[Bug fortran/94221] New: Explicit assignment in type is ignored in some cases

[Bug fortran/94221] New: Explicit assignment in type is ignored in some cases

[r.dilz at tue dot nl]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94221

            Bug ID: 94221
           Summary: Explicit assignment in type is ignored in some cases
           Product: gcc
           Version: 9.2.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
          Assignee: unassigned at gcc dot gnu.org
          Reporter: r.dilz at tue dot nl
  Target Milestone: ---

When working with types, I found some unexpected behavior. 

It seems that the variables in a type do not get assigned their proper values
reliably. I suspect that the compiler does not check well enough whether a
variable is used or not and the whole assignment is then optimized away.



Below is a minimal example: 

module pyimp
  use iso_c_binding
  type,bind(C) :: container
    integer(C_INT) :: val = 3 !assignment here is not reliable in gfortran
  end type container

  contains

    type(container) function test() bind(c,name='test')
      print*,"running test"
      print*,test%val
    end function test

    type(container) function test2() bind(c,name='test2')
      print*,"running test - here test%val is not initialized"
    end function test2

end module pyimp

program main
  use pyimp
  type(container) :: a
  print*,"First: the test%var is initialized by an extra print statement:"
  a = test()
  print*,a%val
  a = test2()
  print*,"But in test2 it is not: "
  print*,a%val
end program main



Compiled with fortran 4.8 and fortran 9.2.1, both give the same output:



 First: the test%var is initialized by an extra print statement:
 running test
           3
           3
 running test - here test%val is not initialized
 But in test2 it is not: 
           0

Whereas in test2 it should also output 3. Using ifort, I do get the expected
result for test2.

[Bug fortran/94221] Explicit assignment in type is ignored in some cases

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=94221

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |10.0
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2020-03-19
     Ever confirmed|0                           |1
           Keywords|                            |wrong-code

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
test2 lacks a return stmt and initialization of the return value:

test2 ()
{
  {
    struct __st_parameter_dt dt_parm.0;

    dt_parm.0.common.filename = &"t.f90"[1]{lb: 1 sz: 1};
    dt_parm.0.common.line = 15;
    dt_parm.0.common.flags = 128;
    dt_parm.0.common.unit = 6;
    _gfortran_st_write (&dt_parm.0);
    _gfortran_transfer_character_write (&dt_parm.0, &"running test - here
test%val is not initialized"[1]{lb: 1 sz: 1}, 47);
    _gfortran_st_write_done (&dt_parm.0);
  }
}

but the caller still expects one:

  a = test ();

indeed a frontend issue.