Postprocessing

[PrzemoF]

I'm observing some transparency problems. Pipeline transparency works, but even 100% doesn't make it fully transparent. Clip transparency 0% works fine, but settig even 1% makes the object almost completely transparent.

[teobo]

What do you guys think, is this something worth pursuing?

Incredible. You did it. Congrats.
to your question: Without doubt. When does it come to master? This is one of the things salome had in favor uptonow: its cloned paraview (paravis) module. And yes FC could use access to that functionality, too, however is smarter even with Python, is not it?

And even though I do not overlook the whole integration: Now it seems like that freecad connects to the state of the art in postprocessing world.
I had not believed it would proceed that fast when I read it the first time.

Thank you

[ickby]

When does it come to master?

ah well, I thought I make it until this week, but I struggle with finishing the property editor expression support, which I mistakenly thought was only a few hours work. So no estimate on this, but it is the next thing afterwards

[ickby]

Hello,

Over the last days I worked further on "freecadifying" the post processing code. Functionality wise nothing changed. The implementation now uses a property to store data, which also makes save and restore work and which fits way better into the undo/redo handling. Furthermore I switched from the vtk pipeline to use the freecad recompute system, it was just too messy having both next to each other. While on it I did fix quite some bugs too, however, the transparency issue is still unsolved, I'm a bit puzzled with that one...

So I thought how to proceed, and I do think that I will keep the current state (bug fixes excluded) for now. There is plenty room for improvement, however, many of the functionality I want for the grouping and handling will be much much easier to implement once the new part design is in master. If I would go further now I would only do double work, removing lots of code again later on. So we can think about moving this to master now and fixing stuff on the go, or waiting for the start of 0.17, fixing the grouping stuff and merge than. But well, for this no timing exist One downside when merging right now is that we may have some trouble loading old files with post processing objects once the grouping changes are done. I'm not 100% sure this change can be done backwards compatible.

What do you guys think?

[triplus]

Hard to say anything without scheduled release cycle. Part Design Next branch was the reason i hoped for FreeCAD 0.16 release in 2015 but it didn't happen. Looking at the list of releases done in past i guess FreeCAD 0.16 release could be scheduled for Jan, Feb or Apr.

Therefore if you finished the work merge it into master. I don't see why you would wait again with this one.

[teobo]

Hi ickby,
thanks for the work.
maybe I am too lazy for learning git branch testing (did it once with one of bernds branches..), but I welcome (can not wait ) to have it in master as sone as possible for learning to know it the first time and to have it (felt) more reliable for cases. Other testers might think the like and could give early corrections, too.
Vote: +1 master, asap

[jmaustpc]

One downside when merging right now is that we may have some trouble loading old files with post processing objects once the grouping changes are done. I'm not 100% sure this change can be done backwards compatible.

If there is going to be pain at some point then I suggest it should happen as soon as possible, as there will only be more work done in the old system as time goes on.

[bernd]

... One downside when merging right now is that we may have some trouble loading old files with post processing objects once the grouping changes are done. I'm not 100% sure this change can be done backwards compatible.

What do you guys think?

FEM has been under heavy development in the last months. There has not been a stable API for FEM yet. Therefore Przemo and me we never made effort in reading obsolete files. --> Go for merging ASAP without reading old files.

Tried to test but on Debian Jessie and Debian Stretch cmake fails with

Code: Select all

hugo@weide:~/Documents/dev/freecad/freecadbhb_ickby/build$ cmake ../
-- prefix: /usr/local
-- datadir: data
-- docdir: doc
-- includedir: include
-- libdir: /usr/local/lib
-- Boost version: 1.55.0
-- Found the following Boost libraries:
--   filesystem
--   program_options
--   regex
--   signals
--   system
--   thread
-- Found Xerces-C: /usr/lib/x86_64-linux-gnu/libxerces-c.so
-- PyCXX found:
--   Headers:  /home/hugo/Documents/dev/freecad/freecadbhb_ickby/src
--   Sources:  /home/hugo/Documents/dev/freecad/freecadbhb_ickby/src/CXX
-- -- OpenCASCADE Community Edition has been found.
-- -- Found OCE/OpenCASCADE version: 6.7.0
-- -- OCE/OpenCASCADE include directory: /usr/lib/x86_64-linux-gnu/oce-0.15/../../../include/oce
-- -- OCE/OpenCASCADE shared libraries directory: 
-- The imported target "vtkCommonCoreTCL" references the file
   "/usr/lib/x86_64-linux-gnu/libvtkCommonCoreTCL-6.1.so.6.1.0"
but this file does not exist.  Possible reasons include:
* The file was deleted, renamed, or moved to another location.
* An install or uninstall procedure did not complete successfully.
* The installation package was faulty and contained
   "/usr/lib/cmake/vtk-6.1/VTKTargets.cmake"
but not all the files it references.

