[Bug fortran/64932] New: [5.0] ICE on module end, gfc_conv_descriptor_data_get

[Bug fortran/64932] New: [5.0] ICE on module end, gfc_conv_descriptor_data_get

[shapero at uw dot edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

            Bug ID: 64932
           Summary: [5.0] ICE on module end, gfc_conv_descriptor_data_get
           Product: gcc
           Version: 5.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: shapero at uw dot edu

Created attachment 34661
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34661&action=edit
internal compiler error backtrace

I have some Fortran 2003 code, hosted here:

https://github.com/danshapero/sigma

It builds and passes all tests on gfortran-4.7.2 and 4.8.2 on Intel 32, AMD64
and ARM with Linux, and Intel/OSX.

I'm attempting to build it with gfortran-5.0, because co-arrays :D.
Unfortunately, I'm getting an internal compiler error. GCC was configured with

../configure --prefix=/home/daniel/programs/gcc --enable-languages=c,fortran
--enable-checking=release --disable-bootstrap
--disable-build-poststage1-with-cxx --disable-libstdcxx-pch --disable-multilib

in accordance with the suggestions from
https://gcc.gnu.org/wiki/GFortranBinaries#FromSource. The GCC test suite
passes, and I've compiled some simple programs to verify that gfortran-5.0
works.

The backtrace is long; I've attached it as a text file. The library is built
with the usual

mkdir build
cd build
FC=gfortran-5.0 CC=gcc-5.0 cmake ..
make

The error also occurs when I do a debug build that includes -Wall -Wextra.

I can attempt to make a reduced test case for this, but I'm a bit out of my
depth here. Thanks for reading!

[Bug fortran/64932] [5.0] ICE on module end, gfc_conv_descriptor_data_get

[burnus at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

Tobias Burnus <burnus at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |ice-on-valid-code
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2015-02-04
                 CC|                            |burnus at gcc dot gnu.org
      Known to work|                            |4.7.2, 4.8.2
     Ever confirmed|0                           |1
      Known to fail|                            |5.0

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> ---
Reduced test case:

module coo_graphs
  implicit none
  type :: dynamic_array
    integer :: length, capacity, min_capacity
    integer, allocatable :: array(:)
  end type
  type :: coo_graph
    type(dynamic_array) :: edges(2)
    integer, private :: ne
  end type coo_graph
contains
  subroutine coo_dump_edges(g, edges)
    class(coo_graph), intent(in) :: g
    integer, intent(out) :: edges(:,:)
  end subroutine coo_dump_edges
end module coo_graphs

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
                 CC|                            |jakub at gcc dot gnu.org
            Summary|[4.9/5.0 Regression] ICE in |[4.9/5 Regression] ICE in
                   |gfc_conv_descriptor_data_ge |gfc_conv_descriptor_data_ge
                   |t for generated finalizer   |t for generated finalizer

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

Paul Thomas <pault at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pault at gcc dot gnu.org
           Assignee|unassigned at gcc dot gnu.org      |pault at gcc dot gnu.org

--- Comment #3 from Paul Thomas <pault at gcc dot gnu.org> ---
Created attachment 34691
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34691&action=edit
Patch for the PR

class.c(finalize_component) is not able to deal correctly with non-allocatable
derived type array components that have finalizable components. Rather than
generating loops in finalize_component, I detected the condition in
trans-stmt.c(gfc_trans_deallocate) and have called gfc_deallocate_alloc_comp
after obtaining the derived type for the array.

Happily, this fix does not generate the error:
Error: Two or more part references with nonzero rank must not be specified at
(1)

which occurs if the code is written explicitly.

Bootstraps and regtests on FC21/x_86_64

OK for trunk?

Paul

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[paul.richard.thomas at gmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

--- Comment #4 from paul.richard.thomas at gmail dot com <paul.richard.thomas at gmail dot com> ---
Dear All,

It would be nice to commit this tonight, if possible. An impetus to do
this is added by Dominique pointing out that it fixes PRs 59765, 60529
and 61766!

