[Bug fortran/51435] New: Bad association status after null() of derived type component

[Bug fortran/51435] New: Bad association status after null() of derived type component

[darmar.xxl at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

             Bug #: 51435
           Summary: Bad association status after null() of derived type
                    component
    Classification: Unclassified
           Product: gcc
           Version: 4.6.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: darmar.xxl@gmail.com


Created attachment 26002
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26002
test code

Here is the code:
!---------------
module arr_m
    type arr_t
        real(8), dimension(:), allocatable :: rsk
        ! if allocatable changed to pointer, all is OK
    end type
end module arr_m
!*********************
module list_m
    use arr_m
    implicit none

    type my_list
        type(arr_t), pointer :: head => null()
        !!!integer, pointer     :: tail => null() !if this is uncommented, all
is OK.
    end type my_list
end module list_m
!***********************
module worker_mod
    use list_m
    implicit none

    type data_all_t
        type(my_list) :: my_data
    end type data_all_t
contains
    subroutine do_job()
        type(data_all_t) :: dum

        if (associated(dum%my_data%head)) then
            print *, 'do_job my_data%head is associated' 
!
! gfortran goes here.
! WRONG. head was nullified.
!
        else
            print *, 'do_job my_data%head is NOT associated' ! OK
        end if

    end subroutine

end module
!***************
program hello
    use worker_mod
    implicit none

    call do_job()

end program
!***************


gfortran reports that dum%my_data%head is associated in the sub "do_job", while
"head" component was nullified using null() in the variable declaration line.

I tested this with gfortran 4.6 and 4.7 on Linux and Windows (produces wrong
results). Version 4.5 outputs correct result. (ifort is OK).

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |wrong-code
                 CC|                            |burnus at gcc dot gnu.org
   Target Milestone|---                         |4.6.3
            Summary|Bad association status      |[4.6/4.7 Regression] Bad
                   |after null() of derived     |association status after
                   |type component              |null() of derived type
                   |                            |component

--- Comment #1 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-06 10:08:32 UTC ---
With GCC 4.3, 4.4 and 4.5, I get:
  do_job my_data%head is NOT associated
while GCC 4.6 prints
 do_job my_data%head is associated

With GCC 4.7 I get the correct result ("NOT associated") but it shows the same
valgrind warning as GCC 4.6. (GCC 4.5 shows no valgrind warning.) Thus, either
the issue is fixed in 4.7 - and 4.6 and 4.7 have additional issues. Or it works
in 4.7 only by chance.

==19540== Conditional jump or move depends on uninitialised value(s)
==19540==    at 0x400776: __worker_mod_MOD_do_job (hjhjff.f90:29)
==19540==    by 0x400849: MAIN__ (hjhjff.f90:47)
==19540==    by 0x40087F: main (hjhjff.f90:44)

Line 29 is:
        if (associated(dum%my_data%head)) then

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-06 10:33:57 UTC ---
Lightly tested patch:

--- a/gcc/fortran/expr.c
+++ b/gcc/fortran/expr.c
@@ -3733,17 +3733,14 @@ gfc_has_default_initializer (gfc_symbol *der)
     if (c->ts.type == BT_DERIVED)
       {
         if (!c->attr.pointer
             && gfc_has_default_initializer (c->ts.u.derived))
          return true;
       }
-    else
-      {
-        if (c->initializer)
-         return true;
-      }
+    if (c->initializer)
+      return true;

   return false;
 }

 /* Get an expression for a default initializer.  */

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
                 CC|                            |jakub at gcc dot gnu.org

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011-12-06
     Ever Confirmed|0                           |1

--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2011-12-06 13:59:21 UTC ---
On x86_64-apple-darwin10, the test always gives

 do_job my_data%head is NOT associated

but valgrind complains as in comment #1.

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

--- Comment #4 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-06 18:10:11 UTC ---
Author: burnus
Date: Tue Dec  6 18:10:01 2011
New Revision: 182059

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=182059
Log:
2011-12-06  Tobias Burnus  <burnus@net-b.de>

        PR fortran/51435
        * expr.c (gfc_has_default_initializer): Fix handling of
        DT with initialized pointer components.

2011-12-06  Tobias Burnus  <burnus@net-b.de>

        PR fortran/51435
        * gfortran.dg/default_initialization_5.f90: New.


Added:
    trunk/gcc/testsuite/gfortran.dg/default_initialization_5.f90
Modified:
    trunk/gcc/fortran/ChangeLog
    trunk/gcc/fortran/expr.c
    trunk/gcc/testsuite/ChangeLog

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

--- Comment #5 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-06 19:26:48 UTC ---
Author: burnus
Date: Tue Dec  6 19:26:44 2011
New Revision: 182062

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=182062
Log:
2011-12-06  Tobias Burnus  <burnus@net-b.de>

        PR fortran/51435
        * expr.c (gfc_has_default_initializer): Fix handling of
        DT with initialized pointer components.

2011-12-06  Tobias Burnus  <burnus@net-b.de>

        PR fortran/51435
        * gfortran.dg/default_initialization_5.f90: New.


Added:
   
branches/gcc-4_6-branch/gcc/testsuite/gfortran.dg/default_initialization_5.f90
Modified:
    branches/gcc-4_6-branch/gcc/fortran/ChangeLog
    branches/gcc-4_6-branch/gcc/fortran/expr.c
    branches/gcc-4_6-branch/gcc/testsuite/ChangeLog

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

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

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

--- Comment #6 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-06 19:28:32 UTC ---
FIXED on the trunk (4.7) and on the 4.6 branch.

Thanks for the report!

[Bug fortran/51435] [4.6/4.7 Regression] Bad association status after null() of derived type component

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51435

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jonathan.hogg at stfc dot
                   |                            |ac.uk

--- Comment #7 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-12-20 17:43:43 UTC ---
*** Bug 51639 has been marked as a duplicate of this bug. ***