[f2py] One dimensional array with unbounded dimension

[pearu at cens.ioc.ee]

On Thu, 13 Oct 2005, olusegun ogunbade wrote:

> 
> Dear All,
> 
> I'm a newbie to f2py. I need someone to help me out with this trivial 
> problem. I created a signature file for a fortran code. The subroutine 
> interface is given as:
> 
> subroutine wclbes(zz,eta1,zlmin,nl,fc,gc,fcp,gcp,sig,kfn,mode1,ifail,ipr).
> 
> fc, gc, fcp, gcp and sig are arrays dimensioned (0 : *). On exit, each of 
> these one-dimensional arrays with dimension (0:d) where d is >= nl + 1  may 
> contain the value of a function of order zlmin + n (n = 0,1,...,nl).
> 
> The modified signature file I generated looks like this:
> !    -*- f90 -*-
> ! Note: the context of this file is case sensitive.
> 
> python module wclbes ! in
>     interface  ! in :wclbes
>         subroutine 
> wclbes(zz,eta1,zlmin,nl,fc,gc,fcp,gcp,sig,kfn,mode1,ifail,ipr) ! in 
> :wclbes:wclbes.f
>             complex*16 intent(in) :: zz
>             complex*16 intent(in) :: eta1
>             complex*16 intent(in) :: zlmin
>             integer intent(in) :: nl
>             complex*16 dimension(*), intent(out) :: fc
>             complex*16 dimension(*), intent(out) :: gc
>             complex*16 dimension(*), intent(out) :: fcp
>             complex*16 dimension(*), intent(out) :: gcp
>             complex*16 dimension(*), intent(out) :: sig
>             integer intent(in) :: kfn
>             integer intent(in) :: mode1
>             integer intent(out) :: ifail
>             integer intent(hide) :: ipr = 1
>             integer :: nfp
>             integer :: n11
>             integer :: npq1
>             integer :: npq2
>             integer :: n20
>             integer :: kas1
>             integer :: kas2
>             common /coulc2/ nfp,n11,npq1,npq2,n20,kas1,kas2
>         end subroutine wclbes
>     end interface
> end python module wclbes
> 
> ! This file was auto-generated with f2py (version:2.45.241_1926).
> ! See http://cens.ioc.ee/projects/f2py2e/
> 
> Afterwards, I created the .so library using f2py -c wclbes.pyf wclbes.f 
> without any problem. When I tried to access this from python I got the 
> following:
> 
> Python 2.3.4 (#2, Aug 19 2004, 15:49:40)
> [GCC 3.4.1 (Mandrakelinux (Alpha 3.4.1-3mdk)] on linux2
> Type "copyright", "credits" or "license()" for more information.
> 
>     ****************************************************************
>     Personal firewall software may warn about the connection IDLE
>     makes to its subprocess using this computer's internal loopback
>     interface.  This connection is not visible on any external
>     interface and no data is sent to or received from the Internet.
>     ****************************************************************
> 
> IDLE 1.0.3
> >>>import Numeric
> >>>import wclbes
> >>>print wclbes.wclbes.__doc__
> wclbes - Function signature:
>   fc,gc,fcp,gcp,sig,ifail = wclbes(zz,eta1,zlmin,nl,kfn,mode1)
> Required arguments:
>   zz : input complex
>   eta1 : input complex
>   zlmin : input complex
>   nl : input int
>   kfn : input int
>   mode1 : input int
> Return objects:
>   fc : rank-1 array('D') with bounds (*)
>   gc : rank-1 array('D') with bounds (*)
>   fcp : rank-1 array('D') with bounds (*)
>   gcp : rank-1 array('D') with bounds (*)
>   sig : rank-1 array('D') with bounds (*)
>   ifail : int
> 
> >>>a, b, c, d, e, f = wclbes.wclbes(19.2j, 0., 0.728, 11, 3, -2)
> 
> Traceback (most recent call last):
>   File "<pyshell#5>", line 1, in -toplevel-
>     a, b, c, d, e, f = wclbes.wclbes(19.2j, 0., 0.728, 11, 3, -2)
> error: failed in converting hidden `fcp' of wclbes.wclbes to C/Fortran array
> 
> What is it that I'm doing wrong? Any suggestion is welcome and will be 
> highly valued.

Output arrays should have defined dimensions (f2py should give a warning
on that). So, try

  complex*16 dimension(0:nl+1), intent(out) :: fc,gc,fco,gcp,sig

HTH,
Pearu