take another pass at the LIC mapper

Try to clean up a few things I did not have time to address in the first conversion. It now doesn't do the wierd thing with setting its Output so that it can map vectors to tcoords. This is now properly passed as an attribute.

There were a couple bugs realted to parallel composite LIC one being a vertex coordinate system issues the other being a MPI deadlock with icet

Basic content checks passed!

Branch-at: fe8b62ac
Acked-by: @kwrobot

Do: test

Testing commands handed to buildbot.

Branch-at: fe8b62ac

Your merge request has been queued for testing. You may view the test results on CDash or Buildbot.

Branch-at: fe8b62ac

@berkgeveci bunch of LIC fixes for parallel composite LIC support plus general cleanup reorg of LIC to be easier to track.

+2

Do: merge

Branch successfully merged and pushed!

mentioned in commit 5f15eaa6

Status changed to merged

This is breaking paraview LIC plugin, especially the cleanup of vtkSurfaceLICMapper.cxx

vtkSurfaceLICRepresentation.cxx: In member function 'void vtkSurfaceLICRepresentation::SetUseLICForLOD(bool)': vtkSurfaceLICRepresentation.cxx:102:30: error: 'class vtkCompositeSurfaceLICMapper' has no member named 'SetEnable' ... https://open.cdash.org/viewBuildError.php?buildid=4347864

@utkarsh.ayachit @bloring

See paraview/paraview!752 (closed)

thanks, i forgot to check open MR.

I hesitate to update PV's VTK, so I made the PV topic but was hoping someone else would commit it when they decided to update PVs VTK version.

allright, i'm trying to get green dashboards in order to update vtk this week. I supose that the right way to go would be to update PV'sVTK as soon as something is merged into VTK master, at least for ParaView/VTK dev. It would avoid the need of going trough other peoples commit to figure out what's wrong in the last 40 commits when one is trying to update PV's VTK.

@ken-martin you should really not hesitate to update ParaVIew's VTK. If you only had added a commit to your MR to update it's VTK SHA we could already have it tested and merged and not have to wait for someone else to catch it.

@mwestphal We have discussed in the past for having such a test PV using VTK master dashboard. The problem becomes no one monitors it. It's easier to just have the one attempting to update the VTK branch take the onus of tracking down who's to blame for the regression and then work with them to get it fixed.

@utkarsh.ayachit allright, happy to see that you're back ! :D

@mwestphal Thanks! Always good to be back :).

Yes,see my PV topic that is open,you will need those changes if you update the VTK version. -Ken