Re: Can someone know the time for a oo implemented F2003 complier

Hi,

On Apr 18, 8:48 am, "James Van Buskirk" <not_va...@[EMAIL PROTECTED]
> wrote:
> Spoken like a true C programmer.  Why write a C wrapper to invoke a
> Fortran procedure when Fortran can do so all by itself?  Here is an
> example with cray pointers, could be updated to use procedure
> pointers if they ever come true:

[Code skipped]

Thanks for the replies. I decided to give a try to Cray pointers as a
tem****ary replacement for pointers to procedures, and I was about to
try out the code you provided, but given that apparently it gave a
segmentation fault when trying to compile it, I decided to go for
something simpler :-)

I read this about Cray pointers and gfortran
http://gcc.gnu.org/onlinedocs/gfortran/Cray-pointers.html#Cray-pointers

but I've never used this feature, neither external procedures, so be
gentle :-)

I got a minimal code to try the Cray pointers, as per the
do***entation above:

[angelv@[EMAIL PROTECTED]
 temp]$ cat temp.f90
PROGRAM test
  IMPLICIT NONE

  EXTERNAL sp
  POINTER (subptr,subpte)
  EXTERNAL subpte

  subptr =3D LOC(sp)

  CALL subpte
END PROGRAM test

SUBROUTINE sp
  PRINT*, "HI"
END SUBROUTINE sp



But compilation is unsuccessfull:

[angelv@[EMAIL PROTECTED]
 temp]$ gfortran -fcray-pointer -o temp temp.f90
 In file temp.f90:6

  EXTERNAL subpte
                1
Error: CRAY POINTEE attribute conflicts with EXTERNAL attribute at (1)
 In file temp.f90:10

  CALL subpte
            1
Error: CRAY POINTEE attribute conflicts with SUBROUTINE attribute in
'subpte' at (1)
 In file temp.f90:5

  POINTER (subptr,subpte)
                       1
Error: Symbol 'subpte' at (1) has no IMPLICIT type


[angelv@[EMAIL PROTECTED]
 temp]$ gfortran --version
GNU Fortran (GCC) 4.1.2 20070925 (Red Hat 4.1.2-33)


Any help appreciated. Thanks,
=C1ngel de Vicente