Arch Linux User Repository

thanks for the welcome! :) importing freecad itself was working but some of the modules were not available from python out of the box, drmacro was having trouble to import PartDesign specifically, there was a recent change regarding the whole site-packages situation so that might have broken stuff. With my changes the modules are available out of the box, no hacks required, I'll push the changes later today once I'm home since I need to setup my ssh key here.

Hey adrianinsaval, happy to add you as a co-maintiner! I'd love to see you get rid of some of my hacks here :-) My goals have been to of course make sure all the features of this freecad work correctly, but also to ensure that its python modules can be imported from anywhere without any python search path manipulation. Seems like something in freecad must have (hopefully) changed recently because everything used to be working correctly with the hacks I had. Would be great if they're not needed any more.

@greyltc all the hacks in the PKGBUILD to get it to work on python were problematic and it took me some time to figure out how to get it to work, here https://github.com/adrianinsaval/freecad-git-AUR a PKGBUILD that works without adding potentially problematic env variables or links, explicitly setting the full path for CMAKE_INSTALL_LIBDIR was the main needed change because that is what is used to set the default value of PATH_TO_FREECAD_LIBDIR at build time in the freecad init.py script. I changed it to be /usr/lib/freecad/lib to be FHS compliant. Please merge into this repo or if you're ok with it you can add me as another maintainer for this package, I can't promise I'll be making this work all the time but I can give it a go whenever needed and I'm very active in the freecad forum

Further experimentation shows that the AUR script leave symlinks in /usr/lib/python3.10/site-packages for most of the FreeCAD modules.

The symlinks point to /usr/lib where most of the FreeCAD modules have been duplicated. (They also exist in /usr/lib/freecad.)

But, PartDesign is not similarly symlinked. In addition it is not duplicated in /usr/lib nor is it a .so module, it uses the Python ini.py resident in /usr/lib/freecad/Mod/PartDesign directory.

It would apear the symlink setup done by AUR is lacking a symlink from to /usr/lib/python3.10/site-packages to /usr/lib/freecad/Mod/PartDesign

Should the freecad modules be duplicated in multiple places? Is the symlink setup the right way for this to be handled? At minimum the symlink for PartDesign directory needs to be added?

More research and experimentation.

Basically the AUR install leave Python with no way to find PartDesign/init.py because it is just not resident in any of the directories where Python searches.

PartDesign/__init__.py does exist in /usr/lib/freecad/Mod

So, the following script allows Blender/sverchok to work with FreeCAD PartDesign objects:

PYTHONPATH=/usr/lib/freecad/Mod
export PYTHONPATH
blender

It appears that the AUR script ends up with parts of freecad in different palces. /usr/lib/python3.10/site-packages/ /usr/share/freecad /usr/lib/freecad

Apps trying to use the freecad api fail to find PartDesign since the system is looking in /usr/lib/python3.10/site-packages/ but PartDesign is not installed there.

Normally the PartDesign directory contains init.py. There is no PartDesign directory in /usr/lib/python3.10/site-packages/ But, _PartDesign and init.py have been deposited there. Attempting to 'import PartDesign' fails the Python mechanism for loading a directory module.

Not sure, but it is possible that the resulting file structure is not quite right. When using the FreeCAD API it appears to not find PartDesign.

See this thread where I attempt to trouble shoot with wmayer: https://forum.freecadweb.org/viewtopic.php?f=22&t=68354&start=10

Edit: tried this on a different PC with Manjaro. It does appear that the file locations are not set up correctly for API use outside FreeCAD.

As you pointed out yourself Kunda qt5-webengine was already added, qt5-webchannel I think should be explicitly added, although currently it will get installed anyways because qt5-webengine depends on it.

Looks like there may be new Qt dep for Addon manager (reported by a user who uses the PPA) https://forum.freecadweb.org/viewtopic.php?f=42&t=68015 This is what is shown on their system:

Addon Manager Warning: Could not import QtWebEngineWidgets, it seems to be missing from your system. 
Please use your system's package manager to install the python3-pyside2.qtwebengine* and python3-pyside2.qtwebchannel packages, 
and if possible alert your package creator to the missing dependency. 
Display of package README will be limited until this dependency is resolved.

I can't test this because freecad-git fails when i built it. But just wanted to give @greyltc a heads up.

Just tried to build latest freecad-git (r28679)
Build completed but failed in the tests (80 tests failed)
Here is the log file: https://forum.freecadweb.org/download/file.php?id=186393

qt5-webengine was added https://aur.archlinux.org/cgit/aur.git/commit/?h=freecad-git&id=2498280e13ca6362fc5c31db486cdc9a85a5d636

@greyltc also please add qt5-webengine as dependency (https://forum.freecadweb.org/viewtopic.php?f=42&t=65971)

@greyltc the INSTALL_TO_SITEPACKAGES thing was fixed upstream (https://forum.freecadweb.org/viewtopic.php?p=586095) and the code you put to deal with that is now causing an error because /usr/lib/freecad/lib doesn't exists

@adrianinsaval English UI with different decimal separator works again: https://github.com/FreeCAD/FreeCAD/commit/bba7719cd15e755a5a431ea339ef535377a0f41b

FYI the abandoned assembly module was recently removed from master branch. Plot was mostly removed and what is left I think is on by default, not sure. Just mentioning if you want to clean those up. FEM I think defaults to on too.

yeah that's a new feature but it has some bugs, go into preferences and set your language to German and check the box to use the same decimal separator as the language you use if you want to use comma separator. Currently there's no way to use English language and comma separator at the same time. Or forget that and check the configuration in FreeCAD to substitute your numpad comma by the decimal separator.

Anyone here having issues with the decimal separator?

Although I have set language to English, I still use local (Germany) number formats (which uses the "," instead of "." as decimal separator).

Until the newest build (from today) FreeCAD always used the system settings and therefore used the comma instead of the decimal point. But now, the locale seems to be ignored and only the decimal point is accepted/used.

https://forum.freecadweb.org/viewtopic.php?p=581878#p581796 https://forum.freecadweb.org/viewtopic.php?p=581878#p581878

apparently enabling JTREADER is pointless and has recently led to build failures, apparently it was fixed but the module seems mostly unmaintained so IMO it would be best to stop enabling it.

@greyltc did you see my patch proposals for pkgver() and to add graphviz python-markdown and python-gitpython as optional dependencies? They are in the comments here, please take a look, the opt deps are useful and the pkver would give a version number in line with FreeCAD's version numbering system.

Seems the latest is broken...

[3449/4279] Linking CXX executable bin/FreeCAD
FAILED: bin/FreeCAD 
: && /usr/bin/c++ -Wall -Wextra -Wno-write-strings -march=x86-64 -mtune=generic -O2 -pipe -fno-plt -fexceptions         -Wp,-D_FORTIFY_SOURCE=2 -Wformat -Werror=format-security         -fstack-clash-protection -fcf-protection -Wp,-D_GLIBCXX_ASSERTIONS -fPIC -w -Wl,-O1,--sort-common,--as-needed,-z,relro,-z,now src/Main/CMakeFiles/FreeCADMain.dir/FreeCADMain_autogen/mocs_compilation.cpp.o src/Main/CMakeFiles/FreeCADMain.dir/MainGui.cpp.o -o bin/FreeCAD -L/usr/lib/openmpi -Wl,-rpath,/usr/lib/openmpi:/var/tmp/pamac-build-mac/freecad-git/src/FreeCAD/build_dir/lib:  -lmpi_cxx  -lmpi  lib/libFreeCADGui.so  lib/libFreeCADApp.so  lib/libFreeCADBase.so  -lmpi_cxx  -lmpi  /usr/lib/libxerces-c.so  /usr/lib/libz.so  /usr/lib/libpython3.10.so  -lutil  -ldl  /usr/lib/libQt5Xml.so.5.15.3  -lCoin  /usr/lib/libboost_filesystem.so.1.78.0  /usr/lib/libboost_atomic.so.1.78.0  /usr/lib/libboost_program_options.so.1.78.0  /usr/lib/libboost_regex.so.1.78.0  /usr/lib/libboost_system.so.1.78.0  /usr/lib/libboost_thread.so.1.78.0  /usr/lib/libboost_date_time.so.1.78.0  -lpthread  /usr/lib/libQt5OpenGL.so.5.15.3  /usr/lib/libQt5PrintSupport.so.5.15.3  /usr/lib/libQt5Svg.so.5.15.3  /usr/lib/libQt5UiTools.a  /usr/lib/libQt5Widgets.so.5.15.3  /usr/lib/libQt5Gui.so.5.15.3  /usr/lib//libQt5Widgets.so  /usr/lib//libQt5Gui.so  /usr/lib//libQt5Core.so  /usr/lib//libQt5Widgets.so  /usr/lib//libQt5Gui.so  /usr/lib//libQt5Core.so  /usr/lib/libGL.so  /usr/lib/libpyside2.cpython-310-x86_64-linux-gnu.so.5.15.3  /usr/lib/libshiboken2.cpython-310-x86_64-linux-gnu.so.5.15.2.1  /usr/lib/libQt5Network.so.5.15.3  /usr/lib/libQt5Core.so.5.15.3 && :
/usr/bin/ld: /usr/lib/libpyside2.cpython-310-x86_64-linux-gnu.so.5.15.3: undefined reference to `setDestroyQApplication'
collect2: error: ld returned 1 exit status
[3450/4279] Building CXX object src/Mod/Inspection/Gui/CMakeFiles/InspectionGui.dir/AppInspectionGui.cpp.o
[3451/4279] Automatic MOC for target MeshGui
[3452/4279] Building CXX object src/Mod/Image/Gui/CMakeFiles/ImageGui.dir/AppImageGui.cpp.o
[3453/4279] Building CXX object src/Mod/Inspection/Gui/CMakeFiles/InspectionGui.dir/VisualInspection.cpp.o
[3454/4279] Building CXX object src/Mod/Inspection/Gui/CMakeFiles/InspectionGui.dir/Command.cpp.o
[3455/4279] Automatic MOC for target RobotGui
[3456/4279] Building CXX object src/Mod/Drawing/Gui/CMakeFiles/DrawingGui.dir/AppDrawingGui.cpp.o
[3457/4279] Building CXX object src/Mod/Drawing/Gui/CMakeFiles/DrawingGui.dir/AppDrawingGuiPy.cpp.o
[3458/4279] Automatic MOC for target PartGui
ninja: build stopped: subcommand failed.
==> ERROR: A failure occurred in build().
    Aborting...

