[Bug fortran/58644] New: [OOP] Missing .data ref in passing a CLASS array as actual argument to a TYPE.

[Bug fortran/58644] New: [OOP] Missing .data ref in passing a CLASS array as actual argument to a TYPE.

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58644

            Bug ID: 58644
           Summary: [OOP] Missing .data ref in passing a CLASS array as
                    actual argument to a TYPE.
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Keywords: wrong-code
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: burnus at gcc dot gnu.org

Found while looking at the dump. The code seems to work at run time on all
targets.


Looking at gfortran.dg/class_to_type_2.f90, one has:

  function g () result(res)
    class(foo), allocatable :: res(:)
    allocate (res(3))
    res(:)%i = 55
  end function g

  subroutine subpr2_array (x)
    type(foo) :: x(:)
    if (any(x(:)%i /= 55)) call abort ()
  end subroutine

  call subpr2_array (g ())



For that one gets the dump:
          struct __class_mod_subpr_Foo_1_0a D.2108;
          D.2108 = g ();
          subpr2_array (&D.2108);

I believe that should be instead:
          subpr2_array (&D.2108.data);

As D.2108 is not a pointer and .data is the first field, it seems to work,
though.

[Bug fortran/58644] [OOP] Missing .data ref in passing a CLASS array as actual argument to a TYPE.

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=58644

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |WAITING
   Last reconfirmed|                            |2014-03-22
     Ever confirmed|0                           |1

--- Comment #1 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
I don't understand this PR. The test succeeds even when compiled with
-fsanitize=address.

[Bug fortran/58644] [OOP] Missing .data ref in passing a CLASS array as actual argument to a TYPE.

[pault at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58644

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

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

--- Comment #2 from Paul Thomas <pault at gcc dot gnu.org> ---
(In reply to Dominique d'Humieres from comment #1)
> I don't understand this PR. The test succeeds even when compiled with
> -fsanitize=address.

I think that the PR is justified. The actual argument should be as Tobias says.
Both &D.2108 and &D.2108.data are valid address.  Not only this, they are the
same address.  However, should we ever change the class ABI to put the vptr
first, for example, then this will no longer work.

I believe that my imminent fix for PR63205 will correct this issue and so
eliminate this PR.

Paul

[Bug fortran/58644] [OOP] Missing .data ref in passing a CLASS array as actual argument to a TYPE.

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=58644

--- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
> I believe that my imminent fix for PR63205 will correct this issue
> and so eliminate this PR.

Does it? If yes, this PR should be closed as FIXED. If no, what is missing?