[Insight-developers] TransformToDeformationField new Test and Valgrind Errors

[Luis Ibanez]

Hi Hans,

We found one of the problems and have committed a fix:
http://www.itk.org/cgi-bin/viewcvs.cgi/Testing/Code/Review/itkTransformToDeformationFieldSourceTest1.cxx?root=Insight&r1=1.1&r2=1.2&sortby=date

The image buffer was not initialized (no call to FillBuffer()),
and the subsequent for-loop was not visiting all the pixels.

This fixes the UMJ reported in the Resample filter.

There are still remaining Valgrind problems related to

                "itk_zlib_deflate"

due to the use of compression in the nifti file.

If you are confident that this are not bugs in nifti then
we could add them to the Valgrind suppression file.

In the meantime, I'm removing the .gz termination of the output
files in order to avoid the Valgrind errors.
http://www.itk.org/cgi-bin/viewcvs.cgi/Testing/Code/Review/CMakeLists.txt?root=Insight&r1=1.184&r2=1.185&sortby=date



   Please let us know what you think,


      Luis



------------------
Hans Johnson wrote:
> Luis,
> 
> These are valgrind issues that are usually suppressed (at least for the
> reading/writing of gzip compressed files, i.e. nifti). Perhaps there is a
> new hole the suppressions where this new test is the only one that actually
> reads/writes nifti files.
> 
> I'm leaving town right now and won't be able to look at it until Monday.
> 
> Hans
> 
> 
> 
> On 5/1/09 2:57 PM, "Luis Ibanez" <luis.ibanez at kitware.com> wrote:
> 
> 
>>Hi Hans,
>>
>>Thanks for taking care of Bug #8930.
>>
>>
>>There is however, a remaning issue,
>>the new tests that was added on Wednesday:
>>
>>http://www.itk.org/cgi-bin/viewcvs.cgi/Testing/Code/Review/CMakeLists.txt?root
>>=Insight&r1=1.179&r2=1.180
>>
>>http://www.itk.org/cgi-bin/viewcvs.cgi/Testing/Code/Review/itkTransformToDefor
>>mationFieldSourceTest1.cxx?root=Insight&sortby=date&view=log
>>
>>
>>Is producing several Valgrind errors today:
>>
>>    http://www.cdash.org/CDash/viewDynamicAnalysisFile.php?id=326791
>>
>>
>>
>>The two other tests for the same class do not produce Valgrind
>>errors, so it would seem that there is something particular
>>about this specific .cxx file that may need fixing.
>>
>>
>>
>>    Any hints on what may be the source of the problem ?
>>
>>
>>        Thanks
>>
>>
>>
>>           Luis
>>
>>
> 
> 
>