Hi all.

If I 'makepkg -s' the build fails on tests

This is the fist fail

====================================================================== ERROR: TestSketcherApp (unittest.loader._FailedTest)

ImportError: Failed to import test module: TestSketcherApp Traceback (most recent call last): File "/usr/lib/python3.10/unittest/loader.py", line 154, in loadTestsFromName module = import(module_name) File "/home/mauro/Scaricati/Artix/freecad-git/src/FreeCAD/build_dir/Mod/Sketcher/TestSketcherApp.py", line 24, in <module> from SketcherTests.TestSketcherSolver import TestSketcherSolver File "/home/mauro/Scaricati/Artix/freecad-git/src/FreeCAD/build_dir/Mod/Sketcher/SketcherTests/TestSketcherSolver.py", line 23, in <module> import FreeCAD, os, sys, unittest, Part, Sketcher ImportError: libtbb.so.2: cannot open shared object file: No such file or directory

In my system I have only libtbb.so.12.5 installed from package tbb.

I look the histoty of tbb in Arch. From here https://archlinux.org/todo/tbb-20214-rebuild/ libtbb.so.2 is renamed to libtbb.so.12

Analisys

I made some analisys on the so files produced bny Freecad.

These files need libtbb.so.2 and use libtbb.so.12

• libDriverSTL.so
• libSMESH.so:
• libSMESHDS.so
• libStdMeshers.so

SOLVED

The problem was in opencascade libraries

@greyltc I've send you an email the 11 but forgot to add the attachment, send it today

no?

@greyltc did you receive @FabioLolix's patch ?

@Kunda no idea, I have successfully builded using makepkg from my home directory and using extra-x86_64-build from devtools with used dir set to /tmp/chrootbuild, tmpfs mounted without nosuid

@FabioLolix nice! any idea why we're seeing the test failures in https://aur.archlinux.org/packages/freecad-git#comment-851196 ?

Hello, need checkdepends=(fmt pugixml) for building in clean chroot. Just send you you a patch with this and other minor changes

Opened a discussion thread on the AUR subforum regarding the test failures: https://bbs.archlinux.org/viewtopic.php?pid=2021620#p2021620

Has anyone figured out why the tests fail? It looks like a permissions issue.
@adrianinsaval thought it may have been an issue with fakeroot? But we're not so sure. Here are the upstream discussions:
https://forum.freecadweb.org/viewtopic.php?f=4&t=66140
https://forum.freecadweb.org/viewtopic.php?f=4&t=66141

Separate issue but relevant: https://forum.freecadweb.org/viewtopic.php?f=4&t=66137

@greyltc upstream py3.10 fix https://github.com/FreeCAD/FreeCAD/pull/5503

Using "pip install PyQtWebEngine" and I still get the same pugixml not found.

So, is "HEAD needs new dep: PySide2.QtWebEngineWidgets" covered by the above or am I just not understanding what this is supposed to fix?

@greyltc heads up, latest HEAD needs new dep: PySide2.QtWebEngineWidgets

ref: https://github.com/ppd/freecad-ppd/issues/17

@greyltc pivy/python3.10 upstream issue: https://github.com/coin3d/pivy/issues/88

@C0rn3j at least one test theoretically should be fixed: https://github.com/FreeCAD/FreeCAD/commit/3dfdc6804f3931b27ebb526d3d87b200874d2a7c

could you and @Wolfenmond and anyone else retest the build (with check()) enabled and report back if there are any other tests that are failing?

Using the workaround mentioned by K4LCIFER this package does compile successfully. But, because of the Python 3.10 and pivy issues features of FreeCAD do not work.

Got the same issue as @Wolfenmond and couldn't find it reported upstream, so I did - https://tracker.freecadweb.org/view.php?id=4825

Since there still seems to be a lot of issues with check(), I would like to reiterate a workaround for anyone encountering issues, which is to install freecad with the --nocheck flag in makepkg to skip check(). You can also run yay with the --nocheck flag by passing a makepkg flag to yay with yay's --mflags option, for example:

yay --mflags --nocheck -S freecad-git

on latest HEAD I'm getting

CMake Warning at /usr/lib64/cmake/vtk/VTK-vtk-module-find-packages.cmake:2718 (find_package):
  By not providing "Findpugixml.cmake" in CMAKE_MODULE_PATH this project has
  asked CMake to find a package configuration file provided by "pugixml", but
  CMake did not find one.

  Could not find a package configuration file provided by "pugixml" with any
  of the following names:

    pugixmlConfig.cmake
    pugixml-config.cmake

  Add the installation prefix of "pugixml" to CMAKE_PREFIX_PATH or set
  "pugixml_DIR" to a directory containing one of the above files.  If
  "pugixml" provides a separate development package or SDK, be sure it has
  been installed.
Call Stack (most recent call first):
  /usr/lib64/cmake/vtk/vtk-config.cmake:150 (include)
  cMake/FreeCAD_Helpers/SetupSalomeSMESH.cmake:53 (find_package)
  CMakeLists.txt:58 (SetupSalomeSMESH)

I'm getting "failures=4, errors=1" while building this on a freshly installed system:

FAIL: test_ccxcantilever_prescribeddisplacement_0_mm femtest.app.test_solver_elmer.TestSolverElmer)
FAIL: test_ccxcantilever_nodeload_0_mm (femtest.app.test_solver_elmer.TestSolverElmer)
FAIL: test_ccxcantilever_faceload_1_si (femtest.app.test_solver_elmer.TestSolverElmer)
FAIL: test_ccxcantilever_faceload_0_mm (femtest.app.test_solver_elmer.TestSolverElmer)
ERROR: test_box_static_0_mm (femtest.app.test_solver_elmer.TestSolverElmer)

Am I missing something here?

EDIT: Markdown

I don't know if this is an upstream issue, or if anyone actually cares. If not cared about, feel free to ignore.

Building this package uses all CPU cores on my system, despite a config in my makepkg.conf telling it to only use 4. This config works with other packages (Tested with kicad-git). It looks like it's related to Freecad using cmake, and that setting not making it through.

Does anyone else have this issue or a workaround?

I get a bunch of missing library dependencies when running check(), e.g.:

======================================================================
ERROR: TestFemApp (unittest.loader._FailedTest)
----------------------------------------------------------------------
ImportError: Failed to import test module: TestFemApp
Traceback (most recent call last):
  File "/usr/lib/python3.10/unittest/loader.py", line 154, in loadTestsFromName
    module = __import__(module_name)
  File "/usr/lib/python3.10/site-packages/shiboken2/files.dir/shibokensupport/__feature__.py", line 142, in _import
    return original_import(name, *args, **kwargs)
  File "/home/joe/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/TestFemApp.py", line 33, in <module>
    from femtest.app.test_mesh import TestMeshCommon as FemTest07
  File "/usr/lib/python3.10/site-packages/shiboken2/files.dir/shibokensupport/__feature__.py", line 142, in _import
    return original_import(name, *args, **kwargs)
  File "/home/joe/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/femtest/app/test_mesh.py", line 33, in <module>
    import Fem
  File "/usr/lib/python3.10/site-packages/shiboken2/files.dir/shibokensupport/__feature__.py", line 142, in _import
    return original_import(name, *args, **kwargs)
ImportError: libfmt.so.8: cannot open shared object file: No such file or directory

This one can be fixed by installing fmt package. Similarly, libpugixml.so.1 can be fixed by installing pugixml. Finally, libhdf5.so.103 can't be easily fixed because the latest hdf5 package installs libhdf5.so.200.

Looks like some dependencies need to be updated?

proposal 2, add some optional dependencies:

diff --git a/PKGBUILD b/PKGBUILD
index cdd87ba..a08b3ff 100644
--- a/PKGBUILD
+++ b/PKGBUILD
@@ -51,6 +51,9 @@ optdepends=(
 'luxcorerender: ray tracing support'
 'libspnav: 3d mouse support'
 'openscad: OpenSCAD support'
+'graphviz: dependency graph support'
+'python-markdown: markdown support in addon manager'
+'python-gitpython: support downloading addons with git'
 )
 provides=('freecad')
 conflicts=('freecad' 'freecad-appimage' 'freecad-appimage-git')

proposal, get version number more similar to what is reported in FreeCAD's about dialog by comparing to 0.19 tag commit, any other arbitrary commit from master could be used too

diff --git a/PKGBUILD b/PKGBUILD
index cdd87ba..356742c 100644
--- a/PKGBUILD
+++ b/PKGBUILD
@@ -60,9 +60,9 @@ md5sums=('SKIP')
 pkgver() {
   cd FreeCAD
   read -d$'/n' -r major minor patch < <(grep -Po "set\(PACKAGE_VERSION_(MAJOR|MINOR|PATCH) \"\K[0-9]*" CMakeLists.txt) || true
-  count=$(git rev-list --count $(git tag --sort=-creatordate|head -1)..HEAD)
+  count=$((24266 + $(git rev-list --count d29fd7d..HEAD) ))
   hash=$(git rev-parse --short HEAD)
-  printf "%d.%d.%d.r%d.g%s" "$major" "$minor" "$patch" "$count" "$hash"
+  printf "%d.%d.%d.%d.g%s" "$major" "$minor" "$patch" "$count" "$hash"
 }

 prepare() {

would the arch equivalent be python-markdown? Maybe it's worth adding that and python-gitpython as optional dependencies. QtNetwork I think is already included in qt5-base and pyside2 so maybe no change is needed?

