I see you reverted the change in the rules file after the failed build. Why?sgrogan wrote:We also need to set FREECAD_USE_OCC_VARIANT="Official Version" in the rules file
I already set it last night.sgrogan wrote:we may also need to set the OCC_INCLUDE_DIR
Completely forgot about switching oce build-deps to occt. Guess that's what lack of sleep will produce.sgrogan wrote:We also need to modify the control file, I'm looking at this now
Well the problem is here:sgrogan wrote:EDIT2: no joy
Remove the "l", I think it stands for "lib". If you look in the OCCT7 source in /adm/MODULES, those are part of Data Exchange.sgrogan wrote:/src/Driver/Driver_SMDS_Mesh.cpp"
/usr/bin/ld: cannot find -lTKIGES
/usr/bin/ld: cannot find -lTKSTL
/usr/bin/ld: cannot find -lTKXSBase
/usr/bin/ld: cannot find -lTKSTEP
/usr/bin/ld: cannot find -lTKSTEPBase
/usr/bin/ld: cannot find -lTKSTEPAttr
/usr/bin/ld: cannot find -lTKXCAF
/usr/bin/ld: cannot find -lTKXDESTEP
/usr/bin/ld: cannot find -lTKXDEIGES
Following gdo35's script, those toolkits are provided in the libopencascade-data-exchange-7.1.0 package and its -dev package. In liboce there is no such package, for some reason the Debian maintainers split them between modeling and ocaf.
Looking at the buildlog, I don't know why, the binary data-exchange package was installed, but not the dev package. That must be the reason for the build failure.
But what about netgen-headers? Will it conflict if we leave it there?