[Bug fortran/59941] New: ICE with polymorphic types in [4.7.0|4.7.1|4.73|4.7.4]

[Bug fortran/59941] New: ICE with polymorphic types in [4.7.0|4.7.1|4.73|4.7.4]

[juergen.reuter at desy dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

            Bug ID: 59941
           Summary: ICE with polymorphic types in [4.7.0|4.7.1|4.73|4.7.4]
           Product: gcc
           Version: 4.7.4
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: juergen.reuter at desy dot de

This is the code that triggers the ICE:
$ gfortran -c ICE.f90
f951: internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <http://gcc.gnu.org/bugs.html> for instructions.

Here is the code (also attached, 147 lines):


module tao_random_numbers
  integer, parameter, private :: tao_i32 = selected_int_kind (9)
  integer, parameter, private :: DEFAULT_BUFFER_SIZE = 1009
  integer(kind=tao_i32), dimension(DEFAULT_BUFFER_SIZE), save, private ::
s_buffer
  integer, save, private :: s_buffer_end = size (s_buffer)
end module tao_random_numbers

!!!!

module interactions
  public :: interaction_t

  type :: external_link_t
     type(interaction_t), pointer :: int => null ()
  end type external_link_t

  type :: interaction_t
     type(external_link_t), dimension(:), allocatable :: source
  end type interaction_t

end module interactions

!!!!!

module polarizations
  public :: smatrix_t
  public :: pmatrix_t

  type :: smatrix_t
     integer :: dim = 0
     integer, dimension(:,:), allocatable :: index
  end type smatrix_t

  type, extends (smatrix_t) :: pmatrix_t
   contains
     procedure :: p_from_s => pmatrix_assign_from_smatrix
  end type pmatrix_t

contains

  subroutine pmatrix_assign_from_smatrix (pmatrix, smatrix)
    class(pmatrix_t), intent(out) :: pmatrix
    type(smatrix_t), intent(in) :: smatrix
    pmatrix%smatrix_t = smatrix
  end subroutine pmatrix_assign_from_smatrix

end module polarizations

!!!!!

module beam_structures
  use polarizations
  public :: beam_structure_t

  type :: beam_structure_entry_t
     logical :: is_valid = .false.
  end type beam_structure_entry_t

  type :: beam_structure_record_t
     type(beam_structure_entry_t), dimension(:), allocatable :: entry
  end type beam_structure_record_t

  type :: beam_structure_t
     private
     integer :: n_beam = 0
     type(beam_structure_record_t), dimension(:), allocatable :: record
     type(smatrix_t), dimension(:), allocatable :: smatrix
   contains
     procedure :: get_smatrix => beam_structure_get_smatrix
  end type beam_structure_t

contains

  function beam_structure_get_smatrix (beam_structure) result (smatrix)
    class(beam_structure_t), intent(in) :: beam_structure
    type(smatrix_t), dimension (size (beam_structure%smatrix)) :: smatrix
    smatrix = beam_structure%smatrix
  end function beam_structure_get_smatrix

end module beam_structures

!!!!

module beams
  use polarizations
  use beam_structures
  public :: beam_data_t
  type :: beam_data_t
     type(pmatrix_t), dimension(:), allocatable :: pmatrix
  end type beam_data_t
end module beams

!!!!!

module sf_base
  use interactions
  use beams
  public :: sf_data_t
  public :: sf_int_t

  type, abstract :: sf_data_t
  end type sf_data_t

  type, abstract, extends (interaction_t) :: sf_int_t
   contains
     procedure (sf_int_init), deferred :: init
  end type sf_int_t

  abstract interface
     subroutine sf_int_init (sf_int, data)
       import
       class(sf_int_t), intent(out) :: sf_int
       class(sf_data_t), intent(in), target :: data
     end subroutine sf_int_init
  end interface
end module sf_base

!!!!!

module sf_circe1
  use tao_random_numbers !NODEP!
  use sf_base

  type, extends (sf_data_t) :: circe1_data_t 
     private 
   contains  
       procedure :: init => circe1_data_init 
  end type circe1_data_t 

  type, extends (sf_int_t) :: circe1_t
     type(circe1_data_t), pointer :: data => null ()
   contains
     procedure :: init => circe1_init
  end type circe1_t 

contains

  subroutine circe1_data_init (data)
    class(circe1_data_t), intent(out) :: data 
  end subroutine circe1_data_init 

  subroutine circe1_init (sf_int, data) 
    class(circe1_t), intent(out) :: sf_int 
    class(sf_data_t), intent(in), target :: data
  end subroutine circe1_init

end module sf_circe1

[Bug fortran/59941] ICE with polymorphic types in [4.7.0|4.7.1|4.73|4.7.4]

