[Bug fortran/47984] New: Pointer dummy argument mismatch not detected by Fortran compiler

[Bug fortran/47984] New: Pointer dummy argument mismatch not detected by Fortran compiler

[thenlich at users dot sourceforge.net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984

           Summary: Pointer dummy argument mismatch not detected by
                    Fortran compiler
           Product: gcc
           Version: 4.6.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: thenlich@users.sourceforge.net


Created attachment 23539
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=23539
Test case

The Fortran compiler does not detect a mismatch between an actual argument (not
a pointer) and a dummy argument in a procedure (which is a pointer, by mistake)
in this example:

program test_pointer
    type :: t
        integer :: i = 7
    end type
    type(t), target :: a
    call testsub(a%i)
contains
    subroutine testsub(p)
        integer, pointer, intent(in) :: p
        print *, p
    end subroutine
end

Result: The program compiles without errors.

Expected result: The compiler should report an error like

test_pointer.f90:6.17:

    call testsub(a%i)
                 1
Error: Actual argument for 'p' must be a pointer at (1)

[Bug fortran/47984] Pointer dummy argument mismatch not detected by Fortran compiler

[kargl at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984

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 2011-03-04 15:46:46 UTC ---
troutmask:sgk[205] gfc4x -o z -std=f2003 a.f90
a.f90:6.17:

    call testsub(a%i)
                 1
Error: Fortran 2008: Non-pointer actual argument at (1) to pointer dummy 'p'


troutmask:sgk[212] svn blame interface.c | more

163262     burnus       if (a->expr->expr_type != EXPR_NULL
163262     burnus         && (gfc_option.allow_std & GFC_STD_F2008) == 0
163262     burnus         && compare_pointer (f->sym, a->expr) == 2)
163262     burnus       {
163262     burnus         if (where)
163262     burnus           gfc_error ("Fortran 2008: Non-pointer actual
argument at %L to "
163262     burnus                      "pointer dummy '%s'",
&a->expr->where,f->sym->name);
163262     burnus         return 0;
163262     burnus       }

troutmask:sgk[213] svn log -r 163262 |more
------------------------------------------------------------------------
r163262 | burnus | 2010-08-15 08:47:11 -0700 (Sun, 15 Aug 2010) | 12 lines

2010-08-15  Tobias Burnus  <burnus@net-b.de>

        * interface.c (compare_pointer, ): Allow passing TARGETs to
        pointers dummies with intent(in).

[Bug fortran/47984] Pointer dummy argument mismatch not detected by Fortran compiler

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |RESOLVED
                 CC|                            |burnus at gcc dot gnu.org
         Resolution|                            |INVALID

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-03-04 17:09:50 UTC ---
And without INTENT(IN):

    call testsub(a%i)
                 1
Error: Actual argument for 'p' must be a pointer at (1)


Close as INVALID. If you think that's an error, feel free to reopen.

[Bug fortran/47984] Pointer dummy argument mismatch not detected by Fortran compiler

[thenlich at users dot sourceforge.net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984

--- Comment #3 from Thomas Henlich <thenlich at users dot sourceforge.net> 2011-03-04 18:58:06 UTC ---
Sorry, I don't understand why you consider the bug report invalid. You may very
well be correct, but please explain. I am not such a Fortran expert, so I may
have missed something here.

[Bug fortran/47984] Pointer dummy argument mismatch not detected by Fortran compiler

[sgk at troutmask dot apl.washington.edu]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984

--- Comment #4 from Steve Kargl <sgk at troutmask dot apl.washington.edu> 2011-03-04 19:16:27 UTC ---
On Fri, Mar 04, 2011 at 06:58:19PM +0000, thenlich at users dot sourceforge.net
wrote:
> http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984
> 
> --- Comment #3 from Thomas Henlich <thenlich at users dot sourceforge.net> 2011-03-04 18:58:06 UTC ---
> Sorry, I don't understand why you consider the bug report invalid. You may very
> well be correct, but please explain. I am not such a Fortran expert, so I may
> have missed something here.
> 

gfortran has some support for Fortran 2003 and Fortran 2008
features.  Under Fortran 2008, your program is a valid program.
Under Fortran 2003 (and earlier), your program is invalid.
By default gfortran accepts everything including vendor
extensions.  If you want to restrict gfortran to a particular
standard, then you need to use the -std option.

-std=gnu    Default behavior. Accept a whole lotta garbage.
-std=f95    Restricted to Fortran 95 standard.
-std=f2003  Restricted to Fortran 2003 standard.
-std=f2008  Restricted to Fortran 2008 standard.

[Bug fortran/47984] Pointer dummy argument mismatch not detected by Fortran compiler

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984

--- Comment #5 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-03-04 19:39:32 UTC ---
(In reply to comment #3)
> Sorry, I don't understand why you consider the bug report invalid.
> You may very well be correct, but please explain. I am not such a
> Fortran expert, so I may have missed something here.


I thought the error message (with -std=f2003) in comment 1 was rather clear,
seemingly that's not the case:


    call testsub(a%i)
                 1
Error: Fortran 2008: Non-pointer actual argument at (1) to pointer dummy 'p'


That means: It is allowed in Fortran 2008, but not before. (See below for the
details.)


  * * *


If one looks into the Fortran 2008 standard
(http://gcc.gnu.org/wiki/GFortranStandards), one finds the following:


"12.5.2.7 Pointer dummy variables"


"If the dummy argument does not have the INTENT (IN), the actual argument shall
be a pointer. Otherwise, the actual argument shall be a pointer or a valid
target for the dummy pointer in a pointer assignment statement. If the actual
argument is not a pointer, the dummy pointer becomes pointer associated with
the actual argument."


Thus, the crucial points are:

(a) the dummy argument has INTENT(IN) - which "p" has
(b) the actual argument has the TARGET attribute which "a" and thus "a%i" has


The INTENT(IN) for the pointer ensures that the pointer association is not
modified -- and TARGET avoids alias issues. Thus, there is no technical reason
for not allowing it -- which the standardization committee seemingly also
realized. Thus, it is now allowed in Fortran 2008.

 * * *

For an overview about the new Fortran 2008 features, see also
  ftp://ftp.nag.co.uk/sc22wg5/N1801-N1850/N1828.pdf

For the already implemented F2008 features, see
  http://gcc.gnu.org/wiki/Fortran2008Status
or in the release notes:
  http://gcc.gnu.org/wiki/GFortran#news
  http://gcc.gnu.org/gcc-4.6/changes.html

[Bug fortran/47984] Pointer dummy argument mismatch not detected by Fortran compiler

[thenlich at users dot sourceforge.net]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=47984

--- Comment #6 from Thomas Henlich <thenlich at users dot sourceforge.net> 2011-03-07 06:26:24 UTC ---
RTFM, I should have.

Thank you for your help.