[Bug fortran/99125] New: [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[Bug fortran/99125] New: [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

            Bug ID: 99125
           Summary: [9/10/11 Regression] ICE: gimplification failed
                    (gimplify.c:15068)
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: gscfq@t-online.de
  Target Milestone: ---

Changed between 20190310 and 20190331 :


$ cat z1.f90
program p
   type t
      character(:), allocatable :: a(:)
   end type
   type(t) :: x
   character(8) :: c(2) = '12 45 78'
   x%a = c
   print *, x%a(7:8)
end


$ gfortran-9-20190310 -c z1.f90
$
$ gfortran-11-20210214 -c z1.f90
gimplification failed:
0 <integer_cst 0x7fed577da288 type <integer_type 0x7fed577d1738
integer(kind=8)> constant 0>
z1.f90:8:20:

    8 |    print *, x%a(7:8)
      |                    ^
internal compiler error: gimplification failed
0xaaad94 gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        ../../gcc/gimplify.c:15068
0xabe043 gimplify_modify_expr
        ../../gcc/gimplify.c:5833
0xaa9e9a gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        ../../gcc/gimplify.c:14070
0xaad8a8 gimplify_stmt(tree_node**, gimple**)
        ../../gcc/gimplify.c:6876
0xaa9793 gimplify_statement_list
        ../../gcc/gimplify.c:1879
0xaa9793 gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        ../../gcc/gimplify.c:14514
0xaad8a8 gimplify_stmt(tree_node**, gimple**)
        ../../gcc/gimplify.c:6876
0xaae243 gimplify_bind_expr
        ../../gcc/gimplify.c:1421
0xaaa612 gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        ../../gcc/gimplify.c:14271
0xaad8a8 gimplify_stmt(tree_node**, gimple**)
        ../../gcc/gimplify.c:6876
0xaa9793 gimplify_statement_list
        ../../gcc/gimplify.c:1879
0xaa9793 gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        ../../gcc/gimplify.c:14514
0xaad8a8 gimplify_stmt(tree_node**, gimple**)
        ../../gcc/gimplify.c:6876
0xaae243 gimplify_bind_expr
        ../../gcc/gimplify.c:1421
0xaaa612 gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        ../../gcc/gimplify.c:14271
0xaad8a8 gimplify_stmt(tree_node**, gimple**)
        ../../gcc/gimplify.c:6876
0xaa9793 gimplify_statement_list
        ../../gcc/gimplify.c:1879
0xaa9793 gimplify_expr(tree_node**, gimple**, gimple**, bool (*)(tree_node*),
int)
        ../../gcc/gimplify.c:14514
0xaad8a8 gimplify_stmt(tree_node**, gimple**)
        ../../gcc/gimplify.c:6876
0xaae243 gimplify_bind_expr
        ../../gcc/gimplify.c:1421

[Bug fortran/99125] [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

--- Comment #1 from G. Steinmetz <gscfq@t-online.de> ---

Works with valid code :


$ cat z0.f90
program p
   type t
      character(:), allocatable :: c(:)
   end type
   character(8) :: a(2) = '12 45 78'
   type(t) :: x
   x%c = a
   print *, x%c
   print *, x%c(1)(7:8)
end


$ gfortran-11-20210214 z0.f90 && ./a.out
 12 45 7812 45 78
 78

[Bug fortran/99125] [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2021-02-16
                 CC|                            |marxin at gcc dot gnu.org,
                   |                            |pault at gcc dot gnu.org

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
Confirmed, started with r9-6726-gd5f48c7c62d3d8cf.

[Bug fortran/99125] [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |9.4
           Priority|P3                          |P4

[Bug fortran/99125] [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           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 50229
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=50229&action=edit
Provisional patch

This regtests OK but needs a small amount of reorganisation before submission.
It willl happen tomorrow.

BTW Although the ICE is a regression, the testcase never gave the intended
result.

What was the reason for the (7:8) in the testcase? -fcheck=all picks it up
correctly.

Cheers

Paul

[Bug fortran/99125] [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

--- Comment #4 from G. Steinmetz <gscfq@t-online.de> ---

> What was the reason for the (7:8) in the testcase?
If this question was directed at me, some background :

The problem occurred during verification tests, 
in other code constellation and with other numbers. 
(7:8) is simply beyond the array extension (1:2).

The example in comment #0 was the result of some 
simplifications in order to get a very simple test case.

I wondered about the ICE, and that it could be narrowed 
down to (well) a time range in the recent past.

