[Insight-developers] A few minor consistency issues

[Zachary Pincus]

Hi folks,

In the course of working on the new ITK wrapping system that Gatean  
and Benoit have developed, we've found a few little consistency  
problems in the ITK codebase that I'd like to fix.

I'll list the issues here, along with the proposed fix. Barring  
suggestions to the contrary, I'll file bugs and commit the fixes here  
proposed later this week.

(1) itkTernaryFunctorImageFilter.h
The SetFunctor() method doesn't compile under gcc4, because operator! 
= isn't being provided for the functor by the compiler.

Here's what the method looks like:
     void SetFunctor(const FunctorType& functor)
     {
      if ( m_Functor != functor )
        {
         m_Functor = functor;
         this->Modified();
        }
     }

The corresponding method in itkBinaryFunctorImageFilter.h is as follows:
   void SetFunctor(const FunctorType& functor)
   {
     m_Functor = functor;
     this->Modified();
   }

I propose removing the test in the Ternary case to make it look like  
the Binary case. Are there any other suggestions, or would this be  
the best approach?
(I can provide test code to show the error.)

(2) itkSpatialObject.h does not have a private, unimplemented copy  
constructor and operator=. I propose adding the same.

(3) itkSpatialObjectTreeNode.h does not have a private, unimplemented  
copy constructor and operator=. I propose adding the same.


Thanks,

Zach Pincus

Program in Biomedical Informatics and Department of Biochemistry
Stanford University School of Medicine