@greyltc I'll try to re-compile (my poor poor laptop)

BTW, upstream is also recommending python3-markdown dependency for the updated addon manager as well (source)

Edit: recompiling I see this the CMake output

CMake Warning at /usr/lib64/cmake/vtk/VTK-vtk-module-find-packages.cmake:2718 (find_package):
  By not providing "Findpugixml.cmake" in CMAKE_MODULE_PATH this project has
  asked CMake to find a package configuration file provided by "pugixml", but
  CMake did not find one.

  Could not find a package configuration file provided by "pugixml" with any
  of the following names:

    pugixmlConfig.cmake
    pugixml-config.cmake

  Add the installation prefix of "pugixml" to CMAKE_PREFIX_PATH or set
  "pugixml_DIR" to a directory containing one of the above files.  If
  "pugixml" provides a separate development package or SDK, be sure it has
  been installed.
Call Stack (most recent call first):
  /usr/lib64/cmake/vtk/vtk-config.cmake:150 (include)
  cMake/FreeCAD_Helpers/SetupSalomeSMESH.cmake:53 (find_package)
  CMakeLists.txt:58 (SetupSalomeSMESH)

@Kunda hmm. I had that check fail too when K4LCIFER reported it, but my commit here:
https://aur.archlinux.org/cgit/aur.git/commit/?h=freecad-git&id=880ae77dc6b0497a0a36dbffdc6cb67e12970918
fixed it. I'm surprised it's still an issue for you.

heads up @greyltc QtNetwork is now a required dependency https://forum.freecadweb.org/viewtopic.php?p=554851#p554851

The failed FEM test that @K4LCIFER reported got me as well.

Opened a thread on the FEM subforum: https://forum.freecadweb.org/viewtopic.php?f=18&t=64525

Does this version includes RT TNP fixes?

actually, I totally had a mistake in check()
Hopefully it's fixed for us now @K4LCIFER!

I think check() seems to be working alright here. Probably if there's a failure in check() you should report that upstream instead of here.

Will not install due to errors in check().

FAILED (errors=1)
==> ERROR: A failure occurred in check().
    Aborting...
 -> error making: freecad-git

I believe this is the error message:

======================================================================
ERROR: TestFemApp (unittest.loader._FailedTest)
----------------------------------------------------------------------
ImportError: Failed to import test module: TestFemApp
Traceback (most recent call last):
  File "/usr/lib/python3.9/unittest/loader.py", line 154, in loadTestsFromName
    module = __import__(module_name)
  File "/usr/lib/python3.9/site-packages/shiboken2/files.dir/shibokensupport/__feature__.py", line 142, in _import
    return original_import(name, *args, **kwargs)
  File "/home/<username>/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/TestFemApp.py", line 33, in <module>
    from femtest.app.test_mesh import TestMeshCommon as FemTest07
  File "/usr/lib/python3.9/site-packages/shiboken2/files.dir/shibokensupport/__feature__.py", line 142, in _import
    return original_import(name, *args, **kwargs)
  File "/home/<username>/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/femtest/app/test_mesh.py", line 33, in <module>
    import Fem
  File "/usr/lib/python3.9/site-packages/shiboken2/files.dir/shibokensupport/__feature__.py", line 142, in _import
    return original_import(name, *args, **kwargs)
ImportError: /usr/lib64/libSMDS.so: undefined symbol: _ZN19vtkUnstructuredGrid16GetCellNeighborsExP9vtkIdListS1_


----------------------------------------------------------------------

Temporary solution: run makepkg with the --nocheck option to skip check() in the PKGBUILD.

The package no longer builds without BUILD_ENABLE_CXX_STD=C++17 as they started using std::string_view.

pkgver(): git tag is lagging behind the project version: tag:0.19.2 where the project version is at 0.20.0.

patch: http://ix.io/3ASD/diff

Merge with git am -s < <(curl -s http://ix.io/3ASD)

Dear maintainers,

From the FreeCAD developer forums, "Your flags are a bit out of date, you should be using -D BUILD_ENABLE_CXX_STD=C++17, and the QT5 define is no longer used, we don't support Qt4 in 0.20 dev"

@adrianinsaval feel free to build the package with --nocheck if you're not interested in testing if anything is broken!

why do we need the freecad tests in check() if this is a development package? I don't see the benefit

the tip of master

@greyltc I think bernd fixed the upstream FEM tests. Is this package pulling freecad HEAD?

@greyltc, that's perfect! It installed just fine now. Thanks!

@K4LCIFER I expect those are upstream's bugs, but I'm not 100% sure. Have a look at lines 68 and 69 of the PKGBUILD as it is now. When I eliminate those two failing tests in my builds, check() passes.

@greyltc, here is the output of check() from the freecad-git installation.

You should be able to find the errors if you do a text search for FAIL, error or ERROR.

EDIT: Upon further inspection, I have found the following:

======================================================================
FAIL: test_box_frequency (femtest.app.test_ccxtools.TestCcxTools)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/<username>/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/femtest/app/test_ccxtools.py", line 102, in test_box_frequency
    self.result_reading_test(
  File "/home/<username>/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/femtest/app/test_ccxtools.py", line 323, in result_reading_test
    self.assertFalse(
AssertionError: True is not false : Invalid results read from .frd file

======================================================================
FAIL: test_thermomech_flow1D (femtest.app.test_ccxtools.TestCcxTools)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/<username>/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/femtest/app/test_ccxtools.py", line 158, in test_thermomech_flow1D
    self.result_reading_test(
  File "/home/<username>/.cache/yay/freecad-git/src/FreeCAD/build_dir/Mod/Fem/femtest/app/test_ccxtools.py", line 323, in result_reading_test
    self.assertFalse(
AssertionError: True is not false : Invalid results read from .frd file

----------------------------------------------------------------------
Ran 543 tests in 37.991s

FAILED (failures=2)
==> ERROR: A failure occurred in check().
    Aborting...
error making: freecad-git

@greyltc I'm ashamed to say... it turned out I had another version of freecad installed and that one was running. I think I had a git built version that I didn't realize.

@cculpepper1214 I suspect the libboost version error is actually Arch's fault. I had to recompile vtk to fix that.

@K4LCIFER are those errors in my packaging or upstream's errors?

Will not install because of multiple errors in check().

I am getting a compile/link error relating to pthreads. Not sure if upstream or my system is broken or what. Haven't been able to get any flavor of arch package of freecad to work for the past week or so. git commit for the freecad repo is 82e7ae0e35b7.

Performing C SOURCE FILE Test CMAKE_HAVE_LIBC_PTHREAD failed with the following output:
Change Dir: /home/plyons/.cache/yay/freecad-git/src/FreeCAD/build_dir/CMakeFiles/CMakeTmp

Run Build Command(s):/bin/ninja cmTC_ad2a2 && [1/2] Building C object CMakeFiles/cmTC_ad2a2.dir/src.c.o
[2/2] Linking C executable cmTC_ad2a2
FAILED: cmTC_ad2a2 
: && /bin/cc -march=x86-64 -mtune=generic -O2 -pipe -fno-plt -fexceptions         -Wp,-D_FORTIFY_SOURCE=2 -Wformat -Werror=format-security         
-fstack-clash-protection -fcf-protection -fPIC -w -Wl,-O1,--sort-common,--as-needed,-z,relro,-z,now CMakeFiles/cmTC_ad2a2.dir/src.c.o -o cmTC_ad2a2
   && :
/bin/ld: CMakeFiles/cmTC_ad2a2.dir/src.c.o: in function `main':
src.c:(.text.startup+0x28): undefined reference to `pthread_create'
/bin/ld: src.c:(.text.startup+0x32): undefined reference to `pthread_detach'
/bin/ld: src.c:(.text.startup+0x3c): undefined reference to `pthread_cancel'
/bin/ld: src.c:(.text.startup+0x48): undefined reference to `pthread_join'
collect2: error: ld returned 1 exit status
ninja: build stopped: subcommand failed.

@cculpepper1214 I had the same issue. Reinstalling all dependencies and a complete recompile solved the problem for me.

This does not run for me, it appears that the recent boost upgrade needs to be accounted for. FreeCAD: error while loading shared libraries: libboost_filesystem.so.1.75.0: cannot open shared object file: No such file or directory

If I downgrade to 1.75.0 of boost and boost-libs, I cannot compile, it fails because OpenSCAD expects 1.76...

I had the same issue with the aur appimage package, so this might be an upstream arch/freecad issue?

This doesn't build for me at the moment: it appears to be complaining that ospray has the wrong version (in the repos currently is 2.5.0).

~ » aur sync freecad-git
==> Using [custom] repository
  -> freecad-git: 0.19_pre.r5362.gcbc3fbd157-1 -> 0.19.r582.g0cfea3fee3-1
==> Retrieving package files
From https://aur.archlinux.org/freecad-git
 = [up to date]      master     -> origin/master
Running makepkg --syncdeps --clean
==> Making package: freecad-git 0.19.r796.g4a9963a03f-1 (Thu 10 Jun 2021 08:56:52 PM)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
  -> Updating FreeCAD git repo...
Fetching origin
==> Validating source files with md5sums...
    FreeCAD ... Skipped
==> Extracting sources...
  -> Creating working copy of FreeCAD git repo...
Reset branch 'makepkg'
==> Starting prepare()...
==> Starting pkgver()...
==> Removing existing $pkgdir/ directory...
==> Starting build()...
-- Compiler: GNU, version: 11.1.0
-- prefix: /usr/lib/freecad
-- bindir: bin
-- datadir: ../../share/freecad
-- docdir: ../../share/doc/freecad
-- includedir: include
-- libdir: ../../lib
-- cmake: 3.20.3
-- PyCXX found:
--   Headers:  /home/jade/.cache/aurutils/sync/freecad-git/src/FreeCAD/src
--   Sources:  /home/jade/.cache/aurutils/sync/freecad-git/src/FreeCAD/src/CXX
--   Version:  6.2.8
-- -- Found OCE/OpenCASCADE version: 7.5.0
-- -- OCE/OpenCASCADE include directory: /usr/include/opencascade
-- -- OCE/OpenCASCADE shared libraries directory: /usr/lib
-- VTK components: not found or used
CMake Error at /usr/lib64/cmake/vtk-9.0/VTK-vtk-module-find-packages.cmake:162 (find_package):
  Could not find a configuration file for package "ospray" that is compatible
  with requested version "1.8".