In the case of a regression, there must be a trigger for 
the change, and knowing this trigger can be quite helpful.

It sometimes happens that there are gray zones where 
something "works" incorrectly with some numbers, 
or does not work with other numbers and remains undetected, 
or is detected with some combinations of options with again 
other numbers, or leads to an ICE with again other numbers.

Cheers,
Gerhard

[Bug fortran/99125] [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[gscfq@t-online.de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

--- Comment #5 from G. Steinmetz <gscfq@t-online.de> ---

I also checked this invalid (probably a known issue) :


$ cat za.f90
program p
   character(:), allocatable :: a(:)
   character(8) :: c(2) = '12 45 78'
   a = c
   print *, a(7:8)
end

$ gfortran-11-20210221 za.f90 && ./a.out   # surprising
 12 45 7812 45 78


Detected at runtime :

$ gfortran-11-20210221 za.f90 -Wall -fcheck=all && ./a.out
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
Warning: '.a' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:4:8:

    4 |    a = c
      |        ^
Warning: 'a.offset' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
note: 'a' declared here
za.f90:4:8:

    4 |    a = c
      |        ^
Warning: 'a.dim[0].lbound' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
note: 'a' declared here
za.f90:4:8:

    4 |    a = c
      |        ^
Warning: 'a.dim[0].ubound' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
note: 'a' declared here
za.f90:4:8:

    4 |    a = c
      |        ^
Warning: 'a.dim[0].lbound' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
note: 'a' declared here
za.f90:4:8:

    4 |    a = c
      |        ^
Warning: 'a.dim[0].ubound' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
note: 'a' declared here
za.f90:4:8:

    4 |    a = c
      |        ^
Warning: 'a.dim[0].ubound' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
note: 'a' declared here
za.f90:4:8:

    4 |    a = c
      |        ^
Warning: 'a.dim[0].lbound' may be used uninitialized [-Wmaybe-uninitialized]
za.f90:2:36:

    2 |    character(:), allocatable :: a(:)
      |                                    ^
note: 'a' declared here
At line 5 of file za.f90
Fortran runtime error: Index '7' of dimension 1 of array 'a' outside of
expected range (1:2)

Error termination. Backtrace:
#0  0x4025d1 in ???
#1  0x402738 in ???
#2  0x7fa0f5697f89 in ???
#3  0x4022b9 in ???
        at ../sysdeps/x86_64/start.S:120
#4  0xffffffffffffffff in ???


---

The very basic simplification gives :

$ cat zc.f90
program p
   character(8) :: c(2) = '12 45 78'
   print *, c(7:8)
end

$ gfortran-11-20210221 zc.f90 -Wall -fcheck=all && ./a.out
zc.f90:3:14:

    3 |    print *, c(7:8)
      |              1
Warning: Lower array reference at (1) is out of bounds (7 > 2) in dimension 1
At line 3 of file zc.f90
Fortran runtime error: Index '7' of dimension 1 of array 'c' outside of
expected range (1:2)

Error termination. Backtrace:
#0  0x4023f7 in ???
#1  0x40242b in ???
#2  0x7fc15fc2af89 in ???
#3  0x4022b9 in ???
        at ../sysdeps/x86_64/start.S:120
#4  0xffffffffffffffff in ???

[Bug fortran/99125] [9/10/11 Regression] ICE: gimplification failed (gimplify.c:15068)

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Paul Thomas <pault@gcc.gnu.org>:

https://gcc.gnu.org/g:7987beec679898cfa75839551d55ae5234a216bd

commit r11-7657-g7987beec679898cfa75839551d55ae5234a216bd
Author: Paul Thomas <pault@gcc.gnu.org>
Date:   Sat Mar 13 11:39:57 2021 +0000

    Fortran: Fix for class defined operators [PR99125].

    2021-03-13  Paul Thomas  <pault@gcc.gnu.org>

    gcc/fortran
            PR fortran/99125
            * trans-array.c (gfc_conv_expr_descriptor): For deferred length
            length components use the ss_info string length instead of
            gfc_get_expr_charlen. Make sure that the deferred string length
            is a variable before assigning to it. Otherwise use the expr.
            * trans-expr.c (gfc_conv_string_length): Make sure that the
            deferred string length is a variable before assigning to it.

    gcc/testsuite/
            PR fortran/99125
            * gfortran.dg/alloc_deferred_comp_1.f90: New test.

