[Bug fortran/106856] New: [12/13 Regression] ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977

[Bug fortran/106856] New: [12/13 Regression] ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

            Bug ID: 106856
           Summary: [12/13 Regression] ICE in gfc_conv_expr_present, at
                    fortran/trans-expr.cc:1977
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Started between 20211010 and 20211017, with option -fcheck=all :


$ cat z1.f90
program p
   class(*), allocatable :: x, y
   y = 'abc'
   call s1(x, y)
contains
   subroutine s1(x, y)
      class(*) :: x, y
   end
   subroutine s2(x, y)
      class(*), allocatable :: x, y
      optional :: x
   end
end


$ gfortran-13-20220904 -c z1.f90 -fcheck=all
z1.f90:4:16:

    4 |    call s1(x, y)
      |                1
internal compiler error: in gfc_conv_expr_present, at
fortran/trans-expr.cc:1977
0x894a80 gfc_conv_expr_present(gfc_symbol*, bool)
        ../../gcc/fortran/trans-expr.cc:1977
0x8a2b86 gfc_conv_procedure_call(gfc_se*, gfc_symbol*, gfc_actual_arglist*,
gfc_expr*, vec<tree_node*, va_gc, vl_embed>*)
        ../../gcc/fortran/trans-expr.cc:7192
0x8e1130 gfc_trans_call(gfc_code*, bool, tree_node*, tree_node*, bool)
        ../../gcc/fortran/trans-stmt.cc:422
0x867f18 trans_code
        ../../gcc/fortran/trans.cc:2018
0x8910ae gfc_generate_function_code(gfc_namespace*)
        ../../gcc/fortran/trans-decl.cc:7654
0x81353e translate_all_program_units
        ../../gcc/fortran/parse.cc:6670
0x81353e gfc_parse_file()
        ../../gcc/fortran/parse.cc:6976
0x860b1f gfc_be_parse_file
        ../../gcc/fortran/f95-lang.cc:229

[Bug fortran/106856] [12/13 Regression] ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |12.3
           Priority|P3                          |P4

[Bug fortran/106856] [12/13 Regression] ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[12/13 Regression] ICE in   |[12/13 Regression] ICE in
                   |gfc_conv_expr_present, at   |gfc_conv_expr_present, at
                   |fortran/trans-expr.cc:1977  |fortran/trans-expr.cc:1977
                   |                            |since
                   |                            |r12-4346-geb92cd57a1ebe7cd
   Last reconfirmed|                            |2022-09-07
             Status|UNCONFIRMED                 |NEW
                 CC|                            |burnus at gcc dot gnu.org,
                   |                            |marxin at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r12-4346-geb92cd57a1ebe7cd.

[Bug fortran/106856] [12/13 Regression] ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[burnus at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Created attachment 53548
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53548&action=edit
draft patch

Draft patch - the actual issue seems to work fine (first testcase, extended) –
however, I have still once ICE and bogus attribute errors (both: second
testcase)

Currently running out of time to work on this → deferred.

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |anlauf at gcc dot gnu.org

--- Comment #3 from anlauf at gcc dot gnu.org ---
Created attachment 54030
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54030&action=edit
Incremental patch

The above patch on top of Tobias' initial one fixes some minor issues with
it and improves the behavior for testcase class_75.f90 modified as follows:

! { dg-do compile }
! { dg-additional-options "-fcoarray=single" }
!
! PR fortran/106856
!
!
!
subroutine foo(x,y)
  class(*), optional :: x, y
  optional    :: x    ! { dg-error "Duplicate OPTIONAL attribute" }
  target      :: x
  allocatable :: x
  target      :: x    ! { dg-error "Duplicate TARGET attribute" }
  allocatable :: x    ! { dg-error "Duplicate ALLOCATABLE attribute" }
  pointer     :: y
  contiguous  :: y
  pointer     :: y    ! { dg-error "Duplicate POINTER attribute" }
  contiguous  :: y    ! { dg-error "Duplicate CONTIGUOUS attribute" }
  dimension   :: x(:,:)
  dimension   :: y(:,:,:)
  codimension :: x[:]
    ! < duplicat error for dimension/codimension: TODO add decl.