  The following configuration files were considered but not accepted:

    /usr/lib64/cmake/ospray-2.5.0/osprayConfig.cmake, version: 2.5.0
    /usr/lib/cmake/ospray-2.5.0/osprayConfig.cmake, version: 2.5.0
    /lib64/cmake/ospray-2.5.0/osprayConfig.cmake, version: 2.5.0
    /lib/cmake/ospray-2.5.0/osprayConfig.cmake, version: 2.5.0

Call Stack (most recent call first):
  /usr/lib64/cmake/vtk-9.0/vtk-config.cmake:131 (include)
  cMake/FreeCAD_Helpers/SetupSalomeSMESH.cmake:45 (find_package)
  CMakeLists.txt:58 (SetupSalomeSMESH)


-- Configuring incomplete, errors occurred!

Just tried to compile and I get broken OpenSCAD but not: AssertionError: 25.499999999999996 != 17.0 in check() mentioned by @greyltc in https://aur.archlinux.org/packages/freecad-git/#comment-807928

https://paste.ubuntu.com/p/DkwY9NY6Bf/

I think everything is sorted now. Please post here if you still have issues. If you're getting the AssertionError: 25.499999999999996 != 17.0 in check(), it's a known issue with OpenSCAD support and upstream is trying to fix it.

Just to confirm, getting similar issues:

Getting a mostly empty app window and the following messages in the log window:

12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Arch/Resources/icons/ArchWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Draft/Resources/icons/DraftWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Drawing/Resources/icons/DrawingWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Fem/Resources/icons/FemWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Image/Resources/icons/ImageWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Inspection/Resources/icons/InspectionWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Mesh/Resources/icons/MeshWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/OpenSCAD/Resources/icons/OpenSCADWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/PartDesign/Resources/icons/PartDesignWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Part/Resources/icons/PartWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Path/Resources/icons/PathWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Plot/resources/icons/PlotWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Points/Resources/icons/PointsWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Raytracing/Resources/icons/RaytracingWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/ReverseEngineering/Resources/icons/ReverseEngineeringWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Robot/Resources/icons/RobotWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Ship/resources/icons/ShipWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Sketcher/Resources/icons/SketcherWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Spreadsheet/Resources/icons/SpreadsheetWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Start/Resources/icons/StartWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Surface/Resources/icons/Surface_Workbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/TechDraw/Resources/icons/preferences-techdraw.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Test/Resources/icons/TestWorkbench.svg
12:20:04  Cannot find icon: /usr/lib/freecad/share/Mod/Web/Resources/icons/WebWorkbench.svg
12:20:05  Wizard shaft module cannot be loaded
12:20:05  No module named 'PartDesignGui'

@greyltc Thanks a lot for the quick response. Unfortunately it's not yet working correctly. I now get the following error messages:

Cannot find icon: /usr/lib/freecad/share/Mod/Arch/Resources/icons/ArchWorkbench.svg
[... for all mods ...]
No module named 'StartGui'

FreeCAD seems to really rely on not having it's folder structure changed. Also having something like "StartGui.so" or "Mesh.so" in "/usr/lib" seems confusing to me because it suggests it's something "global" and not related to a specific program. What's your take on this?

Ack! sorry about that, should hopefully be fixed now!

@greyltc: with the latest changes in the PKGBUILD my FreeCAD doesn't find any modules (message: "No modules found in /usr/Mod"). It seems "/usr/bin/FreeCAD" used to be a symlink to "/usr/local/freecad/bin/FreeCAD" and FreeCAD uses the executable path (stripped of the executable name and the bin folder) to determine it's "home path" (see https://github.com/FreeCAD/FreeCAD/blob/master/src/App/Application.cpp line 2972). However with the new folder structure the FreeCAD executable sits directly in "/usr/bin" and it will therefore search in "/usr/Mod" (instead of "/usr/lib/freecad/Mod") for it's modules.

@bartus Oh yeah. The file layout here is totally wrong now (thought I checked that). Thanks for pointing that out. I'll fix it when I have a free minute.

@Woodhousezb if your check function is failing with AssertionError: 25.499999999999996 != 17.0 within 1 places (8.499999999999996 difference) then I guess that's upstream's bug. You can move to some commit that does not have this or try to help them understand/fix it, or build with --nocheck if you want to ignore the check function failure. I made a bug report on it here: https://tracker.freecadweb.org/view.php?id=4622

@greyltc: would be nice to move out of /usr/local/freecad to e.g: /usr/lib/freecad/ ( check out Debian package file-list ) ( debian package cmake flags )

Hi there,

Kind of got the impression the problem got tackled, but I still get the same errors when trying to compile. @greylts: Is tis supposed to woork after your last comment?

@bartus sorry about that. I didn't mean to leave that testing line active in my commit.

@greyltc: Sorry but this won't work 10fb2eb61467 if users has no github config (username/email) or if build inside chroot (also it breaks pkgver() )

Please switch to --no-commit flag.

Generally for me check() fails when there's an upstream bug and passes when there's not.

There still might be an issue with the current install somehow leaking into the check() environment when you rebuild the package, any help with addressing that is welcome.

I think if the check function always(?) prevents a successful package build, it should be commented out, until a solution is available. If everyone builds with "--nocheck" it's of no use anyway. Or are there people out there for which the package builds fine with the all the checks successful?

I think the TestImportCSG failure is an upstream bug. I've reported it here: https://tracker.freecadweb.org/view.php?id=4622

Running makepkg with --nocheck worked, however uninstalling freecad did not. ( I just pacman -R'd it, no reboots though. I'll just run with --nocheck for now. Thank you!

I suspect the check() function is not being sandboxed properly and so part of the existing install is leaking into check() when it runs here. If anyone has any ideas for fixing that I'd love to hear them.

@cculpepper1214 you might try uninstalling freecad, then building this again or you can build it with --nocheck to hopefully avoid the issue for now.

Hello! I've been having trouble building this package, for about a week. The length of time makes me suspect something in my install is causing it, but I don't know what/how to find it. I've built from raw github source, and that worked fine.

Here's the output I'm getting:

testAll (TestPythonSyntax.PythonSyntaxTestCase) ... ok

======================================================================
FAIL: test_import_resize (OpenSCADTest.app.test_importCSG.TestImportCSG)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/chris/.cache/yay/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/OpenSCADTest/app/test_importCSG.py", line 352, in test_import_resize
    self.assertAlmostEqual (object.Shape.BoundBox.XLength, 2*8.5, 1)
AssertionError: 25.499999999999996 != 17.0 within 1 places (8.499999999999996 difference)

======================================================================
FAIL: test_known_quantity_units (femtest.app.test_material.TestMaterialUnits)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/chris/.cache/yay/freecad-git/src/FreeCAD/build_dir/check/Mod/Fem/femtest/app/test_material.py", line 77, in test_known_quantity_units
    self.assertTrue(
AssertionError: False is not true : Unit of quantity material parameter ElectricalConductivity is not known by FreeCAD unit system.

I've also updated to the latest openscad (from openscad-git). Does anyone have any ideas?

EDIT: Tried it again today, and it built without problem.

Hi, I'm encountering a build error in check(), not sure if it's the package's fault or mine:

======================================================================
ERROR: test_import_surface (OpenSCADTest.app.test_importCSG.TestImportCSG)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/tmp/makepkg/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/OpenSCADTest/app/test_importCSG.py", line 317, in test_import_surface
    doc = self.utility_create_scad(f"surface(file = \"{testfile}\", center = true, convexity = 5);", "surface_simple_dat")
  File "/tmp/makepkg/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/OpenSCADTest/app/test_importCSG.py", line 204, in utility_create_scad
    return importCSG.open(filename)
  File "/tmp/makepkg/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/importCSG.py", line 119, in open
    processcsg(tmpfile)
  File "/tmp/makepkg/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/importCSG.py", line 177, in processcsg
    result = parser.parse(f.read())
  File "/usr/lib/python3.9/site-packages/ply/yacc.py", line 333, in parse
    return self.parseopt_notrack(input, lexer, debug, tracking, tokenfunc)
  File "/usr/lib/python3.9/site-packages/ply/yacc.py", line 1120, in parseopt_notrack
    p.callable(pslice)
  File "/tmp/makepkg/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/importCSG.py", line 785, in p_surface_action
    obj.Shape,xoff,yoff=makeSurfaceVolume(p[3]['file'])
  File "/tmp/makepkg/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/OpenSCADFeatures.py", line 506, in makeSurfaceVolume
    with open(filename) as f1:
FileNotFoundError: [Errno 2] No such file or directory: 'makepkg/freecad-git/src/FreeCAD/build_dir/check/Mod/OpenSCAD/OpenSCADTest/data/Surface.dat'

----------------------------------------------------------------------
Ran 560 tests in 29.802s

FAILED (errors=1)
==> ERROR: A failure occurred in check().   (100.0 %)   
    Aborting...
error making: freecad-git

I did try removing yay's cache and the old installation, as mentioned in other comments.

