fresh master pull failing to build [SOLVED]

[freman]

Hi ,

I just updated from Fed33 to Fed34. So I'm rebuilding my FreeCAD master.

I've rebuilt the ocl and opencascade libs which I build from source and deleted both the source and build trees for FreeCAD.

With fresh git pull from earlier today it is failing like this:

Code: Select all

[ 11%] Copying /~/FreeCAD/src/Mod/Start/TestStart/__init__.py to /~/freecad-build/freecad-build/Mod/Start/TestStart/__init__.py
[ 11%] Built target PartDesignScripts
make[2]: *** [src/3rdParty/salomesmesh/CMakeFiles/SMDS.dir/build.make:134: src/3rdParty/salomesmesh/CMakeFiles/SMDS.dir/src/SMDS/SMDS_FaceOfNodes.cpp.o] Error 1
make[1]: *** [CMakeFiles/Makefile2:2187: src/3rdParty/salomesmesh/CMakeFiles/SMDS.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....

From the cmake output:

Code: Select all

-- OCC:                 7.5.0 [TKFillet;TKMesh;TKernel;TKG2d;TKG3d;TKMath;TKIGES;TKSTL;TKShHealing;TKXSBase;TKBool;TKBO;TKBRep;TKTopAlgo;TKGeomAlgo;TKGeomBase;TKOffset;TKPrim;TKSTEPBase;TKSTEPAttr;TKSTEP209;TKSTEP;TKHLR;TKFeat] [/usr/local/lib] [/usr/local/include/opencascade]
-- SMESH:               build internal
--  MEDFile:            4.1.0 [/usr/lib64/libmedC.so;/usr/lib64/libmed.so] [/usr/include/med]

Can anyone see what is wrong?

It looks a bit like this problem but it's nearly a year old and was on 0.18. Is this still not fixed?
https://forum.freecadweb.org/viewtopic. ... 10#p408121

TIA.

[openBrain]

I guess googling the error message would give you some valuable clues.

[freman]

Well the actual error seems to be higher up, googling that is what led me to the FC issue I linked at the end.

Code: Select all

/~/FreeCAD/src/3rdParty/salomesmesh/inc/SMDS_UnstructuredGrid.hxx: In member function ‘vtkCellLinks* SMDS_UnstructuredGrid::GetLinks()’:
/~/FreeCAD/src/3rdParty/salomesmesh/inc/SMDS_UnstructuredGrid.hxx:98:12: error: invalid user-defined conversion from ‘vtkSmartPointer<vtkAbstractCellLinks>’ to ‘vtkCellLinks*’ [-
fpermissive]

do you have anything more pertinent?

[openBrain]

May that help ? https://github.com/waebbl/waebbl-gentoo/issues/268

[freman]

Thanks for the suggestion oB but that does not seem at all related to the errors I posted above.

There is a syntax error in building smesh and it seems related to vtk9 having made an API change as detailed in the FC forum link I posted.

[openBrain]

Thanks for the suggestion oB but that does not seem at all related to the errors I posted above.

There is a syntax error in building smesh and it seems related to vtk9 having made an API change as detailed in the FC forum link I posted.

As a coincidence, the link I gave links to this other one : https://github.com/waebbl/waebbl-gentoo/pull/289

[freman]

Thanks so it's definitely the vtk9 issue.

Fed34 only has vtk9 which is my dumb fault for updating to "current version" and getting to deal with this kind of crap.

For now I probably don't SMESH anyway, so -DBUILD_SMESH=OFF

[openBrain]

Thanks so it's definitely the vtk9 issue.

Fed34 only has vtk9 which is my dumb fault for updating to "current version" and getting to deal with this kind of crap.

For now I probably don't SMESH anyway, so -DBUILD_SMESH=OFF

Great. Please just mark the topic [Solved] if not building SMESH is enough.

[freman]

Why would I mark something as solved when quite clearly it is NOT solved ?

Fedora may be the first distro to only offer vtk9 and are thus providing advanced warning of the API change which is causing SMESH to fail to compile.

Avoiding building SMESH is a work around to allow the rest of FC to compile. It is not solved. It needs to be fixed.

Also making others aware of this is helpful. If I'd known before hand I would not have moved to Fed34. Marking it solved would be misleading.

[openBrain]

Why would I mark something as solved when quite clearly it is NOT solved ?

Fedora may be the first distro to only offer vtk9 and are thus providing advanced warning of the API change which is causing SMESH to fail to compile.

Avoiding building SMESH is a work around to allow the rest of FC to compile. It is not solved. It needs to be fixed.

Also making others aware of this is helpful. If I'd known before hand I would not have moved to Fed34. Marking it solved would be misleading.

Forum isn't a bugtracker. Marking [Solved] just means you found a "solution" (answer), even it being workaround. Use '[Solved-workaround]' if you prefer or whatever.