[Bug fortran/99125] [9/10/11/12 Regression] ICE: gimplification failed (gimplify.c:15068)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |ASSIGNED

[Bug fortran/99125] [9/10/11/12 Regression] ICE: gimplification failed (gimplify.c:15068)

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|9.4                         |9.5

--- Comment #7 from Richard Biener <rguenth at gcc dot gnu.org> ---
GCC 9.4 is being released, retargeting bugs to GCC 9.5.

[Bug fortran/99125] [9/10/11/12 Regression] ICE: gimplification failed (gimplify.c:15068)

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |11.2.1, 12.0
                 CC|                            |anlauf at gcc dot gnu.org

--- Comment #8 from anlauf at gcc dot gnu.org ---
AFAICT the invalid testcase in comment#0 does not ICE after the commit
in comment#6 on gcc-12 and 11-branch, so Paul simply needs to backport
to 10- and 9-branches.

The bounds violation is correctly detected at runtime:

At line 8 of file pr99125.f90
Fortran runtime error: Index '7' of dimension 1 of array 'x' outside of
expected range (1:2)

The testcase in comment#1 still works.

Also, the invalid testcase in comment#5 are handled at runtime.

@Paul: do you intent to backport your patch so that we can close this one?

[Bug fortran/99125] [9/10/11/12 Regression] ICE: gimplification failed (gimplify.c:15068)

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

--- Comment #9 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-10 branch has been updated by Harald Anlauf
<anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:4cecfae4b57e052008e2febe25607d5e2cad5dbb

commit r10-10075-g4cecfae4b57e052008e2febe25607d5e2cad5dbb
Author: Paul Thomas <pault@gcc.gnu.org>
Date:   Sat Mar 13 11:39:57 2021 +0000

    Fortran: Fix for class defined operators [PR99125].

    2021-03-13  Paul Thomas  <pault@gcc.gnu.org>

    gcc/fortran
            PR fortran/99125
            * trans-array.c (gfc_conv_expr_descriptor): For deferred length
            length components use the ss_info string length instead of
            gfc_get_expr_charlen. Make sure that the deferred string length
            is a variable before assigning to it. Otherwise use the expr.
            * trans-expr.c (gfc_conv_string_length): Make sure that the
            deferred string length is a variable before assigning to it.

    gcc/testsuite/
            PR fortran/99125
            * gfortran.dg/alloc_deferred_comp_1.f90: New test.

    (cherry picked from commit 7987beec679898cfa75839551d55ae5234a216bd)

[Bug fortran/99125] [9 Regression] ICE: gimplification failed (gimplify.c:15068)

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

anlauf at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[9/10/11/12 Regression]     |[9 Regression] ICE:
                   |ICE: gimplification failed  |gimplification failed
                   |(gimplify.c:15068)          |(gimplify.c:15068)

--- Comment #10 from anlauf at gcc dot gnu.org ---
Cherry-picked to 10-branch for Paul after regtesting.

[Bug fortran/99125] [9 Regression] ICE: gimplification failed (gimplify.c:15068)

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

--- Comment #11 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Harald Anlauf
<anlauf@gcc.gnu.org>:

https://gcc.gnu.org/g:3e59c3c6f4c4f42a158d8ab936c995754bf22dee

commit r9-9711-g3e59c3c6f4c4f42a158d8ab936c995754bf22dee
Author: Paul Thomas <pault@gcc.gnu.org>
Date:   Sat Mar 13 11:39:57 2021 +0000

    Fortran: Fix for class defined operators [PR99125].

    2021-03-13  Paul Thomas  <pault@gcc.gnu.org>

    gcc/fortran
            PR fortran/99125
            * trans-array.c (gfc_conv_expr_descriptor): For deferred length
            length components use the ss_info string length instead of
            gfc_get_expr_charlen. Make sure that the deferred string length
            is a variable before assigning to it. Otherwise use the expr.
            * trans-expr.c (gfc_conv_string_length): Make sure that the
            deferred string length is a variable before assigning to it.

    gcc/testsuite/
            PR fortran/99125
            * gfortran.dg/alloc_deferred_comp_1.f90: New test.

    (cherry picked from commit 7987beec679898cfa75839551d55ae5234a216bd)

[Bug fortran/99125] [9 Regression] ICE: gimplification failed (gimplify.c:15068)

[anlauf at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=99125

anlauf at gcc dot gnu.org changed:

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

--- Comment #12 from anlauf at gcc dot gnu.org ---
Should be fixed on all open branches.  Closing.

Thanks for the report!