Kunda, occ can't be patched here because it will need to be rebuilt with that patch. That would need to be proposed to the official occ package or submit a PKGBUILD for opencascade patched to the AUR but I don't think many people would be willing to compile occ themselves. Here is the pkgbuild for the official package: https://github.com/archlinux/svntogit-community/blob/packages/opencascade/trunk/PKGBUILD maybe submit a PR to patch the source there?

If opencascade is broken, seems like the official package deserves the patch no?

Upstream just cut a new release called 0.19. Can we use the patched OCCT repo (it has the fillet/chamfer fix in it) https://gitlab.com/blobfish/occt

Here's the patch by itself: https://gitlab.com/blobfish/occt/-/commit/ad0ba55f55b36dc957f66192c4766ace83f82b7e

It looks like installing to /usr/local is recommended against by the Arch Package Guidelines. Would you mind having freecad-git either install to /opt or split between /usr/share and /usr/lib like the official package does?

Apparently upstream FreeCAD has patches for OCCT that aren't official but fix some key issues. Lately there is a patch floating around for fixing a long standing chamfer/fillet OCC crash. I know that the OCC package maintainer for Debian is working to incorporate said patches in to their package. Perhaps we here at Arch could do the same?

Context
Patch announced: https://twitter.com/thekurtwk/status/1357887322264662022

The OCCT mirror with patches: https://gitlab.com/blobfish/occt (maintained by a member of the FC community 'tanderson' AKA 'blobfish')

Tweet by Debian maintainer: https://twitter.com/thekurtwk/status/1357887322264662022

Seems like it's a problem with the package. If you remove freecad-git and then reinstall it works.

Indeed it works after uninstalling the current freecad-git package. Thanks for the tip! :-)

Seems like it's a problem with the package. If you remove freecad-git and then reinstall it works.

Yes, I've tried, but didn't find anything. But they don't use Github as Bugtracker and I may be to stupid to use Mantis the right way. To be honest, it kept me from reporting from bugs in the past, since you need to report to the forums first (register there and in the bugtracker etc.) and stuff like that and it's really an endeavor to report a simple bug, it's getting frustrating fast.

Yeah, check() fails for me too. I assume this test failure is upstream's bug. Have you tried looking for or reporting an issue there?

It's me again. :-) Last time it was my own fault for not completely deleting the src and pkg directories. After that everything worked fine... for a while.

But since a few days I only get this:

Ran 246 tests in 31.672s

FAILED (failures=1, errors=49)
==> ERROR: A failure occurred in check().
    Aborting...

The build itself seems to work fine, but the tests always fail (which they didn't before).

Should I just ignore/comment them in the PKGBUILD or do I need to actually care about the result of the unit tests?

I've tried to delete everything and start fresh, I also waited a few days and tried again a few times. But it's always the same...

Example output:

======================================================================
ERROR: test_draft_to_sketch (drafttests.test_modification.DraftModification)
Convert a Draft object to a Sketch and back.
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/cf/abs/aur/freecad-git/src/FreeCAD/build_dir/check/Mod/Draft/drafttests/test_modification.py", line 375, in test_draft_to_sketch
    obj = Draft.make_sketch(wire, autoconstraints=True)
  File "/home/cf/abs/aur/freecad-git/src/FreeCAD/build_dir/check/Mod/Draft/draftmake/make_sketch.py", line 79, in make_sketch
    from Sketcher import Constraint
  File "/usr/lib/python3.9/site-packages/shiboken2/files.dir/shibokensupport/__feature__.py", line 142, in _import
    return original_import(name, *args, **kwargs)
ImportError: /home/cf/abs/aur/freecad-git/src/FreeCAD/build_dir/check/lib/Sketcher.so: undefined symbol: _ZN3App4Part15getPartOfObjectEPKNS_14DocumentObjectEb

======================================================================
FAIL: testExtensions (Document.DocumentBasicCases)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/cf/abs/aur/freecad-git/src/FreeCAD/build_dir/check/Mod/Test/Document.py", line 248, in testExtensions
    grp.addExtension("App::GroupExtensionPython")
TypeError: function takes exactly 2 arguments (1 given)

During handling of the above exception, another exception occurred:

Traceback (most recent call last):
  File "/home/cf/abs/aur/freecad-git/src/FreeCAD/build_dir/check/Mod/Test/Document.py", line 255, in testExtensions
    self.failUnless(False)
AssertionError: False is not true

News: upstream has announced feature freeze with release for v0.19 pending next month.

packages fine for me. tested with both makepkg and yay

Anyone getting this error after the build and test have finished?:

mv: cannot stat '/home/xxx/abs/aur/freecad-git/pkg/freecad-git/usr/local/freecad/share/doc': No such file or directory
==> ERROR: A failure occurred in package().
Aborting...

I just made an update that I hope will address any outstanding issues.

@greyltc I suggest to remove the execution of the test suite. As it is now the PKGBUILD should be considered broken.

@alhirzel, seems like you're right about that recource directory. Does anyone know how to set that at runtime?

@Deewiant mentioned your observation in the FreeCAD forum: https://forum.freecadweb.org/viewtopic.php?f=4&t=50810&p=461930#p461930

I'm seeing a few issues in check() right now, which could be upstream. I am building on a PC where FreeCAD is not installed, and the first set of errors are like the following:

  File "/home/alhirzel/build_freecad/freecad-git/src/build/Mod/Arch/ArchIFCSchema.py", line 34, in <module>
    with open(os.path.join(FreeCAD.getResourceDir(), "Mod", "Arch", "Presets",
FileNotFoundError: [Errno 2] No such file or directory: '/usr/share/freecad/Mod/Arch/Presets/ifc_contexts_IFC4.json'

I think it's because FreeCAD.getResourceDir() follows Application::getResourceDir which depends on RESOURCEDIR at build-time. Perhaps the tests should not use this directory or something. I'm reporting this here just in case it is due to our packaging or our invocation of the test suite. (I did not see any mention of RESOURCEDIR in the PKGBUILD.)

Second set of issues are as follows, which are filesystem-related though I struggle to see how they could be due to the PKGBULID:

======================================================================
ERROR: testHatchCase (TestTechDrawApp.TechDrawTestCases)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/alhirzel/build_freecad/freecad-git/src/build/Mod/TechDraw/TestTechDrawApp.py", line 53, in testHatchCase
    rc = DHatchTest()
  File "/home/alhirzel/build_freecad/freecad-git/src/build/Mod/TechDraw/TDTest/DHatchTest.py", line 45, in DHatchTest
    hatch = FreeCAD.ActiveDocument.addObject('TechDraw::DrawHatch','Hatch')
Base.FreeCADError: {'sclassname': 'N4Base15FileSystemErrorE', 'sErrMsg': 'File /tmp/FreeCAD_Doc_6a76e492-caf2-464f-af89-d311de155634_da39a3_47564/FreeCAD_Doc_6a76e492-caf2-464f-af89-d311de155634_da39a3_47564HatchSvgHatch.svgpg7I9M does not exist.', 'sfile': '', 'iline': 0, 'sfunction': '', 'swhat': 'File /tmp/FreeCAD_Doc_6a76e492-caf2-464f-af89-d311de155634_da39a3_47564/FreeCAD_Doc_6a76e492-caf2-464f-af89-d311de155634_da39a3_47564HatchSvgHatch.svgpg7I9M does not exist.', 'btranslatable': False, 'breported': True}

======================================================================
ERROR: testSectionCase (TestTechDrawApp.TechDrawTestCases)
----------------------------------------------------------------------
Traceback (most recent call last):
  File "/home/alhirzel/build_freecad/freecad-git/src/build/Mod/TechDraw/TestTechDrawApp.py", line 85, in testSectionCase
    rc = DVSectionTest()
  File "/home/alhirzel/build_freecad/freecad-git/src/build/Mod/TechDraw/TDTest/DVSectionTest.py", line 42, in DVSectionTest
    section = FreeCAD.ActiveDocument.addObject('TechDraw::DrawViewSection','Section')
Base.FreeCADError: {'sclassname': 'N4Base15FileSystemErrorE', 'sErrMsg': 'File /tmp/FreeCAD_Doc_9ddf8b87-511a-47e5-a564-d1935042ba5b_da39a3_47564/FreeCAD_Doc_9ddf8b87-511a-47e5-a564-d1935042ba5b_da39a3_47564SectionSvgHatch.svgD7AOwM does not exist.', 'sfile': '', 'iline': 0, 'sfunction': '', 'swhat': 'File /tmp/FreeCAD_Doc_9ddf8b87-511a-47e5-a564-d1935042ba5b_da39a3_47564/FreeCAD_Doc_9ddf8b87-511a-47e5-a564-d1935042ba5b_da39a3_47564SectionSvgHatch.svgD7AOwM does not exist.', 'btranslatable': False, 'breported': True}

Doesn't build with both ccache and distcc enabled. The CMakeLists.txt has some custom logic to use ccache in front of every compilation command if it's found, which combined with makepkg's logic to set up the PATH for ccache results in invocations like /usr/bin/ccache /usr/lib/ccache/bin/c++. With distcc also enabled, both the "outer" and "inner" ccache programs trigger distcc which causes it to bail out with "CRITICAL! distcc seems to have invoked itself recursively!".

This seems like FreeCAD's bug, but here's a patch:

diff --git CMakeLists.txt CMakeLists.txt
index f4dabf7657..ded1023532 100644
--- CMakeLists.txt
+++ CMakeLists.txt
@@ -8,11 +8,6 @@ if (POLICY CMP0072)
     cmake_policy(SET CMP0072 OLD)
 endif(POLICY CMP0072)

-find_program(CCACHE_PROGRAM ccache)  #This check should occur before project()
-if(CCACHE_PROGRAM)
-    set_property(GLOBAL PROPERTY RULE_LAUNCH_COMPILE "${CCACHE_PROGRAM}")
-endif()
-
 project(FreeCAD)

 set(PACKAGE_VERSION_NAME "Vulcan")

I'm getting the following error while compiling https://0x0.st/iCeH.txt

@bartus Thanks very much for the tips!

