[Insight-developers] Compile error associated with wrapping -- Linux GCC 4.0.2

[Andinet Enquobahrie]

Kent,

thanks for pointing out the problem...

> .../Insight/Code/Common/itkImageFunction.h: In member function ‘void 
> itk::ImageFunction<TInputImage, TOutput, 
> TCoordRep>::ConvertPointToContinousIndex(const itk::Point<TCoordRep, 
> itk::ImageFunction<TInputImage,TOutput,TCoordRep>::ImageDimension>&, 
> typename TInputImage::IndexType&) const [with TInputImage = 
> itk::Image<float, 2u>, TOutput = double, TCoordRep = double]’:
> /scratch/kent/KWW/build/Insight/Wrapping/CSwig/CommonA/wrap_ITKInterpolatorsTcl.cxx:1623: 
> instantiated from here
> .../Insight/Code/Common/itkImageFunction.h:178: error: no matching 
> function for call to ‘itk::Image<float, 
> 2u>::TransformPhysicalPointToContinuousIndex(const itk::Point<double, 
> 2u>&, itk::Index<2u>&) const’
> /scratch/kent/KWW/build/Insight/Wrapping/CSwig/CommonA/wrap_ITKInterpolatorsTcl.cxx:1623: 
> instantiated from here
> .../Insight/Code/Common/itkImageFunction.h:178: error: no matching 
> function for call to ‘itk::Image<float, 
> 2u>::TransformPhysicalPointToContinuousIndex(const itk::Point<double, 
> 2u>&, itk::Index<2u>&) const’
>
> Apparently Andinet added ConvertPointToContinuousIndex to 
> itkImageFunction.h, today, which is what is freaking out GCC4.

There was a bug in the new method added (incorrect continuous index 
type). I have fixed the problem ( revision 1.38 )

Let me know if you still have the problem

-Andinet



>
> The definition of 
> itk::Image::TransformPhysicalPointToContinuousIndex() is a bit tricky, 
> in that it's a templated member function --- I presume it's templated 
> over TCoordRep so that the point parameter can be any itk::Point type 
> of the right dimension, instead of being stuck with one type:
>
> template<class TCoordRep>
> bool TransformPhysicalPointToContinuousIndex(
> const Point<TCoordRep, VImageDimension>& point,
> ContinuousIndex<TCoordRep, VImageDimension>& index ) const
> {
> [ implementation ...]
> }
>
> But in the case of wrapping, that gets the compiler's panties all in a 
> bunch. Any suggestions for getting this working with GCC4 and wrapping?
>



> Any suggestions?
>
> _______________________________________________
> Insight-developers mailing list
> Insight-developers at itk.org
> http://www.itk.org/mailman/listinfo/insight-developers
>
>