[Insight-developers] Re: integrating the new algorithms for existing filters in ITK (preparing ITK 3.2)

[Gaëtan Lehmann]

Hi,

I have commited the ConnectedComponentImageFilter changes.
I'll test the changes made by Richard in the reconstruction filters and  
commit the changes tomorrow if everything is fine with the connected  
component filter.

Gaetan


Le Tue, 12 Dec 2006 22:37:12 +0100, Luis Ibanez <luis.ibanez at kitware.com>  
a écrit:

>
> Hi Gaetan,
>
> The changes that you describe don't seem to raise
> any backward compatibility issues.
>
>
> We will just need to verify the coding style and
> make sure that they work in all the platforms.
>
>
>    Please go ahead and commit your changes.
>
>
> The best time to do this is early in the morning
> so that we have the opportunity to fix any issues
> during the day by following the continuous builds.
>
>
>
> BTW, Are these the same files from which Richard
>       wanted to remove the code intended for profiling ?
>
>
>
>     Thanks
>
>
>       Luis
>
>
> ------------------------
> Gaetan Lehmann wrote:
>> On Tue, 12 Dec 2006 00:43:15 +0100, Luis Ibanez  
>> <luis.ibanez at kitware.com>  wrote:
>>
>>>
>>> Hi Gaetan,
>>>
>>>
>>>      Thanks for pointing this out.
>>>
>>>
>>> When dealing with enhancements, they will go directly into the
>>> respective filters. That is, as long as they do not involve the
>>> addition of new classes or files, and do not require the
>>> modification of the external API of existing classes.
>>   The api is not modified, but there is some new files:
>> In order to avoid code duplication,  
>> ReconstructionByDilationImageFilter  and  
>> ReconstructionByErosionImageFilter are implemented in a common class,   
>> ReconstructionImageFilter, which comes with 2 new files   
>> itkReconstructionImageFilter.h and itkReconstructionImageFilter.txx,  
>> and a  specialization to get one result or the other is done in the   
>> itkReconstructionByDilationImageFilter.h and   
>> itkReconstructionByErosionImageFilter.h. the corresponding .txx file  
>> no  more exist.
>> Also, there is a itkConnectedComponentAlgorithm.h file to store some   
>> functions to initialize the iterators depending of the connectivity   
>> option. This file is shared by several of our contributions.
>>  so there is 3 new files, and 2 removed.
>> Please let me known if that's a problem.
>>
>>>
>>>
>>> We must in any case go through a code review and a verification
>>> of backward compatibility. Please make sure that there are tests
>>> already in place for the classes that you plan to modify. If any
>>> of the classes is missing a test, then let's work out to add a
>>> test before making any modifications to the class.
>>>
>>> One interesting place to check is in the current code coverage
>>> of the classes:
>>> http://www.itk.org/Testing/Sites/zion.kitware/Linux-g++-3.4/20061211-0100-Nightly/CoverageByName.html  
>>>
>>> If Richard's papers do not require to modify the API, then please
>>> go ahead and commit them into the repository.
>>>
>>> The best way to do it is to start in the morning of the Dashboard
>>> (e.g. avoid an evening submission). You probably want to commit
>>> them tomorrow *after* 2:00am EST (8:00am France), so that we will
>>> have time to fix any issues that may appear during the continuous
>>> builds before we hit the time of Nightly submissions (9:00pm EST).
>>>
>>>
>>> Submitting an Experimental first from your machine will be also
>>> a great idea.
>>   Done
>> If the addition of the new files is ok, I'll commit the changes  
>> tomorrow  morning.
>>  Thanks,
>>  Gaetan
>>
>>>
>>>
>>>   Thanks
>>>
>>>
>>>     Luis
>>>
>>>
>>> -----------------------------
>>> Gaëtan Lehmann wrote:
>>>
>>>> Le Tue, 12 Dec 2006 00:16:03 +0100, Luis Ibanez   
>>>> <luis.ibanez at kitware.com>  a écrit:
>>>>
>>>>> Please let us know if you have any questions,
>>>>
>>>>   The Richard's filters are enhancements of existing filters. Should   
>>>> they  also go in the Review directory ?
>>>>  Thanks,
>>>>  gaetan
>>>>
>>



-- 
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