  ! Currently gives an ICE :-(
end


Note that these patches "regress" as follows:

- gfortran.dg/interface_41.f90 - this had a bogus error on valid code!
  (Removing the dg-error fixes it.)

- gfortran.dg/pr69398.f90 - this ICEs due to the duplicate DIMENSION attribute

I tried to modify restore_old_symbol() to handle CLASS arrays, but then
we ICE along a different path.  Not sure if this is the right approach.

Nevertheless, the patch as is already appears to be a great step forward!

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #4 from anlauf at gcc dot gnu.org ---
Created attachment 54567
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=54567&action=edit
Updated patch

The original patches started to bit-rot, so here is an updated version.

The current situation is as follows:

- it fixes a couple of ice-on-valid, ice-on-invalid, rejects-valid cases,
  see below

- testcase interface_41.f90 turns out to be actually valid Fortran and is
  now correctly accepted (-> testsuite needs adjustment)

- testcase pr69398.f90 ICE's - as expected, see comment in new testcase
  class_75.f90.  Shall we mark this one as XFAIL for the time being and
  reopen pr69398?

- it appears that the following PRs are addressed or fixed with this patch:
  pr53951, pr101101, pr104229, pr107380

So applying the patch would be a big achievement over the current situation.

@Thomas: do you think this path should/could be followed?

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

anlauf at gcc dot gnu.org changed:

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

--- Comment #5 from anlauf at gcc dot gnu.org ---
(In reply to anlauf from comment #4)
> - testcase pr69398.f90 ICE's - as expected, see comment in new testcase
>   class_75.f90.  Shall we mark this one as XFAIL for the time being and
>   reopen pr69398?

This intermediate regression appears to result from an invalid double-free.
I plan to package an updated patch for submission.

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #6 from anlauf at gcc dot gnu.org ---
Submitted: https://gcc.gnu.org/pipermail/fortran/2023-March/059003.html

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #7 from anlauf at gcc dot gnu.org ---
(In reply to anlauf from comment #6)
> Submitted: https://gcc.gnu.org/pipermail/fortran/2023-March/059003.html

As discussed in this thread, this version of the patch works for

subroutine s1 (x)
  class(*)    :: x
  allocatable :: x
  dimension   :: x(:)
  if (allocated (x)) print *, size (x)
end

subroutine s2 (x)
  class(*)    :: x
  allocatable :: x(:)
  if (allocated (x)) print *, size (x)
end

but still ICEs for the variations / line permutations

subroutine s3 (x)
  class(*)    :: x(:)
  allocatable :: x
  if (allocated (x)) print *, size (x)
end

subroutine s4 (x)
  class(*)    :: x
  dimension   :: x(:)
  allocatable :: x
  if (allocated (x)) print *, size (x)
end

I think Mikael's findings in the thread fix this.

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #8 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Harald Anlauf <anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:6aa1f40a3263741d964ef4716e85a0df5cec83b6

commit r13-6497-g6aa1f40a3263741d964ef4716e85a0df5cec83b6
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Mar 2 22:37:14 2023 +0100

    Fortran: fix CLASS attribute handling [PR106856]

    gcc/fortran/ChangeLog:

            PR fortran/106856
            * class.cc (gfc_build_class_symbol): Handle update of attributes of
            existing class container.
            (gfc_find_derived_vtab): Fix several memory leaks.
            (find_intrinsic_vtab): Ditto.
            * decl.cc (attr_decl1): Manage update of symbol attributes from
            CLASS attributes.
            * primary.cc (gfc_variable_attr): OPTIONAL shall not be taken or
            updated from the class container.
            * symbol.cc (free_old_symbol): Adjust management of symbol versions
            to not prematurely free array specs while working on the declation
            of CLASS variables.

    gcc/testsuite/ChangeLog:

            PR fortran/106856
            * gfortran.dg/interface_41.f90: Remove dg-pattern from valid
testcase.
            * gfortran.dg/class_74.f90: New test.
            * gfortran.dg/class_75.f90: New test.

