[Bug fortran/47054] New: Compilation error when cray pointers are declared in both host and internal subroutines

[Bug fortran/47054] New: Compilation error when cray pointers are declared in both host and internal subroutines

[deji_aking at yahoo dot ca]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

           Summary: Compilation error when cray pointers are declared in
                    both host and internal subroutines
           Product: gcc
           Version: unknown
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: deji_aking@yahoo.ca


The following code produces compilation error when compiled with gfortran-4.5.1
(gcc version 4.5.1 20101130 (Red Hat 4.5.1-6) (GCC)). It compiles fine if I
remove the 'contains' statement, and replace it with 'end subroutine host_sub'.

>>>>>
subroutine host_sub(F_su,F_nk)
   implicit none

   integer :: F_nk
   real,dimension(F_nk) :: F_su
      integer G_ni, G_nj
      real*8 G_xg_8, G_yg_8
      pointer (paxg_8, G_xg_8(G_ni))
      pointer (payg_8, G_yg_8(G_nj))
      common / G_p / paxg_8,payg_8
!
      common / G / G_ni, G_nj

   call internal_sub(F_su,F_nk)
   return
contains 

   subroutine internal_sub(F_su,F_nk)
      implicit none
!
      integer G_ni, G_nj
      real*8 G_xg_8, G_yg_8
      pointer (paxg_8, G_xg_8(G_ni))
      pointer (payg_8, G_yg_8(G_nj))
      common / G_p / paxg_8,payg_8
!
      common / G / G_ni, G_nj

      integer :: F_nk
      real,dimension(F_nk) :: F_su 
      integer k,k2

      k2 = 0
      do k = 1, F_nk, 2
         k2 = k2+1
               F_su(k) = F_su(k) + 1.0
      enddo
      return
   end subroutine internal_sub
end subroutine host_sub
>>>>>

[Bug fortran/47054] Compilation error when cray pointers are declared in both host and internal subroutines

[kargl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

kargl at gcc dot gnu.org changed:

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

--- Comment #1 from kargl at gcc dot gnu.org 2010-12-23 19:23:14 UTC ---
What compilation error?

[Bug fortran/47054] Compilation error when cray pointers are declared in both host and internal subroutines

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

--- Comment #2 from Dominique d'Humieres <dominiq at lps dot ens.fr> 2010-12-23 19:33:04 UTC ---
> What compilation error?

For me it is

[macbook] f90/bug% gfc -fcray-pointer pr47054.f90
pr47054.f90:23.29:

      pointer (paxg_8, G_xg_8(G_ni))
                             1
Error: Symbol 'g_xg_8' at (1) has no IMPLICIT type
pr47054.f90:24.29:

      pointer (payg_8, G_yg_8(G_nj))
                             1
Error: Symbol 'g_yg_8' at (1) has no IMPLICIT type

[Bug fortran/47054] Compilation error when cray pointers are declared in both host and internal subroutines

[deji_aking at yahoo dot ca]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

Deji Akingunola <deji_aking at yahoo dot ca> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|unknown                     |4.5.1

--- Comment #3 from Deji Akingunola <deji_aking at yahoo dot ca> 2011-01-03 15:50:58 UTC ---
(In reply to comment #1)
> What compilation error?

Sorry I missed stating the actual error earlier. It is the same as in comment
#2 above.

[Bug fortran/47054] Compilation error when cray pointers are declared in both host and internal subroutines

[kargl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

--- Comment #4 from kargl at gcc dot gnu.org 2011-01-03 22:39:37 UTC ---
Reduced testcase.

subroutine host_sub
   implicit none
   real xg
   pointer (paxg, xg)
   call internal_sub
   contains 
      subroutine internal_sub
         implicit none
         real xg
         pointer (paxg, xg)
      end subroutine internal_sub
end subroutine host_sub


It seems that gfortran is getting confused
by host association, the symbol names, and
the pointer() in internal_sub.

[Bug fortran/47054] Compilation error when cray pointers are declared in both host and internal subroutines

[kargl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

kargl at gcc dot gnu.org changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P4
      Known to fail|                            |4.2.5, 4.3.6, 4.4.6, 4.5.1,
                   |                            |4.6.0

[Bug fortran/47054] Compilation error when cray pointers are declared in both host and internal subroutines

[kargl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

--- Comment #5 from kargl at gcc dot gnu.org 2011-01-03 23:58:41 UTC ---
The problem appears to be in decl.c(variable_decl)

  /*  If this symbol has already shown up in a Cray Pointer declaration,
      then we want to set the type & bail out.  */
  if (gfc_option.flag_cray_pointer)
    {
      gfc_find_symbol (name, gfc_current_ns, 1, &sym);
      if (sym != NULL && sym->attr.cray_pointee)
    {
      sym->ts.type = current_ts.type;
      sym->ts.kind = current_ts.kind;

When gfc_find_symbol() returns, sym == NULL, so the
if block is skipped.  current_ts.type is BT_REAL and
for the reduced testcase current_ts.kind = 4, which 
is want we want.  I think we may want to have a

if (sym == NULL &&  sym->attr.cray_pointee)
   insert_name_into_current_namespace.

Unfortnately, I can't remember how to do that
and have no time to investigate further.

[Bug fortran/47054] Compilation error when cray pointers are declared in both host and internal subroutines

[dominiq at lps dot ens.fr]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=47054

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-05-13
     Ever confirmed|0                           |1

--- Comment #6 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Still failing at r210330 (4.10.0).