[juergen.reuter at desy dot de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #1 from Jürgen Reuter <juergen.reuter at desy dot de> ---
Created attachment 31951
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=31951&action=edit
Code triggering the ICE
>From gcc-bugs-return-441526-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sat Jan 25 11:09:37 2014
Return-Path: <gcc-bugs-return-441526-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 29491 invoked by alias); 25 Jan 2014 11:09:37 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 29478 invoked by uid 48); 25 Jan 2014 11:09:34 -0000
From: "juergen.reuter at desy dot de" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug fortran/59941] ICE with polymorphic types in [4.7.0|4.7.1|4.73|4.7.4]
Date: Sat, 25 Jan 2014 11:09:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: fortran
X-Bugzilla-Version: 4.7.4
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: juergen.reuter at desy dot de
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-59941-4-0qTeV3KZ6p@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-59941-4@http.gcc.gnu.org/bugzilla/>
References: <bug-59941-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-01/txt/msg02668.txt.bz2
Content-length: 193

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #2 from Jürgen Reuter <juergen.reuter at desy dot de> ---
The bug appears only with 4.7.x, with 4.8 and 4.9 it is working.
>From gcc-bugs-return-441527-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sat Jan 25 12:43:36 2014
Return-Path: <gcc-bugs-return-441527-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 29523 invoked by alias); 25 Jan 2014 12:43:35 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 29503 invoked by uid 48); 25 Jan 2014 12:43:31 -0000
From: "janus at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug fortran/59941] [OOP] ICE with polymorphic types in 4.7.x
Date: Sat, 25 Jan 2014 12:43:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: fortran
X-Bugzilla-Version: 4.7.4
X-Bugzilla-Keywords: ice-on-valid-code
X-Bugzilla-Severity: normal
X-Bugzilla-Who: janus at gcc dot gnu.org
X-Bugzilla-Status: NEW
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields: keywords bug_status cf_reconfirmed_on cc short_desc everconfirmed
Message-ID: <bug-59941-4-7ZwL8Cl7GP@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-59941-4@http.gcc.gnu.org/bugzilla/>
References: <bug-59941-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-01/txt/msg02669.txt.bz2
Content-length: 761

http://gcc.gnu.org/bugzilla/show_bug.cgi?idY941

janus at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-01-25
                 CC|                            |janus at gcc dot gnu.org
            Summary|ICE with polymorphic types  |[OOP] ICE with polymorphic
                   |in [4.7.0|4.7.1|4.73|4.7.4] |types in 4.7.x
     Ever confirmed|0                           |1

--- Comment #3 from janus at gcc dot gnu.org ---
Confirmed. Also seems to work with 4.6.4 here.

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

janus at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[OOP] ICE with polymorphic  |[4.7 Regression] [OOP] ICE
                   |types in 4.7.x              |with polymorphic types

--- Comment #4 from janus at gcc dot gnu.org ---
The current bug list shows three 4.7-only regressions, two of which are
OOP-related:

 * PR 53718
 * PR 56731

At first sight the problem here seems to be different, though.

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #5 from janus at gcc dot gnu.org ---
Reduced test case:


module tao_random_numbers
  integer, dimension(10), private :: s_buffer
  integer, private :: s_buffer_end = size (s_buffer)
end module


module beam_structures

  type :: smatrix_t
  end type

  type :: beam_structure_t
     type(smatrix_t), dimension(:), allocatable :: smatrix
   contains
     procedure :: get_smatrix
  end type

contains

  function get_smatrix (beam_structure) result (smatrix)
    class(beam_structure_t), intent(in) :: beam_structure
    type(smatrix_t), dimension (size (beam_structure%smatrix)) :: smatrix
  end function

end module


module sf_circe1
  use tao_random_numbers
  use beam_structures
end module

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

janus at gcc dot gnu.org changed:

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

--- Comment #6 from janus at gcc dot gnu.org ---
Definitely some kind of module madness is at work here. Somewhat reminds me of
PR 58007.

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #7 from janus at gcc dot gnu.org ---
Another variant with the same behavior (ICEs only with 4.7):


module tao_random_numbers
  integer, dimension(10), private :: s_buffer
  integer, private :: s_buffer_end = size (s_buffer)
end module


module beam_structures

  private

  type :: beam_structure_t
     integer, dimension(:), allocatable :: smatrix
   contains
     procedure :: get_smatrix
  end type

contains

  function get_smatrix (beam_structure) result (matrix)
    class(beam_structure_t), intent(in) :: beam_structure
    integer, dimension (size (beam_structure%smatrix)) :: matrix
  end function

end module


program p
  use tao_random_numbers
  use beam_structures
end



Everything in the modules is private, so the USE statements should not have any
effect in principle.

Btw, maybe a warning for all-private modules would be appropriate? A module
without any public entities is not very useful, I guess ...

