Re: [f2py] line length

Pearu Peterson <[email protected]> Wed, 11 Aug 2010 21:18:24 +0300
Newsgroups gmane.comp.python.f2py.user
Message-ID <[email protected]>
The problem is that by default f2py considers .f files as Fortran 77
source files and F77 does not define module block.
To resolve this problem, add `! -*- fix -*-` to the header of all .f
files. Then f2py considers such .f files as Fortran 90 source files
with fixed format. With this addition, your original example compiles
and imports fine.

Pearu

On Wed, Aug 11, 2010 at 8:59 PM, Al Niessner <[email protected]> wrote:
>
> Ooops. There are two errors in the example I sent. The first is that you
> have to chagne REAL(pr) to REAL(8) in the function declaration. It works
> fine as REAL(pr) with the older f2py on my laptop but fails with a big
> compiler crash on the SVN version. I am guessing that is bug 1. With it
> defined as REAL(8) I get this when loading:
>
> In [1]: import mathsub
> ---------------------------------------------------------------------------
> ImportError                               Traceback (most recent call
> last)
>
> /home/niessner/Scratch/Example/<ipython console> in <module>()
>
> ImportError: ./mathsub.so: undefined symbol: rng_
>
>
> On Wed, 2010-08-11 at 10:44 -0700, Al Niessner wrote:
>> Here it is. I use this command
>>
>> f2py -c kinds.f90 mathsub.f -m mathsub
>>
>> Then when I 'import mathsub' I get the missing _rng symbol. Enjoy.
>>
>> On Tue, 2010-08-10 at 21:56 -0700, Pearu Peterson wrote:
>> > On Wed, Aug 11, 2010 at 12:48 AM, Al Niessner <[email protected]> wrote:
>> > >
>> > > rng is defined as 'REAL(pr) FUNCTION RNG(nts_in)' in mathsub.f. So, in
>> > > the process of debugging this, I created another small test with RNG in
>> > > checks.f90 that you already have. I could not get it to compile -- yes I
>> > > did a 'use kinds' in the checks module -- until I mad it REAL(8)
>> > > FUNCTION RNG(nts_in)'. I can send you the example if you want, but I am
>> > > pretty sure this is an f2py thing.
>> >
>> > Yes, I always look forward for receiving minimal but complete examples
>> > that demonstrate f2py bugs.
>> >
>> > Pearu
>> >
>> > _______________________________________________
>> > f2py-users mailing list
>> > f2py-users-Y4l6ocDipWCuvFJfX82//[email protected]
>> > http://cens.ioc.ee/mailman/listinfo/f2py-users
>>
>
> --
> Al Niessner
> 818.354.0859
>
> --------
> |  dS  | >= 0
> --------
>
>
> _______________________________________________
> f2py-users mailing list
> f2py-users-Y4l6ocDipWCuvFJfX82//[email protected]
> http://cens.ioc.ee/mailman/listinfo/f2py-users
>