    Co-authored-by: Tobias Burnus  <tobias@codesourcery.com>

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #9 from anlauf at gcc dot gnu.org ---
*** Bug 53951 has been marked as a duplicate of this bug. ***

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #10 from anlauf at gcc dot gnu.org ---
*** Bug 107380 has been marked as a duplicate of this bug. ***

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #11 from anlauf at gcc dot gnu.org ---
*** Bug 104229 has been marked as a duplicate of this bug. ***

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #12 from anlauf at gcc dot gnu.org ---
*** Bug 101101 has been marked as a duplicate of this bug. ***

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #13 from anlauf at gcc dot gnu.org ---
*** Bug 51610 has been marked as a duplicate of this bug. ***

[Bug fortran/106856] [12/13 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #14 from anlauf at gcc dot gnu.org ---
*** Bug 52227 has been marked as a duplicate of this bug. ***

[Bug fortran/106856] [12 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[12/13 Regression][OOP]     |[12 Regression][OOP] CLASS
                   |CLASS attribute handling /  |attribute handling / ICE in
                   |ICE in                      |gfc_conv_expr_present, at
                   |gfc_conv_expr_present, at   |fortran/trans-expr.cc:1977
                   |fortran/trans-expr.cc:1977  |since
                   |since                       |r12-4346-geb92cd57a1ebe7cd
                   |r12-4346-geb92cd57a1ebe7cd  |

--- Comment #15 from anlauf at gcc dot gnu.org ---
Fixed for gcc-13 so far.

I'll have a look at whether this is backportable to 12-branch.
There is at least a dependency on a part of the fix for pr102331 (r13-4934).

[Bug fortran/106856] [12 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #16 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-12 branch has been updated by Harald Anlauf
<anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:4f41c4ff250709219a7c3eba27a62f8a4689412b

commit r12-9322-g4f41c4ff250709219a7c3eba27a62f8a4689412b
Author: Harald Anlauf <anlauf@gmx.de>
Date:   Thu Mar 2 22:37:14 2023 +0100

    Fortran: fix CLASS attribute handling [PR106856]

    gcc/fortran/ChangeLog:

            PR fortran/106856
            * class.cc (gfc_build_class_symbol): Handle update of attributes of
            existing class container.
            (gfc_find_derived_vtab): Fix several memory leaks.
            (find_intrinsic_vtab): Ditto.
            * decl.cc (attr_decl1): Manage update of symbol attributes from
            CLASS attributes.
            * primary.cc (gfc_variable_attr): OPTIONAL shall not be taken or
            updated from the class container.
            * symbol.cc (free_old_symbol): Adjust management of symbol versions
            to not prematurely free array specs while working on the declation
            of CLASS variables.

    gcc/testsuite/ChangeLog:

            PR fortran/106856
            * gfortran.dg/interface_41.f90: Remove dg-pattern from valid
testcase.
            * gfortran.dg/class_74.f90: New test.
            * gfortran.dg/class_75.f90: New test.

    Co-authored-by: Tobias Burnus  <tobias@codesourcery.com>
    (cherry picked from commit 6aa1f40a3263741d964ef4716e85a0df5cec83b6)

[Bug fortran/106856] [12 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

anlauf at gcc dot gnu.org changed:

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

--- Comment #17 from anlauf at gcc dot gnu.org ---
There was another trivial dependency on the (simple) fix for pr108025.

Backported to 12-branch.  Closing.

Thanks for the report!

[Bug fortran/106856] [12 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #18 from anlauf at gcc dot gnu.org ---
*** Bug 96084 has been marked as a duplicate of this bug. ***

[Bug fortran/106856] [12 Regression][OOP] CLASS attribute handling / ICE in gfc_conv_expr_present, at fortran/trans-expr.cc:1977 since r12-4346-geb92cd57a1ebe7cd

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=106856

--- Comment #19 from anlauf at gcc dot gnu.org ---
*** Bug 96655 has been marked as a duplicate of this bug. ***