-- The imported target "vtkCommonMathTCL" references the file
   "/usr/lib/x86_64-linux-gnu/libvtkCommonMathTCL-6.1.so.6.1.0"
but this file does not exist.  Possible reasons include:
* The file was deleted, renamed, or moved to another location.
* An install or uninstall procedure did not complete successfully.
* The installation package was faulty and contained
   "/usr/lib/cmake/vtk-6.1/VTKTargets.cmake"
but not all the files it references.

-- The imported target "vtkCommonMiscTCL" references the file
   "/usr/lib/x86_64-linux-gnu/libvtkCommonMiscTCL-6.1.so.6.1.0"
but this file does not exist.  Possible reasons include:
* The file was deleted, renamed, or moved to another location.
* An install or uninstall procedure did not complete successfully.
* The installation package was faulty and contained
   "/usr/lib/cmake/vtk-6.1/VTKTargets.cmake"
but not all the files it references.

-- The imported target "vtkCommonSystemTCL" references the file
   "/usr/lib/x86_64-linux-gnu/libvtkCommonSystemTCL-6.1.so.6.1.0"
but this file does not exist.  Possible reasons include:
* The file was deleted, renamed, or moved to another location.
* An install or uninstall procedure did not complete successfully.
* The installation package was faulty and contained
   "/usr/lib/cmake/vtk-6.1/VTKTargets.cmake"
but not all the files it references.

... and many more of such errors ...

-- libshiboken built for Release
-- Found PySide Tools: /usr/bin/pyside-uic, /usr/bin/pyside-rcc
-- -- matplotlib-1.4.2 has been found.
-- Platform is 64-bit, set -D_OCC64
-- Build type: 
git
/home/hugo/Documents/dev/freecad/freecadbhb_ickby/build/src/Build/Version.h written
-- setting gcc options: -Wall -Werror -Wno-deprecated -pedantic-errors
-- Boost version: 1.55.0
-- Found the following Boost libraries:
--   python
-- found Boost: 1_55
-- boost-incude dirs are: /usr/include
-- boost-python lib is: /usr/lib/x86_64-linux-gnu/libboost_python.so
-- boost_LIBRARY_DIRS is: /usr/lib/x86_64-linux-gnu
-- Boost_LIBRARIES is: /usr/lib/x86_64-linux-gnu/libboost_python.so
-- area module (for Path Workbench) will be installed to: /usr/local/lib
-- Coin3D doc is installed
--   Found /usr/share/doc/libcoin60-doc/html/coin.tag.gz
--   You should uncompress this file if you want to use it for source doc generation
=======================================
Now run 'make' to build FreeCAD
=======================================

-- Configuring done
CMake Error at src/Mod/Fem/App/CMakeLists.txt:217 (add_library):
  Cannot find source file:

    PropertyPostDataObject.h

  Tried extensions .c .C .c++ .cc .cpp .cxx .m .M .mm .h .hh .h++ .hm .hpp
  .hxx .in .txx


-- Build files have been written to: /home/hugo/Documents/dev/freecad/freecadbhb_ickby/build
hugo@weide:~/Documents/dev/freecad/freecadbhb_ickby/build$ 

but vtk is installed properly

Code: Select all

hugo@weide:~/Desktop$ dpkg -l | grep vtk
ii  libvtk6-dev                            6.1.0+dfsg2-6                        amd64        VTK header files for building C++ code
ii  libvtk6.1                              6.1.0+dfsg2-6                        amd64        Visualization Toolkit - A high level 3D visualization library - runtime
ii  python-vtk6                            6.1.0+dfsg2-6                        amd64        Python bindings for VTK
ii  tcl-vtk6                               6.1.0+dfsg2-6                        amd64        Tcl bindings for VTK
ii  vtk6                                   6.1.0+dfsg2-6                        amd64        Binaries for VTK6
ii  vtk6-doc                               6.1.0+dfsg2-6                        all          VTK class reference documentation
ii  vtk6-examples                          6.1.0+dfsg2-6                        all          C++, Tcl and Python example programs/scripts for VTK
hugo@weide:~/Desktop$ 

I purged and reinstalled and reconfigured but no success

@stefean (ickby): I really would like to test this stuff. Which version of ubuntu are you on?

[ickby]

That's a shame, but IMHO this looks like it is a problem with the vtk packages, as in Freecad we only use the cmake vtk find routines provided by the vtk package. So the error occurs before anything freecad custom happens. Strange nonetheless.

I'm on kubuntu 15.04. I will try google and see if this problem is known later today.

[bernd]

That's a shame, but IMHO this looks like it is a problem with the vtk packages, as in Freecad we only use the cmake vtk find routines provided by the vtk package. So the error occurs before anything freecad custom happens. Strange nonetheless.

I'm on kubuntu 15.04. I will try google and see if this problem is known later today.

Yeah, it is what I assume too and I have no glue how to get around this ... The only thing I found is https://bugs.debian.org/cgi-bin/bugrepo ... 8372;msg=2