[Insight-developers] tests failed but why ?

[Miller, James V (Research)]

I believe that if a test writes to stderr, it is considered a failure.

However, I think I am surprised that the stderr was not captured.  

Jim

-----Original Message-----
From: insight-developers-bounces+millerjv=crd.ge.com at itk.org
[mailto:insight-developers-bounces+millerjv=crd.ge.com at itk.org]On Behalf
Of Gaetan Lehmann
Sent: Wednesday, January 04, 2006 5:52 PM
To: insight-developers at itk.org
Cc: norman-k-williams at uiowa.edu
Subject: Re: [Insight-developers] tests failed but why ?



I have found the problem: I have modified the default values of 
RescaleIntensityImageFilter to be usable without setting values

The problem of the message reported by the insight testing system is that it 
include only the standard output and not the error output.
Can it be modified to include it ?

The other comment made earlier is still valid: it would be nice to be able to 
cancel a submission if the tests are failing

Also, can the default values of RescaleIntensityImageFilter be modified to 
make the filter usable with the default values ? The changes are quite 
simple:

diff -rN 
old-itk-mima2/Insight/Code/BasicFilters/itkRescaleIntensityImageFilter.txx 
new-itk-mima2/Insight/Code/BasicFilters/itkRescaleIntensityImageFilter.txx
36,37c36,37
<   m_OutputMaximum   = NumericTraits<OutputPixelType>::Zero;
<   m_OutputMinimum   = NumericTraits<OutputPixelType>::max();
---
>   m_OutputMaximum   = NumericTraits<OutputPixelType>::max();
>   m_OutputMinimum   = NumericTraits<OutputPixelType>::NonpositiveMin();


Thanks,

Gaetan




On Wednesday 04 January 2006 16:53, Gaetan Lehmann wrote:
> I'd be happy to do the exercise :-)
> Can someone please tell what happened ?
>
> Thanks
>
> Gaetan
>
> On Wed, 04 Jan 2006 16:23:40 +0100, Kent Williams
>
> <kent at psychiatry.uiowa.edu> wrote:
> > It means that your test program raised an exception, or dumped core.
> > That's actually good -- an honest core dump is easy to trace, since
> > something unmistakably bad happened, loudly.  Much better than silently
> > getting the wrong results.
> >
> > Someone who has access to the system upon which the test failed needs to
> > run the test under the debugger, at which point, WHAT happened will be
> > quite clear. HOW and WHY it happened, and how it can be fixed, is left
> > as an exercise for the reader.
> >
> > Gaetan Lehmann wrote:
> >> Hi Luis,
> >>
> >> Richard have submitted the new revision of the paper, which included
> >> some  changes to use the same images in the test and the article, and
> >> to verify  the images used in the article.
> >> The new tests I have added have failed on the insight automatic
> >> testing  system, but works without problem on my host.
> >>
> >> It would be nice to be able to run the test before validating the
> >> submission. For example, a mail can be sent if one are several test
> >> have  failed to ask to the author if he want to submit its revision
> >> anyway.
> >>
> >> Also, the error message is quite non informative "*** Exception
> >> executing:  Child aborted". Any idea about how to debug that ?
> >>
> >> Regards,
> >>
> >> Gaetan
> >
> > _______________________________________________
> > Insight-developers mailing list
> > Insight-developers at itk.org
> > http://www.itk.org/mailman/listinfo/insight-developers