[Bug fortran/49111] New: Unnecessary warning for private interfaces having the BIND(C) attribute

[Bug fortran/49111] New: Unnecessary warning for private interfaces having the BIND(C) attribute

[jwmwalrus at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49111

           Summary: Unnecessary warning for private interfaces having the
                    BIND(C) attribute
           Product: gcc
           Version: 4.6.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: fortran
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: jwmwalrus@gmail.com


When the following module:

module mod1
    use iso_c_binding
    implicit none
    save

    interface
        function strerror(errnum) bind(C, NAME = 'strerror')
            import
            type(C_PTR) :: strerror
            integer(C_INT), value :: errnum
        end function
    end interface

    private strerror
end module mod1


gfortran warns about the private symbol 'strerror' being given a binding label.
 The exact warning is:

...:~$ gfortran -c  test_gfortran_private_cbinding.f90 
test_gfortran_private_cbinding.f90:8.8:

        function strerror(errnum) bind(C, NAME = 'strerror')
        1
Warning: Symbol 'strerror' at (1) is marked PRIVATE but has been given the
binding label 'strerror'


In sections 5.3.2 and 5.3.5 of the Fortran 2008 standard, there seems to be no
restriction between the BIND and PRIVATE attributes ---i.e., the BIND(C)
attribute is compatible with the PRIVATE attribute, and therefore no warning
should be issued by the gfortran.


The version information is:

...:~$ gfortran -v
Using built-in specs.
COLLECT_GCC=gfortran
COLLECT_LTO_WRAPPER=/usr/lib/x86_64-linux-gnu/gcc/x86_64-linux-gnu/4.6.1/lto-wrapper
Target: x86_64-linux-gnu
Configured with: ../src/configure -v --with-pkgversion='Ubuntu 4.6.0-3~ppa1'
--with-bugurl=file:///usr/share/doc/gcc-4.6/README.Bugs
--enable-languages=c,c++,fortran,objc,obj-c++,go --prefix=/usr
--program-suffix=-4.6 --enable-shared --enable-multiarch
--with-multiarch-defaults=x86_64-linux-gnu --enable-linker-build-id
--with-system-zlib --libexecdir=/usr/lib/x86_64-linux-gnu
--without-included-gettext --enable-threads=posix
--with-gxx-include-dir=/usr/include/c++/4.6 --libdir=/usr/lib/x86_64-linux-gnu
--enable-nls --with-sysroot=/ --enable-clocale=gnu --enable-libstdcxx-debug
--enable-libstdcxx-time=yes --enable-plugin --enable-objc-gc --disable-werror
--with-arch-32=i686 --with-tune=generic --enable-checking=release
--build=x86_64-linux-gnu --host=x86_64-linux-gnu --target=x86_64-linux-gnu
Thread model: posix
gcc version 4.6.1 20110409 (prerelease) (Ubuntu 4.6.0-3~ppa1) 


The system information (for Ubuntu 11.04) is:

...:~$ uname -srvmpio
Linux 2.6.38-8-generic #42-Ubuntu SMP Mon Apr 11 03:31:24 UTC 2011 x86_64
x86_64 x86_64 GNU/Linux

[Bug fortran/49111] Unnecessary warning for private interfaces having the BIND(C) attribute

[jwmwalrus at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49111

--- Comment #1 from John <jwmwalrus at gmail dot com> 2011-05-22 16:14:48 UTC ---
Created attachment 24328
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=24328
Test case

[Bug fortran/49111] Unnecessary warning for private interfaces having the BIND(C) attribute

[burnus at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49111

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
  Attachment #24328|application/octet-stream    |text/plain
          mime type|                            |

--- Comment #2 from Tobias Burnus <burnus at gcc dot gnu.org> 2011-05-22 17:04:39 UTC ---
Comment on attachment 24328
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=24328
Test case

(In reply to comment #0)
> gfortran warns about the private symbol 'strerror' being given a binding label.
> 
> In sections 5.3.2 and 5.3.5 of the Fortran 2008 standard, there seems to be no
> restriction between the BIND and PRIVATE attributes ---[...] and therefore no
> warning should be issued by the gfortran.

The Fortran standard only states whether something is valid or not - not
whether a compiler should issue a warning or not.

The idea is that having PRIVATE (making it inaccessible from outside the module
by direct means) and having a C binding (make it available out side) are in a
way opposite concepts - and thus possible unintended.

However, for an INTERFACE - as opposed to a module procedure - a warning indeed
does not make much sense.

[Bug fortran/49111] Unnecessary warning for private interfaces having the BIND(C) attribute

[craig.powers at gmail dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49111

Craig Powers <craig.powers at gmail dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |craig.powers at gmail dot
                   |                            |com

--- Comment #3 from Craig Powers <craig.powers at gmail dot com> 2012-10-12 20:53:53 UTC ---
I see the accessibility denoted by PRIVATE/PUBLIC as conceptually different
from the accessibility denoted by BIND(C).  In addition to the fact that
BIND(C) does not distinguish an import from an export, there is also the
consideration that one might wish to hide the C-callable routine (marked with
BIND(C)) from other Fortran code by also marking it PRIVATE.

I'm in the process of producing a Fortran interface to a C library, and I find
that there are some instances where I want to declare the import as PRIVATE and
then provide a wrapper routine to take care of the Fortran-to-C stuff that is
mechanical e.g. converting a Fortran array.

Because my Fortran is a little rusty, and this is my first time making any
extended use of the BIND(C) features, the warning made me concerned that I
hadn't accomplished what I intended to accomplish.

I see this warning as far more likely to be a spurious complaint about valid
and intended usage than to be a useful hint about something unintended.

[Bug fortran/49111] Unnecessary warning for private interfaces having the BIND(C) attribute

[dominiq at lps dot ens.fr]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=49111

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |diagnostic
           Priority|P3                          |P4
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2013-12-29
     Ever confirmed|0                           |1
           Severity|normal                      |enhancement

--- Comment #4 from Dominique d'Humieres <dominiq at lps dot ens.fr> ---
Still present at r206227.

[Bug fortran/49111] Unnecessary warning for private interfaces having the BIND(C) attribute

[sandra at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=49111

sandra at gcc dot gnu.org changed:

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

--- Comment #8 from sandra at gcc dot gnu.org ---
Still present on master, so no, this bug hasn't spontaneously fixed itself. 
:-(