STYLE: Make PrintSelf implementation style consistent

[jhlegarreta]

Make PrintSelfimplementation style consistent following the ITK SWG
coding style guideline and the available ITK macros:

• Always print the superclass first.
• Print only the class instance variables.
• Use the os << indent << "{ivarName}: " << m_ivarName << std::endl
  recipe: do not print the leading m_ of the ivar; use a colon after
  printing its name; leave a whitespace between the colon and the
  printed value of the ivar.
• For boolean ivars print On/Off according to their values using the
  ternary operator.
• Use itkPrintSelfObjectMacro to print objects inheriting from
  itk::LightObject that can be null.
• For pointers that do not inherit from itk::LightObject check whether
  they are non-null before attempting to print them and print the
  (empty) string otherwise.
• Rely on the self printing ability of the ITK classes to avoid
  boilerplate code when printing the corresponding ivars in other
  classes' PrintSelf methods.
• Take advantage of the output stream << insertion operator overload
  for std::vector types in the itk::print_helper namespace defined
  in itkPrintHelper.h to print such types consistently.
• Cast custom types to their numerical values using
  itk::NumericTraits::PrinType so that they are printed as expected.
• Remove statements printing the hard-coded class name and its own
  pointer, e.g.

  os << indent << "ArrowSpatialObject(" << this << ")" << std::endl;
  

  PrintSelf does this without requiring the removed line.
• Save typing this-> to point to the instance of the current class.

Apply the same principles to the << operator overloads.

Overload the ostream operator for the FixedImageSamplePoint class used
by the itk::ImageToImageMetric class.

Overload the ostream operator for the ImageVoxel class used by the
itk::DeformableSimplexMesh3DGradientConstraintForceFilter class.

Overload the ostream operator for the ImageVoxel class used by the
itk::DeformableSimplexMesh3DGradientConstraintForceFilter class.

PR Checklist

• No API changes were made (or the changes have been approved)
• No major design changes were made (or the changes have been approved)

[dzenanz]

Looks great! There are still a few compile errors.

[jhlegarreta]

There are still a few compile errors.

I know. Still WIP/draft. I have other changes locally, but unsure whether I should split it across at least 2 PRs, as it is becoming difficult to deal with. Also it is not intended to be comprehensive, as unfortunately, this is not an automated change.

I also grant that some changes belong to ENH, while others are more STYLE, but it is not evident to me that they should be split/how to split them without creating dozens of commits/PRs then even more difficult to deal with locally.

[dzenanz]

Having it as a single commit is fine. I know that splitting into themed commits requires extra effort.

[blowekamp]

Amazing work getting this more consistent!

Use itkPrintSelfObjectMacro to print objects that can be null.
It may be clearer to say that this should be used for all pointer like objects, as pointers could be null.

One other issue that some time occurs with the print self methods is when "pixel types" are char. When they are a char then they are printed as an ASCII char and not a numeric value as expected. This is a case where the NumbericTraits::PrintType must be used. This is clearly separate work.

[jhlegarreta]

Amazing work getting this more consistent!

Thanks Brad.

Use itkPrintSelfObjectMacro to print objects that can be null.
It may be clearer to say that this should be used for all pointer like objects, as pointers could be null.

Will rework the commit message as time permits and as I revise all the changes contained.

One other issue that some time occurs with the print self methods is when "pixel types" are char. When they are a char then they are printed as an ASCII char and not a numeric value as expected. This is a case where the NumbericTraits::PrintType must be used. This is clearly separate work.

Yes, had that in mind, but did not take the chance to make the changes. Will try to do that for the ivars that I'll be streaming, but will not be checking the existing ones as a general rule.

Note that this process is not automated, and I do not intend it to be comprehensive across the entire code base, but just a first step: I have mainly looked at some given patterns (like containing equal signs, etc.). Also, it will take me some time until I can get a working version. For now, it is still WIP.

[dzenanz]

Looks good on a glance.

[issakomi]

M:\Dashboard\ITK\Modules\Numerics\FEM\include\itkImageToRectilinearFEMObjectFilter.hxx(302,3): error C3861: 'itkPrintSelfObject': identifier not found [M:\Dashboard\ITK-build\Modules\Registration\FEM\test\ITKFEMRegistrationTestDriver.vcxproj]

  itkPrintSelfObject(Material);
  itkPrintSelfObjectMacro(Element);

Does itkPrintSelfObject exist? S. this line

[jhlegarreta]

Again, thanks for the heads-up @issakomi. You're right, it should be itkPrintSelfObjectMacro. I'll push another patch set.

[issakomi]

Thanks @jhlegarreta.
BTW, there are many other errors from this PR in Modules/Registration/FEM, s.
https://open.cdash.org/viewBuildError.php?buildid=8462352

[jhlegarreta]

BTW, there are many other errors from this PR in Modules/Registration/FEM, s.
https://open.cdash.org/viewBuildError.php?buildid=8462352

Oh, my bad. I did not have the FEM modules turned on and blindly relied on my local build/CI results, and did not spot those in the yesterday dashboard result. Thanks for the heads-up @issakomi.

[jhlegarreta]

Again, thanks for the heads-up @issakomi. You're right, it should be itkPrintSelfObjectMacro. I'll push another patch set.

Oh, my bad. I did not have the FEM modules turned on and blindly relied on my local build/CI results, and did not spot those in the yesterday dashboard result. Thanks for the heads-up @issakomi.

PR #3918.