• I've added the python-ply dep
• I never knew the environment was manipulated differently between prepare() and build() with respect to how rules in the options array are handled! (this honestly feels like a build system design flaw to me) I've moved the build preparation stuff from prepare() to build() now. Seems like I've been wrong to do that kind of preparation in prepare().
• Yeah, I should put my patches in the source array and not fetch anything in prepare(), sometimes I'm just too lazy to do it correctly
• Not sure I agree with what you've done with MAKEFLAGS and ninja there. I think I won't do that.

I patched my PKGBUILD with the patch posted by @bartus; it helped with building (set makeflags correctly, much slower build because of this I assume) but it still was failing the self tests so I think maybe this is an issue with FreeCAD. I commented out check() and then it just skips to package() which runs fine and the program is installed. I can open it and there are no obvious issues (I checked for the features that were failing test, all related to helix patterning). Maybe we should remove the self test until it's known to pass? Otherwise it's impossible to install, unless it's just a problem with my machine. Also, during check() we should remove the directory change; it enters the correct one at the end of Ninja build anyway.

To patch the makefile you can do like this, for anyone reading before it gets pushed:
curl http://ix.io/2IKZ >> freecad.patch
patch -ruN < freecad.patch

@greyltc; few things:

• missing python-ply dependency ##

• cmake should be called form build() function as prepare() is called in sanitized environment (no flags defined in /etc/makepkg.conf present) ##

• to recreate make behaviour with options=(!makeflags) using ninja you should add ${MAKEFLAGS:--j1} to ninja call ##

• since cmake:3.12 you can used cmake -S source_dir -B build_dir call creplacing all mkdir build, cd build, cmake ../ ##

• and as bonus, instead of calling curl to get patch you can add one to source list with filename::url syntax ## also you can combine multiple python packages using bash bracket expansions ##

patch

# merge with:
git am -s < <(curl -s http://ix.io/2IKZ)

btw. gendesktop is pretty neat - will try to use it in my PKGBUILDs (ღ˘⌣˘ღ)

I think python-yaml needs to be a dependency of some sort

From Arch freecad package: Adding -D BUILD_ENABLE_CXX_STD=C++14 to cmake changes the std as needed.

@khumarahn, I was able to make it build by replacing all occurrences of -std=gnu++11 in build/build.ninja with -std=gnu++14 after prepare() have been run. I could not figure out how to get CMake to spit out a correct build.ninja.

Installing python-netcdf4 I could start the compilation, but it fails with

[2379/3992] Building CXX object src/Mod/Path/App/CMakeFiles/Path.dir/Area.cpp.o
FAILED: src/Mod/Path/App/CMakeFiles/Path.dir/Area.cpp.o 
/usr/bin/c++ -DBOOST_ALL_NO_LIB -DBOOST_FILESYSTEM_DYN_LINK -DBOOST_PROGRAM_OPTIONS_DYN_LINK -DBOOST_REGEX_DYN_LINK -DBOOST_SYSTEM_DYN_LINK -DBOOST_THREAD_DYN_LINK -DHAVE_CONFIG_H -DHAVE_FFMPEG -DHAVE_FREEIMAGE -DHAVE_LIMITS_H -DHAVE_RAPIDJSON -DHAVE_TBB -DHAVE_VTK -DOCC_CONVERT_SIGNALS -DPath_EXPORTS -DQT_CORE_LIB -DQT_NO_DEBUG -DQT_XML_LIB -DVTK_OPENGL2_BACKEND -D_OCC64 -DvtkDomainsChemistry_AUTOINIT="2(vtkDomainsChemistryOpenGL2,vtkDomainsParallelChemistry)" -DvtkFiltersCore_AUTOINIT="1(vtkFiltersParallelDIY2)" -DvtkFiltersFlowPaths_AUTOINIT="1(vtkFiltersParallelFlowPaths)" -DvtkFiltersParallel_AUTOINIT="2(vtkFiltersParallelDIY2,vtkFiltersParallelGeometry)" -DvtkIOExodus_AUTOINIT="1(vtkIOParallelExodus)" -DvtkIOExport_AUTOINIT="2(vtkIOExportOpenGL2,vtkIOExportPDF)" -DvtkIOGeometry_AUTOINIT="1(vtkIOMPIParallel)" -DvtkIOImage_AUTOINIT="1(vtkIOMPIImage)" -DvtkIOParallel_AUTOINIT="1(vtkIOMPIParallel)" -DvtkIOSQL_AUTOINIT="2(vtkIOMySQL,vtkIOPostgreSQL)" -DvtkIOXdmf3_AUTOINIT="1(vtkIOParallelXdmf3)" -DvtkRenderingContext2D_AUTOINIT="1(vtkRenderingContextOpenGL2)" -DvtkRenderingCore_AUTOINIT="3(vtkInteractionStyle,vtkRenderingFreeType,vtkRenderingOpenGL2)" -DvtkRenderingFreeType_AUTOINIT="2(vtkRenderingFreeTypeFontConfig,vtkRenderingMatplotlib)" -DvtkRenderingLICOpenGL2_AUTOINIT="1(vtkRenderingParallelLIC)" -DvtkRenderingOpenGL2_AUTOINIT="1(vtkRenderingGL2PSOpenGL2)" -DvtkRenderingVolume_AUTOINIT="1(vtkRenderingVolumeOpenGL2)" -I. -Isrc -I../src -Isrc/Mod/Path/App -I/usr/include/opencascade -I/usr/include/eigen3 -I/usr/include/python3.9 -isystem /usr/include/qt -isystem /usr/include/qt/QtCore -isystem /usr/lib/qt/mkspecs/linux-g++ -isystem /usr/include/qt/QtXml -Wall -Wextra -Wno-write-strings -march=x86-64 -mtune=generic -O2 -pipe -fno-plt -fPIC -w -O3 -DNDEBUG -fPIC   -fext-numeric-literals -pthread -fPIC -std=gnu++11 -MD -MT src/Mod/Path/App/CMakeFiles/Path.dir/Area.cpp.o -MF src/Mod/Path/App/CMakeFiles/Path.dir/Area.cpp.o.d -o src/Mod/Path/App/CMakeFiles/Path.dir/Area.cpp.o -c ../src/Mod/Path/App/Area.cpp
In file included from /usr/include/boost/geometry/core/tag.hpp:23,
                 from /usr/include/boost/geometry/core/ring_type.hpp:28,
                 from /usr/include/boost/geometry/core/closure.hpp:24,
                 from /usr/include/boost/geometry/geometry.hpp:34,
                 from /usr/include/boost/geometry.hpp:17,
                 from ../src/Mod/Path/App/Area.cpp:38:
/usr/include/boost/geometry/util/type_traits_std.hpp:54:25: error: expected template-name before ‘<’ token
   54 |     : std::conditional_t<Trait::value, conjunction<Traits...>, Trait>
      |                         ^

It seems, the dependency python-netcdf4 is missing.

It might just be a bug in the Summary report printer?
Above I can see that Coin is found and has the expected version:

-- Found OpenGLU: /usr/lib/libGLU.so
-- Checking for module 'Coin'
--   Found Coin, version 4.0.1
-- Coin3D libraries found
-- Found Spnav: /usr/lib/libspnav.so  

@greyltc, opened a thread on the FreeCAD forum: https://forum.freecadweb.org/viewtopic.php?f=4&t=52921 to see if we can glean any insight from the packagers there.

@Kunda, that's a good question and I was wondering that myself. Maybe the system coin is not being detected properly? Any suggestions welcome.

Quick question: in the cmake summary report, why doesn't Coin3D library output a version number?

-- Freetype:            2.10.4
-- OpenGLU:             /usr/lib/libGLU.so [/usr/lib/libGLU.so][/usr/include]
-- Coin3D:              [Coin] []
-- SPNAV:               [/usr/lib/libspnav.so] [/usr/include]

@greyltc thank you so much for maintaining this package!

@bartus
What problem does that solve?

Could we have ninja ${MAKEFLAGS:--j1} to recreate options=(!makeflags) expected behaviour (by default ninja build with nproc+2 threads)?

Thank you! Build works for me again, too! :)

Builds fine for me in a clean chroot as of today.

However, please add qt5-xmlpatterns to depends.

Does anyone else has difficulties building?

It seems related (again) to the slightly annoying smesh lib.

[ 37%] Linking CXX shared library ../../../lib/libSMESHDS.so
[ 39%] Built target FreeCADApp
[ 39%] Built target FreeCADMainPy
[ 39%] Built target FreeCADMainCmd
[ 39%] Built target Image
[ 39%] Built target Points
[ 39%] Built target Start
[ 39%] Built target Web
[ 39%] Built target Spreadsheet
[ 42%] Built target Mesh
/usr/bin/ld: CMakeFiles/SMESHDS.dir/src/SMESHDS/SMESHDS_GroupBase.cpp.o: in function `SMESHDS_GroupBase::SMESHDS_GroupBase(int, SMESHDS_Mesh const*, SMDSAbs_ElementType)':
SMESHDS_GroupBase.cpp:(.text+0x461): undefined reference to `Quantity_Color::Quantity_Color()'
/usr/bin/ld: CMakeFiles/SMESHDS.dir/src/SMESHDS/SMESHDS_GroupBase.cpp.o: in function `SMESHDS_GroupBase::GetColorGroup() const':
SMESHDS_GroupBase.cpp:(.text+0x653): undefined reference to `Quantity_Color::Red() const'
/usr/bin/ld: SMESHDS_GroupBase.cpp:(.text+0x661): undefined reference to `Quantity_Color::Green() const'
/usr/bin/ld: SMESHDS_GroupBase.cpp:(.text+0x670): undefined reference to `Quantity_Color::Blue() const'
collect2: error: ld returned 1 exit status
make[2]: *** [src/3rdParty/salomesmesh/CMakeFiles/SMESHDS.dir/build.make:350: lib/libSMESHDS.so] Error 1
make[1]: *** [CMakeFiles/Makefile2:2403: src/3rdParty/salomesmesh/CMakeFiles/SMESHDS.dir/all] Error 2
make[1]: *** Waiting for unfinished jobs....
[ 55%] Built target FreeCADGui
[ 61%] Built target Part
make: *** [Makefile:149: all] Error 2
==> ERROR: A failure occurred in build().
Aborting...