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #8 from janus at gcc dot gnu.org ---
Backtrace from gdb:


Program received signal SIGSEGV, Segmentation fault.
0x00000000004d19d6 in replace_comp (expr=expr@entry=0x146cc10,
sym=sym@entry=0x146f7a0, i=i@entry=0x7fffffffdc6c)
    at /home/jweil/gcc47/branch/gcc/fortran/expr.c:4199
4199           && !gfc_is_intrinsic (expr->symtree->n.sym, 0, expr->where)))
(gdb) bt
#0  0x00000000004d19d6 in replace_comp (expr=expr@entry=0x146cc10,
sym=sym@entry=0x146f7a0, i=i@entry=0x7fffffffdc6c)
    at /home/jweil/gcc47/branch/gcc/fortran/expr.c:4199
#1  0x00000000004d46ba in gfc_traverse_expr (expr=0x146cc10,
sym=sym@entry=0x146f7a0, 
    func=func@entry=0x4d19a0 <replace_comp>, f=f@entry=0) at
/home/jweil/gcc47/branch/gcc/fortran/expr.c:3929
#2  0x00000000004d6c9c in gfc_expr_replace_comp (expr=<optimized out>,
dest=dest@entry=0x146f7a0)
    at /home/jweil/gcc47/branch/gcc/fortran/expr.c:4218
#3  0x0000000000524615 in resolve_fl_derived0 (sym=sym@entry=0x146e590)
    at /home/jweil/gcc47/branch/gcc/fortran/resolve.c:11724
#4  0x000000000052c80d in resolve_fl_derived (sym=sym@entry=0x146e590)
    at /home/jweil/gcc47/branch/gcc/fortran/resolve.c:12046
#5  0x0000000000522238 in resolve_symbol (sym=0x146e590) at
/home/jweil/gcc47/branch/gcc/fortran/resolve.c:12308
#6  0x0000000000540a84 in do_traverse_symtree (st=0x146ca50,
st_func=st_func@entry=0x0, 
    sym_func=sym_func@entry=0x521ac0 <resolve_symbol>) at
/home/jweil/gcc47/branch/gcc/fortran/symbol.c:3387
#7  0x000000000054440d in gfc_traverse_ns (ns=ns@entry=0x146bf80,
sym_func=sym_func@entry=0x521ac0 <resolve_symbol>)
    at /home/jweil/gcc47/branch/gcc/fortran/symbol.c:3412
#8  0x00000000005209d1 in resolve_types (ns=ns@entry=0x146bf80) at
/home/jweil/gcc47/branch/gcc/fortran/resolve.c:13885
#9  0x00000000005213c1 in gfc_resolve (ns=0x146bf80) at
/home/jweil/gcc47/branch/gcc/fortran/resolve.c:13985
#10 0x000000000052f7a5 in gfc_resolve (ns=<optimized out>) at
/home/jweil/gcc47/branch/gcc/fortran/resolve.c:13978
#11 0x0000000000516153 in resolve_all_program_units
(gfc_global_ns_list=0x145ab60)
    at /home/jweil/gcc47/branch/gcc/fortran/parse.c:4392

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #9 from janus at gcc dot gnu.org ---
(In reply to janus from comment #6)
> Somewhat reminds me of PR 58007.

Unfortunately r207119 does not seem to fix it on the 4.7 branch. Apparently
it's a different problem after all ...

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #11 from janus at gcc dot gnu.org ---
The reason why the ICE is gone in 4.8 is that 'replace_comp' was completely
removed in r195562 (for PR 54107). However this commit is a bit too heavy for a
backport, I guess.

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

janus at gcc dot gnu.org changed:

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

--- Comment #12 from janus at gcc dot gnu.org ---
Here is a more lightweight fix (pretty much straightforward):

Index: gcc/fortran/expr.c
===================================================================
--- gcc/fortran/expr.c    (revision 207198)
+++ gcc/fortran/expr.c    (working copy)
@@ -4195,7 +4195,7 @@ replace_comp (gfc_expr *expr, gfc_symbol *sym, int
   gfc_component *comp;
   comp = (gfc_component *)sym;
   if ((expr->expr_type == EXPR_VARIABLE 
-       || (expr->expr_type == EXPR_FUNCTION
+       || (expr->expr_type == EXPR_FUNCTION && !expr->value.function.isym
        && !gfc_is_intrinsic (expr->symtree->n.sym, 0, expr->where)))
       && expr->symtree->n.sym->ns == comp->ts.interface->formal_ns)
     {


Removes the ICE on comment 0, comment 5 and comment 7.

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

--- Comment #14 from janus at gcc dot gnu.org ---
Author: janus
Date: Mon Feb  3 08:16:06 2014
New Revision: 207417

URL: http://gcc.gnu.org/viewcvs?rev=207417&root=gcc&view=rev
Log:
2014-02-03  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/59941
    * expr.c (replace_comp): Check for isym to avoid ICE.

2014-02-03  Janus Weil  <janus@gcc.gnu.org>

    PR fortran/59941
    * gfortran.dg/typebound_proc_26.f90: New.

Added:
    branches/gcc-4_7-branch/gcc/testsuite/gfortran.dg/typebound_proc_26.f90
Modified:
    branches/gcc-4_7-branch/gcc/fortran/ChangeLog
    branches/gcc-4_7-branch/gcc/fortran/expr.c
    branches/gcc-4_7-branch/gcc/testsuite/ChangeLog

[Bug fortran/59941] [4.7 Regression] [OOP] ICE with polymorphic types

[janus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=59941

janus at gcc dot gnu.org changed:

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

--- Comment #15 from janus at gcc dot gnu.org ---
Fixed on the 4.7 branch with r207417 (for a future release 4.7.4).

Closing. Thanks for the report!