Fixing arithmetic ring issue

[jlevallois]

PR Description

The aim of this PR is to remove arithmetic ring issue when we change the integer type of SpaceND/KhalimskySpaceND.
By convention, we use SpaceND<dim, int32_t> and KhalimskySpaceND<dim, int32_t>.
But, in some case (if we want to reduce the memory footprint for example), we could be led to use int16_t instead of int32_t (or other).
The issue is that some algorithms are hard-coded to use int32_t. This PR try to resolve this issue.

(I'm sorry, the reviewing process will not be easy because my editor has deleted some unnecessary space in comments ...)

Critical files modified

These files need to be closely reviewed.

• AlphaThickSegmentComputer.h
• PowerMap.h / PowerMap.ih
• VoronoiMap.h / VoronoiMap.ih
• RigidTransformation3D.h
• Display2DFactory.ih
• DrawWithDisplay3DModifier.h
• BasicPointFunctors.h

Test files modified

• testAlphaThickSegmentComputer.cpp
• testFrechetShortcut.cpp
• testFreemanChain.cpp
• testDistanceTransformationND.cpp
• testImageSimple.cpp
• testImageSpanIterators.cpp
• testSliceImageFromFunctor.cpp
• testDigitalSet.cpp
• testHyperRectDomain-snippet.cpp
• testHyperRectDomain.cpp
• testCellularGridSpaceND.cpp
• testSurfaceHelper.cpp

Checklist

• [N/A] Unit-test of your feature with Catch.
• [N/A] Doxygen documentation of the code completed (classes, methods, types, members...)
• [N/A] Documentation module page added or updated.
• New entry in the ChangeLog.md added.
• No warning raised in Debug cmake mode (otherwise, Travis C.I. will fail).
• All continuous integration tests pass (Travis & appveyor)

[dcoeurjo]

why did you change this one ?
Even if int32_t is already the default type for SpacND, maybe the author as a good reason to want a int32_t type (if for instance we change the default type in SpaceND

[jlevallois]

Agree, in this case I need to change testAlphaThickSgementComputer who uses Z2i::Point/Z2i::Domain

[dcoeurjo]

No I think it should be SpaceND<2, InputPoint::Coordinate> here

[jlevallois]

We can't, because of AlphaThickSegmentComputer< Z2i::RealPoint> in testAlphaThickSgementComputer (line 199)

[dcoeurjo]

Ok I'll have a look.

[jlevallois]

It works when hardcoding ParallelStrip<SpaceND< 2, int32_t>, ...> because his operator() has a surcharge for Point and RealPoint (always double) from SpaceND.

[dcoeurjo]

Ok... And there is no implicit cast double -> int when used ?

[jlevallois]

double doesn't follow the concept CInteger, require by SpaceND

[dcoeurjo]

Yeah of course. I meant in the code, are input points casted to "primitive" points at some places (which would mean an implicit cast if input points are RealPoints)

[dcoeurjo]

nice PR BTW

[dcoeurjo]

thanks again @jlevallois
Can you please revert your edit in Alpha... (reset the int32_t template type) and add an entry to the changeling ?
Beside, the PR is fine, thanks again

[jlevallois]

It's done :)

[dcoeurjo]

Thanks Merging