0.19 is the git version not 0.18 ... Please update

Вы напрасно боретесь за работу конкретно этой сборки FC - она имеет родовую проблему = топологии наименования (когда ломаются те эскизы и детали, которые ссылаются на другие эскизы после того, как нижерасположенные эскизы меняются). Эта проблема решена (или почти решена) в сборке linkstage, которая тоже здесь есть https://aur.archlinux.org/packages/freecad-linkstage3-git/ и является вполне рабочей, реально собирается и работает. You are fighting in vain for the work of this particular assembly FC - it has a generic problem = name topology (when those sketches and details that refer to other sketches after the below sketches change). This problem is solved (or almost solved) in the linkstage assembly, which is also there and is quite working https://aur.archlinux.org/packages/freecad-linkstage3-git/ , actually assembled and working.

when i try to build freecad-git, i am asked to choose a provider for med with options:

• med-openmpi 4.0.0
• med-salome 3.3.1
• med3 3.3.1

if i choose med3 or med-openmpi, i get the error:

unable to satisfy dependency 'python-pivy'

When I try to build the package, I get a bunch of cmake errors related to OpenCascade. Adding PATH=/usr/bin:$PATH to build() as in the the PKGBUILD of the freecad package fixes it for me.

Could you please add that to this PKGBUILD as well?

Some package update broke scrollwheel zoom for me in FreeCAD 0.19-git. Scroll wheel works fine for scrolling in dialogs, report view and the welcome screen etc, but doesn't work for zooming in/out in models in the 3d view. Tried different Navigation Styles to no avail. Does anyone else have this problem? What (QT5?) package is responsible for the 3d view? Latest 0.19 appimage works fine btw.

Since I still couldn't create meshes using the standard mesher, I've asked in the FreeCAD forums: https://forum.freecadweb.org/viewtopic.php?f=4&t=47375

A user mentioned, that he is using the package med-openmpi instead of med and hdf5-openmpi instead of hdf5. I've tried that and it did indeed solve the issue for now. Maybe this can help others, too.

@greyltc you need to rebuild this package after updating to qt5 5.15.

@bartus: looks like your patch has already been accepted upstream. Thanks!

Wont build with latest qt5:5.15.0.

• upstream:PR
• PKGBUILD:patch

#apply patch with.
git am -s < <(curl -s http://ix.io/2ogn)

@SunDoctor that seems to be a problem with Nouveau. Did you try with the proprietary Nvidia driver?

Freecad is very unstable on my Arch. :-( In Sketch mode it always breaks with log:

Program received signal SIGSEGV, Segmentation fault.

0 /usr/lib/libc.so.6(+0x3bd70) [0x7f4670d53d70]

1 /usr/lib/dri/nouveau_dri.so(+0xe86395) [0x7f4669636395]

2 /usr/lib/dri/nouveau_dri.so(+0xc360fe) [0x7f46693e60fe]

3 0x7f4672ce2da6 in NaviCubeImplementation::drawNaviCube(bool) from /usr/lib/freecad/lib/libFreeCADGui.so+0xa46

4 0x7f4672cc18b5 in Gui::View3DInventorViewer::renderScene() from /usr/lib/freecad/lib/libFreeCADGui.so+0x285

5 0x7f4672c93855 in SIM::Coin3D::Quarter::QuarterWidget::paintEvent(QPaintEvent*) from /usr/lib/freecad/lib/libFreeCADGui.so+0x1b5

6 0x7f4672c9b84e in SIM::Coin3D::Quarter::SoQTQuarterAdaptor::paintEvent(QPaintEvent*) from /usr/lib/freecad/lib/libFreeCADGui.so+0x1e

7 0x7f46717cf2ee in QWidget::event(QEvent*) from /usr/lib/libQt5Widgets.so.5+0x27e

8 0x7f4671880cbf in QFrame::event(QEvent*) from /usr/lib/libQt5Widgets.so.5+0x1f

9 0x7f4671af4b23 in QGraphicsView::viewportEvent(QEvent*) from /usr/lib/libQt5Widgets.so.5+0x163

....

83 0x7f467139c066 in QCoreApplication::exec() from /usr/lib/libQt5Core.so.5+0x96

84 0x7f46729edcea in Gui::Application::runApplication() from /usr/lib/freecad/lib/libFreeCADGui.so+0x153a

85 freecad(main+0x6f9) [0x55e9d1fda539]

86 /usr/lib/libc.so.6(__libc_start_main+0xf3) [0x7f4670d3f023]

87 freecad(_start+0x2e) [0x55e9d1fda84e

@lonaowna: Thank you for the tip! After an short attempt at downgrading didn't work, I decided to download the current AppImage version to create meshes. I can live with that, until the issue is resolved. :-)

@neo2001 yes, there are some compatibility issues with HDF5 1.12 and MED. See also the discussion over at the med package.

I think downgrading HDF5 should provide a temporary solution.

Although I did not recompile in the last weeks, suddenly I can't create meshes using the default mesher anymore (also for older, existing models).

I just get an error: libhdf5.so.103: cannot open shared object file: No such file or directory

I did recompile it (without errors) after that, but it didn't change anything.

Can anyone confirm this?

The PYSIDE2RCCBINARY issue was fixed upstream a while ago: https://github.com/FreeCAD/FreeCAD/pull/2846

If you are still having trouble building, remove your build files and try again. If there are still issues, leave a comment here.

Could we have this patch as a quick fix, package won't build without it. https://github.com/FreeCAD/FreeCAD/pull/2964

Just add -DPYSIDE2RCCBINARY=/usr/bin/rcc to cmake options in PKGBUILD to resolve issue with PYSIDE2RCCBINARY-NOTFOUND

This package does not build with PySide2 5.14 (currently in testing). It fails with the following message:

[  4%] Generating Material_rc.py
/bin/sh: PYSIDE2RCCBINARY-NOTFOUND: command not found

See FreeCAD issue 4229 for more information.

Tried with latest PySide2 and Shiboken2 but problem still exists

@s218, it's great that you found a workaround. but for me, tweaking pkgbuild and stuff is counterproductive. Plus, the time to compile this package takes a long time. I did use distcc to speed the process but then comes compilation issues... I'll currently stick with the AppImage version, and it's delivered by Freecad developers, rather official. Thanks.

upd 2019-11-24 уже исправлено, т.е. можно не читать is already fixed, so you don't have to read -

this package is built, installed and running if you change the string -

-DPYSIDE_LIBRARY="/usr/lib/libpyside2.cpython-${PYVER//.}m-${CARCH}-linux-gnu.so" \

to this one (delete the letter m)

-DPYSIDE_LIBRARY="/usr/lib/libpyside2.cpython-${PYVER//.}-${CARCH}-linux-gnu.so" \

https://forum.freecadweb.org/viewtopic.php?f=4&t=40986&sid=deb0a69eb4a7450e8e49be29332fec17

Pyside2 + python3.8 is not yet working. pyside2 version 5.14 will adress these issue. You can follow these links to get more information: https://github.com/conda-forge/freecad- ... -551993747 https://bugreports.qt.io/browse/PYSIDE-939

This program assembly breaks the related sketches. That is, if you make a sketch of 1 figure A and then in the next sketch (2) to use a link to this figure A from the sketch 1, and then in the first sketch to change the same figure A, then in the sketch 2 we get an accident. This problem is solved in https://aur.archlinux.org/packages/freecad-linkstage3-git, which, however, also is not going to. So it makes sense to finish it off.

oh well. after all that, I still have this problem; can't import DWG or DXF

Traceback (most recent call last):
  File "<input>", line 1, in <module>
  File "/usr/lib/freecad/Mod/Draft/importDWG.py", line 55, in <module>
    from DraftTools import translate
  File "/usr/lib/freecad/Mod/Draft/DraftTools.py", line 44, in <module>
    from DraftGui import todo, translate, utf8_decode
  File "/usr/lib/freecad/Mod/Draft/DraftGui.py", line 279, in <module>
    class DraftBaseWidget(QtGui.QWidget):
TypeError: 'Shiboken.ObjectType' object is not iterable[/code]

@springwurm much appreciated - I did not notice this issue and have been struggling all afternoon to compile successfully with git versions of pyside and shiboken - both of which were redundant with the change you noted.

Hi got this only compiled, when i change the cmake option DPYSIDE_LIBRARY to "/usr/lib/libpyside2.cpython-${PYVER//.}-${CARCH}-linux-gnu.so". Removed the 'm'

Make Error: make[2]: *** Keine Regel vorhanden, um das Ziel „/usr/lib/libpyside2.cpython-38m-x86_64-linux-gnu.so“, benötigt von „lib/libFreeCADGui.so“, zu erstellen. Schluss.

Oh ok I understand now! Thanks again for answering and for providing this package ;)

@bebR uhh yeah but the diff shows r2626 instead of r2526 (like I was saying)..

In any case; it doesn't matter. Since this is a -git package, the pkgver is calculated dynamically based on the latest git commits. And the pkgver should increase each time there is a new commit. So the PKGBUILD will also change every time you run makepkg after there has been an upstream commit.

But again, as long as it's working it doesn't really matter. Let me know if you have any other issues!

