Kernel/LHCbMath: add Bernstein3D and Positive3D

Kernel/LHCbMath:

• add generic 3D-polynomial Bernstein3D
• add positive 3D-polynomial Positive3D

• [2017-11-15 00:02] Validation started with lhcb-2017-patches#204
• [2017-11-16 00:03] Validation started with lhcb-2017-patches#205

Hi Marco, @cattanem

1. all member variable are prefixed with m_
2. it is not initialization. It is a single declaration of several constructor with different combination of default arguments. If I initialize members as you suggest, I will need to duplicate (actually multiplicate) constructors, that is not nice at all

Hello @ibelyaev

1. Yes, I fixed my comment when I saw this...
2. Agreed for the constructor, but that does not prevent to initialise the member variables anyway, on lines 2178 to 2106?

resolved all discussions

resolved all discussions

changed title from Kernel/LHCbMath: to Kernel/LHCbMath: add Bernstein3D and Positive3D

changed the description

merged

mentioned in commit 67ab47fe

mentioned in merge request !955 (closed)

@ibelyaev When trying to merge this to master, I get several conflicts (same file modified in both master and 2017-patches), like this:

++<<<<<<< HEAD
 +  double result = 0 ;
 +  for  ( unsigned short ix = 0 ; ix <= m_nx ; ++ix )
 +  {
 +    for  ( unsigned short iy = 0 ; iy <= m_ny ; ++iy )
 +    { result += par ( ix , iy ) * fx[ix] * fy[iy] ; }
 +    //
 +  }
 +  //
 +  const double scalex = ( m_nx + 1 ) / ( xmax() - xmin() ) ;
 +  const double scaley = ( m_ny + 1 ) / ( ymax() - ymin() ) ;
 +  //
 +  return result * ( scalex * scaley ) ;
++=======
+   return calculate (  fx , fy ) ;
++>>>>>>> 2017-patches

Which should I keep?

keep the latest (2017-patches)

@cattanem can we discuss the issues in person? I am a bit confused. If we propagate full file from 2017-patched to master, we'll lose a lot of very useful functionality :-(

mentioned in merge request !956 (merged)

@ibelyaev We don't propagate the full file, only the diffs. We get a merge conflict if the file has changed in both branches since the branches were split. See !956 (merged)

mentioned in commit 71252585

mentioned in commit c2cf6992