Cheers

Paul

On 7 February 2015 at 12:23, pault at gcc dot gnu.org
<gcc-bugzilla@gcc.gnu.org> wrote:
> https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932
>
> Paul Thomas <pault at gcc dot gnu.org> changed:
>
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>                  CC|                            |pault at gcc dot gnu.org
>            Assignee|unassigned at gcc dot gnu.org      |pault at gcc dot gnu.org
>
> --- Comment #3 from Paul Thomas <pault at gcc dot gnu.org> ---
> Created attachment 34691
>   --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=34691&action=edit
> Patch for the PR
>
> class.c(finalize_component) is not able to deal correctly with non-allocatable
> derived type array components that have finalizable components. Rather than
> generating loops in finalize_component, I detected the condition in
> trans-stmt.c(gfc_trans_deallocate) and have called gfc_deallocate_alloc_comp
> after obtaining the derived type for the array.
>
> Happily, this fix does not generate the error:
> Error: Two or more part references with nonzero rank must not be specified at
> (1)
>
> which occurs if the code is written explicitly.
>
> Bootstraps and regtests on FC21/x_86_64
>
> OK for trunk?
>
> Paul
>
> --
> You are receiving this mail because:
> You are on the CC list for the bug.
> You are the assignee for the bug.

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

--- Comment #5 from Paul Thomas <pault at gcc dot gnu.org> ---
Author: pault
Date: Thu Feb 12 19:30:53 2015
New Revision: 220654

URL: https://gcc.gnu.org/viewcvs?rev=220654&root=gcc&view=rev
Log:
2015-02-12  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/64932
    * trans-stmt.c (gfc_trans_deallocate): If a component array
    expression is not a descriptor type and it is a derived type
    that has allocatable components and is not finalizable, then
    deallocate the allocatable components.

2015-02-12  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/64932
    * gfortran.dg/finalize_28.f90: New test


Added:
    trunk/gcc/testsuite/gfortran.dg/finalize_28.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/trans-stmt.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

--- Comment #6 from Paul Thomas <pault at gcc dot gnu.org> ---
Author: pault
Date: Thu Feb 12 21:06:41 2015
New Revision: 220659

URL: https://gcc.gnu.org/viewcvs?rev=220659&root=gcc&view=rev
Log:
2015-02-12  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/64932
    * trans-stmt.c (gfc_trans_deallocate): If a component array
    expression is not a descriptor type and it is a derived type
    that has allocatable components and is not finalizable, then
    deallocate the allocatable components.

2015-02-12  Paul Thomas  <pault@gcc.gnu.org>

    PR fortran/64932
    * gfortran.dg/finalize_28.f90: New test


Added:
    branches/gcc-4_9-branch/gcc/testsuite/gfortran.dg/finalize_28.f90
Modified:
    branches/gcc-4_9-branch/gcc/fortran/ChangeLog
    branches/gcc-4_9-branch/gcc/fortran/trans-stmt.c
    branches/gcc-4_9-branch/gcc/testsuite/ChangeLog

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

Paul Thomas <pault at gcc dot gnu.org> changed:

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

--- Comment #7 from Paul Thomas <pault at gcc dot gnu.org> ---
Fixed on 4.9 and 5.0.

Thanks for the report.

Paul

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[shapero at uw dot edu]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

--- Comment #8 from Daniel Shapero <shapero at uw dot edu> ---
Everything works now, thanks Paul!

[Bug fortran/64932] [4.9/5 Regression] ICE in gfc_conv_descriptor_data_get for generated finalizer

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=64932

Dominique d'Humieres <dominiq at lps dot ens.fr> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |mandrew9 at vt dot edu

--- Comment #9 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
*** Bug 65571 has been marked as a duplicate of this bug. ***