:   required from here
:793:44: error: ‘__gnu_cxx::__normal_iterator<int, int>::__traits_type’ {aka
‘int’} is not a class, struct, or union type
:794:38: error: ‘__gnu_cxx::__normal_iterator<int, int>::__traits_type’ {aka
‘int’} is not a class, struct, or union type
:795:36: error: ‘__gnu_cxx::__normal_iterator<int, int>::__traits_type’ {aka
‘int’} is not a class, struct, or union type
: In instantiation of ‘struct std::reverse_iterator<int>’:
:1875:44:   required from here
:759:36: error: ‘std::reverse_iterator<int>::__traits_type’ {aka ‘int’} is not
a class, struct, or union type
:764:11: error: ‘int’ is not a class, struct, or union type
: In substitution of ‘template<class _Iter, class _Tp> using
std::__detail::__iter_traits = typename std::__detail::__iter_traits_impl::type
[with _Iter = std::incrementable_traits<int>; _Tp =
std::incrementable_traits<int>]’:
:133:7:   required by substitution of ‘template<class _Tp> using
std::__detail::__iter_diff_t = typename
std::__detail::__iter_traits_impl<std::incrementable_traits<_Tp>,
std::incrementable_traits<_Tp> >::type::difference_type [with _Tp = int]’
:136:7:   required by substitution of ‘template<class _Tp> using
std::iter_difference_t = std::__detail::__iter_diff_t<typename
std::remove_cvref<_Tp>::type> [with _Tp = int]’
:767:11:   required from ‘struct std::reverse_iterator<int>’
:1875:44:   required from here
:131:7: error: invalid use of incomplete type ‘struct
std::__detail::__iter_traits_impl<std::incrementable_traits<int>,
std::incrementable_traits<int> >’
:123:40: note: declaration of ‘struct
std::__detail::__iter_traits_impl<std::incrementable_traits<int>,
std::incrementable_traits<int> >’

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jwakely.gcc at gmail dot com
           Keywords|needs-bisection             |

--- Comment #3 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r13-674-g8ccdc7ce33f5e7ca, but I don't think it's the root cause.

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

--- Comment #4 from Arseny Solokha <asolokha at gmx dot com> ---
(In reply to Martin Liška from comment #2)
> Can't compile this due to:

% g++-13.0.0 pr106955.c -std=c++20 -O1 -fsyntax-only -ftree-parallelize-loops=2
-c
% echo $?
% g++-13.0.0 pr106955.c -std=c++20 -O1 -fcompare-debug
-ftree-parallelize-loops=2 -c
g++-13.0.0: error: pr106955.c: '-fcompare-debug' failure

I've just downloaded the attachment and renamed it to pr106955.c.

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

--- Comment #5 from Martin Liška <marxin at gcc dot gnu.org> ---
Take a look please:
https://godbolt.org/z/14YYhb8oe

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

--- Comment #6 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Martin Liška from comment #2)
> (In reply to Arseny Solokha from comment #1)
> > Created attachment 53589 [details]
> > Reduced preprocessed testcase
> > 
> > Unfortunately a 173K file is the best I could come up with after three days
> > of reduction, though it fails the check w/ just -std=c++20 -O1
> > -fcompare-debug -ftree-parallelize-loops=2.
> 
> Can't compile this due to:
> 
> g++ pr106955.c -std=c++20 -O1 -fcompare-debug -ftree-parallelize-loops=2 -c
> : In instantiation of ‘struct __gnu_cxx::__normal_iterator<int, int>’:
> :822:5:   required from here
> :793:44: error: ‘__gnu_cxx::__normal_iterator<int, int>::__traits_type’ {aka
> ‘int’} is not a class, struct, or union type

This is a nonsense reduction.

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[redi at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

--- Comment #7 from Jonathan Wakely <redi at gcc dot gnu.org> ---
(In reply to Martin Liška from comment #3)
> Started with r13-674-g8ccdc7ce33f5e7ca, but I don't think it's the root
> cause.

Certainly not, when the preprocessed code from comment 0 doesn't produce a
-fcompare-debug failure for GCC 12.

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[asolokha at gmx dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

--- Comment #8 from Arseny Solokha <asolokha at gmx dot com> ---
(In reply to Martin Liška from comment #5)
> Take a look please:
> https://godbolt.org/z/14YYhb8oe

OK, then I cannot reduce it at all. My local copy of g++ accepts both this
version and the one I'm reducing right now afresh. If it is indeed nonsensical,
I'm not sure to constrain my copy of g++ to make it correctly rejecting that
code.

[Bug debug/106955] [13 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|WAITING                     |NEW
           Keywords|                            |needs-bisection,
                   |                            |needs-reduction

--- Comment #9 from Richard Biener <rguenth at gcc dot gnu.org> ---
I've tried and the issue from the description reproduces with r13-1926 but not
r13-3214 so it was eventually fixed (but preprocessed source from r13-1926
isn't accepted by r13-3214).

Bisection might help, so might reduction.

That --param ggc-min-expand=55 is required to trigger the compare-debug is
"interesting".

[Bug debug/106955] [13/14 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|13.0                        |13.2

--- Comment #10 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 13.1 is being released, retargeting bugs to GCC 13.2.

[Bug debug/106955] [13/14 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|13.2                        |13.3

--- Comment #11 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 13.2 is being released, retargeting bugs to GCC 13.3.

[Bug debug/106955] [13/14 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[law at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

Jeffrey A. Law <law at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |law at gcc dot gnu.org
           Priority|P3                          |P2

[Bug debug/106955] [13/14/15 Regression] '-fcompare-debug' failure w/ -std=c++20 -O1 -ftree-parallelize-loops=2 -fno-ipa-sra --param ggc-min-expand=55

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106955

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|13.3                        |13.4

--- Comment #12 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
GCC 13.3 is being released, retargeting bugs to GCC 13.4.