[Insight-developers] Module dependcies help : How to move classes out of Review

[M Stauffer (V)]

OK, thanks Luis. Will do. 

>-----Original Message-----
>From: Luis Ibanez [mailto:luis.ibanez at kitware.com] 
>Sent: Saturday, June 18, 2011 2:31 PM
>To: M Stauffer (V)
>Cc: Brad King; Insight Developers
>Subject: Re: [Insight-developers] Module dependcies help : How 
>to move classes out of Review
>
>Hi Michael,
>
>To move classes from the Review directory into their final destination,
>you (or any other developer) must first perform a review of the code
>following this checklist:
>
>http://www.itk.org/Wiki/ITK_Release_4/Modularization/Code_Revie
ws/Checklist
>
>A patch should be prepared addressing any issues where the code fails
>to pass the items of this checklist.
>
>The patch should be submitted to Gerrit, and it should follow the
>standard review procedure.
>
>Please do so independently of any other patch.
>That is, do no mix this "moving out of Review" patch
>with things like adding new classes at the same time.
>
>The destination directories have been identified here:
>
>https://spreadsheets.google.com/a/kitware.com/spreadsheet/ccc?k
>ey=0ApfVzIsZe0VFdFUxdE95WE5kMVE4LTJ6b3ZYQnpRalE&hl=en_US#gid=0
>
>
>    Thanks
>
>
>         Luis
>
>
>--------------------------------
>On Tue, Jun 14, 2011 at 5:35 PM, M Stauffer (V) 
><mstauff at verizon.net> wrote:
>>>From: Brad King [mailto:brad.king at kitware.com]
>>>Sent: Tuesday, June 14, 2011 10:47 AM
>>>To: M Stauffer (V)
>>>Cc: 'ITK-dev-list-mstauff at ver'; Luis Ibanez
>>>Subject: Re: [Insight-developers] Module dependcies help
>>>
>>>On 06/14/2011 10:31 AM, M Stauffer (V) wrote:
>>>> Is the new paradigm to include new classes directly in the 
>module to
>>>> which they most directly belong, forgoing the Review module?
>>>There are
>>>> some classes in Review, e.g. CompositeTransform,  that
>>>> ITK-RegistrationRefactoring depends on, so we'll have to
>>>also move those
>>>> to other modules?
>>>
>>>The Review module exists only to hold classes that came from 
>the Review
>>>directory prior to modularization.  These were left from the
>>>old CVS-based
>>>workflow where the Review directory was used to put things 
>that had not
>>>been reviewed for inclusion.  Now we use Git branches and 
>Gerrit to do
>>>reviews so everything can be committed directly to its final intended
>>>location.
>>>
>>>If you have some classes that need to be moved out of Review 
>into more
>>>permanent modules, please list them.
>>
>> OK that makes sense. I can move some classes out of
>> RegistrationRefactoring to avoid circular dependencies there.
>>
>> The following six classes need to be moved out of Review for 
>our needs.
>> I don't know the official procedure for moving classes out 
>of Review and
>> changing dependencies, but for now I've moved these out and added
>> dependencies as described below to get things to build and test. How
>> should I proceed in terms of eventually merging into itk?
>>
>> 1) itkDeformationFieldTransform* (base class for
>> Core/Transform/itkBSplineDeformationFieldTransform). I moved it to
>> Core/Transform.
>>
>> The following three are needed by DeformationFieldTransform 
>and tests.
>> It looks like they'll require some new dependcies from what 
>I'm seeing
>> while I experiment with moving them.
>>
>> 2) itkBSplineScatteredDataPointSetToImageFilter* - used in
>> itkBSplineDeformationFieldTransform. I moved to Core/Common for now
>> 3) itkCoxDeBoorBSplineKernelFunction* - used in
>> itkBSplineScatteredDataPointSetToImageFilter and in
>> itkBSplineControlPointImageFunction. I moved to Core/Common for now.
>> 4) itkBSplineControlPointImageFunction* - used in
>> itkBSplineScatteredDataPointSetToImageFilterTest. I moved to
>> Numerics/Optimizers.
>>
>> I then added these dependecies:
>>
>> Added ITK-Common to ITK-Transform.
>> Added test dependency ITK-Optimizers to ITK-Common.
>> Added ITK-ImageFunction to ITK-Optimizers.
>>
>> I moved these two to RegistrationRefactoring for now since 
>they're part
>> of our new registration framework:
>>
>> 5)itkCompositeTransform*
>> 6)itkObjectToObjectMetric*
>>
>> Cheers,
>> Michael
>>
>>