[Insight-developers] Moving Insight Journal Paper 176

Tue May 12 09:08:02 EDT 2009

Le 12 mai 09 à 14:31, Wes Turner a écrit :

> Gaetan,
>
> 2009/5/11 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>
>
> Hi Wes,
>
> Le 11 mai 09 à 22:20, Wes Turner a écrit :
>
>
> Gaetan,
>
> Two quick issues.
>
>        • If we try to configure the system with CMAKE_BUILD_TYPE set  
> to Release, the compiler crashes (error message is below).  System  
> is Ubuntu, gcc-4.2.4 oor 4.3.2.
>
> I had this kind of problem on mandriva 2009.0 rc1 last year. A  
> compiler upgrade fixed the problem.
> Now that I see that code, it recall me that et should be moved to an  
> helper function, because, IIRC, msvc 6 doesn't like too much the  
> templated method. This small change may help with your compiler  
> issue as well :-)
>
> It's not clear.  Is this something you can help with, or are you  
> proposing that I move this out?

I can't help with the compiler issue.
But the method where your compiler fail needs a small refactoring to  
build with older compilers. This refactoring may help with the  
compiler crash you're experimenting.

I hope that's much clear.

>
>
>
>
>        • The test AttributeValues2 is failing - same system,  
> CMAKE_BUILD_TYPE not specidfied.  The log file is attached.
>
> This one may be caused by my lack of knowledge on how to compare  
> floats in a portable way: the test is ok only on one of my hosts, so  
> I'm not too surprised that it fails for you.
> I really need a clean way to check those values.
>
> I made a change that solves this at least for now.  The issue was  
> that you are scaling the difference by the expected value -- which  
> is good for reasonable values.  For the particular case that was  
> failing, the comparison value  was on the order of e-28 (effectively  
> zero) which is unstable and blows up for even small absolute  
> differences.  I added an additional check of the absolute difference  
> against vnl_math::eps and the test passes. You can review the  
> addition when I check this in, hopefully later today.

Great!

Gaëtan

>
>
> - Wes
>
> Gaëtan
>
>
>
> Any chance you can try and resolve these issues?
>
> Thanks much!
>
> - Wes
>
> [ 14%] Building CXX object CMakeFiles/binary_fillhole.dir/ 
> binary_fillhole.o
> /home/wesley/Packages/InsightJournal/Insight_journal_176_8/ 
> binaryAttributeMorphology/itkLabelMapMaskImageFilter.txx: In member  
> function ‘void itk::LabelMapMaskImageFilter<TInputImage,  
> TOutputImage>::ThreadedGenerateData(typename  
> TInputImage::LabelObjectType*) [with TInputImage =  
> itk::LabelMap<itk::AttributeLabelObject<long unsigned int, 2u, bool>  
> >, TOutputImage = itk::Image<unsigned char, 2u>]’:
> /home/wesley/Packages/InsightJournal/Insight_journal_176_8/ 
> binaryAttributeMorphology/itkLabelMapMaskImageFilter.txx:419:  
> internal compiler error: in get_biv_step, at loop-iv.c:774
> Please submit a full bug report,
> with preprocessed source if appropriate.
> See <URL:http://gcc.gnu.org/bugs.html> for instructions.
> For Debian GNU/Linux specific bug reporting instructions,
> see <URL:file:///usr/share/doc/gcc-4.2/README.Bugs>.
> make[2]: *** [CMakeFiles/binary_fillhole.dir/binary_fillhole.o]  
> Error 1
> make[1]: *** [CMakeFiles/binary_fillhole.dir/all] Error 2
> make: *** [all] Error 2
>
>
> <wes.turner at kitware.com> wrote:
> Thanks.  I will grab your new code and will keep in touch during the  
> process.
>
> - Wes
>
> 2009/5/8 Gaëtan Lehmann <gaetan.lehmann at jouy.inra.fr>
>
>
> Hi Wes,
>
> Thanks a lot for taking this task: that's a quite important amount  
> of work.
>
> Le 6 mai 09 à 21:23, Wes Turner a écrit :
>
>
> Gaetan:
> Your paper 176, Label object representation and manipulation with  
> ITK, looks like a good candidate to be moved into the review  
> directory.  Does the current submission to the IJ represent the  
> current state of the code?
>
> No. There are some changes since the last submission, that I didn't  
> have time to submit to the IJ.
> There are available at http://voxel.jouy.inra.fr/darcs/contrib-itk/binaryAttributeMorphology/binaryAttributeMorphology.tar.gz
>
> The new code include many style fixes that I'm quite sure you would  
> be pleased to not have to fix :-)
>
>
>  Anything else I should know before taking the plunge?
>
>
> I'm not a native english speaker, and I'm not so good at choosing  
> names. So if you have suggestions on names, I would be pleased to  
> discuss them :-)
>
>
> ITK developers:
> This submission is fairly large (around 65 classes).
>
> I'm afraid I've added a few more - I see 71 .h files in my repository.
>
>
>  I will send out a warning before I commit anything to Review and  
> will work to minimize the extent and duration of any breakages, but  
> I thought it would be a good idea to keep you informed before taking  
> any steps.  Your comments and pointers are appreciated.  The  
> reference submission is http://hdl.handle.net/1926/584 if you want  
> to check it out.
>
>
> Gaëtan
>
>
>
> -- 
> Gaëtan Lehmann
> Biologie du Développement et de la Reproduction
> INRA de Jouy-en-Josas (France)
> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> http://voxel.jouy.inra.fr  http://www.mandriva.org
> http://www.itk.org  http://www.clavier-dvorak.org
>
>
>
>
> -- 
> Wesley D. Turner, Ph.D.
> Kitware, Inc.
> R&D Engineer
> 28 Corporate Drive
> Clifton Park, NY 12065-8662
> Phone: 518-371-3971 x120
>
>
>
> -- 
> Wesley D. Turner, Ph.D.
> Kitware, Inc.
> R&D Engineer
> 28 Corporate Drive
> Clifton Park, NY 12065-8662
> Phone: 518-371-3971 x120
> <LastTest.log>
>
> -- 
> Gaëtan Lehmann
> Biologie du Développement et de la Reproduction
> INRA de Jouy-en-Josas (France)
> tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
> http://voxel.jouy.inra.fr  http://www.mandriva.org
> http://www.itk.org  http://www.clavier-dvorak.org
>
>
>
>
> -- 
> Wesley D. Turner, Ph.D.
> Kitware, Inc.
> R&D Engineer
> 28 Corporate Drive
> Clifton Park, NY 12065-8662
> Phone: 518-371-3971 x120

-- 
Gaëtan Lehmann
Biologie du Développement et de la Reproduction
INRA de Jouy-en-Josas (France)
tel: +33 1 34 65 29 66    fax: 01 34 65 29 09
http://voxel.jouy.inra.fr  http://www.mandriva.org
http://www.itk.org  http://www.clavier-dvorak.org

-------------- next part --------------
A non-text attachment was scrubbed...
Name: PGP.sig
Type: application/pgp-signature
Size: 186 bytes
Desc: Ceci est une signature ?lectronique PGP
URL: <http://www.itk.org/mailman/private/insight-developers/attachments/20090512/f3982e18/attachment.pgp>