Re: Test failures associated with FFTW

[John W. Eaton]

On 03/16/2013 03:33 AM, Daniel J Sebald wrote:
> Rik (maybe),
>
> I ran "make test" and had a failure for
>
> > > > > > processing
> > > > > > /usr/local/src/octave/octave-dialog/build2/libinterp/dldfcn/fftw.cc-tst
> ***** testif HAVE_FFTW
> def_method = fftw ("planner");
> unwind_protect
> method = "estimate";
> fftw ("planner", method);
> assert (fftw ("planner"), method);
> method = "measure";
> fftw ("planner", method);
> assert (fftw ("planner"), method);
> method = "patient";
> fftw ("planner", method);
> assert (fftw ("planner"), method);
> method = "exhaustive";
> fftw ("planner", method);
> assert (fftw ("planner"), method);
> method = "hybrid";
> fftw ("planner", method);
> assert (fftw ("planner"), method);
> unwind_protect_cleanup
> fftw ("planner", def_method);
> end_unwind_protect
> !!!!! test failed
> value on right hand side of assignment is undefined
>
> My conclusion is that this test is not being properly excluded. I run
> lines individually and immediately get:
>
> octave:1> def_method = fftw ("planner");
> warning: fftw: this copy of Octave was not configured to use the FFTW3
> planner
> error: value on right hand side of assignment is undefined
>
> so perhaps somewhere in the "configure" process the HAVE_FFTW should be
> defined as null or logical 0???
>
> Doing a search for "HAVE_FFTW" in addition to some other places I find:
>
> config.log:AMCOND_HAVE_FFTW_FALSE=''
> config.log:AMCOND_HAVE_FFTW_TRUE='#'
> config.status:S["AMCOND_HAVE_FFTW_FALSE"]=""
> config.status:S["AMCOND_HAVE_FFTW_TRUE"]="#"
>
> Is something being overlooked there?

What does

  octave_config_info ("features")

show?  This is a list of all the HAVE_X macros, with the common "HAVE_" 
part omitted.  On my system, it has FFTW3, but not FFTW, so it looks 
like the testif check is wrong.

jwe