@lonaowna, thanks for your answer.

Actually no, there is no typo here. This is the git diff I get:

diff --git i/PKGBUILD w/PKGBUILD
index 81b7f88..b7b87d4 100644
--- i/PKGBUILD
+++ w/PKGBUILD
@@ -3,7 +3,7 @@

 _appname=freecad
 pkgname="${_appname}-git"
-pkgver=0.18.r2576.g953b802f83
+pkgver=0.18.r2626.gc021ff70d
 pkgrel=1
 epoch=2
 pkgdesc='A general purpose 3D CAD modeler - git checkout'

@bebR, good to hear that you managed to fix it. It is expected that the version in the PKBUILD changes as it reflects the latest commit in the upstream git repository.

The .r*. part should always increase though. I think you made a typo and are at r2626 instead of r2526.

Forget my previous comment. I had a lot a threads because I used to use distcc, but the distributed compilation currently does not work anymore in my environment.

Decreasing MAKEOPTS from -j40 value to -j3 solved it.

However, after compilation, the PKGBUILD gets modified, which is a bit curious. The diff shows pkgver=0.18.r2526.gc021ff70d instead of pkgver=0.18.r2576.g953b802f83.

I cannot build Freecad anymore. The cc1plus gets killed by the kernel because of an out of memory. This used to work. I have 8Gb of ram. Any clue if a lib dependency could cause this? too many recursions or so? Thanks

@bartus thanks for the report. Fixed.

pkgver(): git describe generate lower commit count than what's on AUR:

$ git clone https://aur.archlinux.org/freecad-git.git && cd freecad-git
$ makepkg -Cod
==> Making package: freecad-git 1:0.19pre.r2096.gc903de8af7-1
...
==> Updated version: freecad-git 1:0.19pre.r4.g64cf85c21-1
==> Sources are ready

Following the git-describe(1) man page, second version is correct one.

Since there is no response from upstream on the openmpi issue, I've added it as a mandatory dependency for now.

@Ionaowna https://forum.freecadweb.org/viewtopic.php?t=35505#p301527 Does changing:

-pkg_search_module(OPENMPI ompi-cxx)
+pkg_search_module(OPENMPI REQUIRED ompi-cxx)

fix it?

Just a soft heads up: Opencascade 7.4 will be released imminently

http://git.dev.opencascade.org/gitweb/?p=occt.git;a=commit;h=fd47711d682be943f0e0a13d1fb54911b0499c31

No upstream FreeCAD OCC bugs were fixed. But it would be useful nevertheless to test the latest 7.4 with 0.19.x

@Nahouhak, thanks for the follow up. There's something wrong with the MPI detection in FreeCAD. See this forum thread for more info.

I get an error recently (was working about a week ago). A missing unlisted dependency ? [ 0%] Linking CXX shared library ../../../lib/libDriver.so /usr/bin/ld : can't find -lmpi_cxx /usr/bin/ld : can't find -lmpi

Edit : Adding openmpi solves the issue so I think openmpi is the missing unlisted dependency.

An upstream issue that the FreeCAD team is having issue with https://forum.freecadweb.org/viewtopic.php?f=42&t=37216#p316220 (Ticket #4014 - Finding a solution for the 'version.h' issue)

@Ionaowna, I discovered that the version I had installed was the "Staging" one, so I uninstalled it. Then I installed the 1:0.19 offered by manjaro.

No more update notification now, all fine.

Thanks for your work!

@cerdicola thanks for the heads-up! It looks like the FreeCAD folks created a "Staging" tag which messed things up. I have fixed our PKGBUILD to ignore this tag.

I'm not sure if Manjaro will pick this up automatically. Let us know if this is still broken in a day or so.

From manjaro deepin, I got a freecad-git update, after compile and install keeps notifying. New offered version is: (1:Staging.r0.g7b315d3a1-1)

And I have now: 1:0.19pre.r633.g063582241-1

Please can I get some help I was compiling the freecad package 0.18.* , but after my last system update I was not able to get it to recompile I have switched to this one, but am having the similar/same issue, looking at cmake error logs point to pthread.so issues but I have seem to have that lib, but maybe ld is not detecting it correctly, it also oddly complains about opencascade while compiling, but nothing in the cmake error file about that just on screen during compile, everything but freecad is from community unless I missed one double checked opencascade, pyside2 and shiboken2 as these have all caused update issues previously but the community build seemed to work the last time I had to recompile.

cmake error log https://pastebin.com/jmsYTvD4

Cmake log https://pastebin.com/TBZQbckc

Terminal compile time https://pastebin.com/3P4FvM42

FYI, new official FreeCAD Packaging subforum https://forum.freecadweb.org/viewforum.php?f=42

@OOFscr @bebR, sorry for the late response. I've fixed the desktop file.

@greyltc please fix the .desktop issue or disown the package.

I have the same issue than OOFscr, I had to change the .desktop to make it work.

And I agree with Saku, the git clone is extremely long, almost 20 minutes.

Please update the freecad.desktop.

Hi,

Maybe change the pkgbuild to pull only depth 1 and not all the repository which is huge.

For me, the current freecad.desktop file prevents correct opening of files in xfce. I had to change: Path=/usr/lib/freecad

This package has been updated to use Python 3. Please leave a comment if you find something broken.

depends missing qt5-webengine

@dtbaumann thanks. I've incorporated your suggestion and some from the freecad and freecad-python3-git packages into the following PKGBUILD which uses python3.

https://gist.github.com/slokhorst/4ea6a978ad2287724f5c573cbf89a373

Please give it a try. If there are no problems, I'll update the freecad-git PKGBUILD (this package) in a few days.

the build block below works without python2

depends=('boost-libs' 'curl' 'opencascade>=7.2' 'xerces-c' 'libspnav' 'glu' 'netcdf'
     'shared-mime-info' 'hicolor-icon-theme' 'jsoncpp'
     'qt5-base' 'qt5-declarative' 'qt5-svg' 'qt5-tools'
     'med' 'python-pivy' 'python-pyside2' 'python-matplotlib' 'pyside2-tools')
makedepends=('boost' 'eigen' 'gcc-fortran' 'swig' 'xerces-c' 'desktop-file-utils' 'git'
         'cmake' 'coin' 'python-shiboken2' 'pyside2' 'shiboken2')
optdepends=('python-matplotlib' 'python-pyqt5' 'graphviz' 'openscad')

build() {
    cd "${srcdir}/freecad-code"
    cmake . \
        -DCMAKE_BUILD_TYPE=Release \
        -DCMAKE_INSTALL_PREFIX="/usr/lib/freecad" \
        -DCMAKE_INSTALL_DOCDIR="/usr/share/freecad/doc" \
        -DCMAKE_INSTALL_DATADIR="/usr/share/freecad" \
        -DFREECAD_USE_OCC_VARIANT="Official Version" \
        -DFREECAD_USE_EXTERNAL_PIVY=ON \
        -DBUILD_QT5=ON \
        -DBUILD_QT5_WEBKIT=OFF \
        -DBUILD_START=OFF \
        -DBUILD_WEB=OFF\
        -DPYTHON_EXECUTABLE=/usr/bin/python3 \
        -DPYTHON_LIBRARY=/usr/lib/libpython3.7m.so \
        -DPYTHON_INCLUDE_DIR=/usr/include/python3.7m \
        -DPYTHON_PACKAGES_PATH=/usr/lib/python3.7/site-packages

    make
}

@greyltc thanks! I've fixed the dependency issues mentioned below.

Yes, definitely freecad-python3-git should be redundant with this package. And I think it would be great if there was no trace of python2 here (if possible).

Great, thanks @lonaowna, you're a co-maintainer now!

@greyltc I could also help out. It might be also be time to look into merging with freecad-python3-git which seem to works pretty nice now (freecad has also switched to python3).

@zegentz, would you like to co-maintain this?

Please excuse my ignorance, but why the version part from the package name, i.e. freecad-python3-git-0.17.r2599.g87bf7094a-1-x86_64.pkg.tar.xz, differs with the version info provided by the "About Freecad" menu: (redacted) ... Version: 0.18.15455 (Git) ... Hash: ca7dd2e1c369dc23d2a79b3eb5ac6381dfd54990 ... It's just curiosity, maybe someone here knows... Thanks

Oh my, this was a !!FUN!! package to get to work.

Soooo, you guys are missing a couple dependencies when it comes to pyside and shiboken (and a couple others). I can't tell you which, because I ended up installing every AUR package with those two in it's name:

$ pacman -Q | grep "coin\|shiboken\|webkit\|webengine\|pyside\|qtpy"
coin 3.1.3-18
coin-docs 3.1.3-18
kdewebkit 5.56.0-1
pyqtwebengine-common 5.12-2
pyside-tools-common 0.2.15-3
pyside2 5.12.1-1
pyside2-tools 5.12.1-1
python-pyqtwebengine 5.12-2
python-pyside 1.2.4-9
python-pyside-common 1.2.4-9
python-pyside-tools 0.2.15-3
python-shiboken 1.2.4-4
python-shiboken2 5.12.1-1
python2-pyqtwebengine 5.12-2
python2-pyside 1.2.4-9
python2-pyside-tools 0.2.15-3
python2-pyside2 5.12.1-1
python2-qtpy 1.6.0-1
python2-shiboken 1.2.4-4
python2-shiboken2 5.12.1-1
qt5-webengine 5.12.1-3
qt5-webkit 5.212.0alpha2-23
qtwebkit-bin 2.3.4-9
shiboken 1.2.4-4
shiboken2 5.12.1-1
webkit2-sharp 2.10.9-2
webkit2gtk 2.22.7-1

Then that still didn't work, so I manually compiled the {python-,}shiboken2 packages in the extra/community repos, but with a slightly modified PKGBUILD to allow for support for llvm-svn:

Index: PKGBUILD
===================================================================
--- PKGBUILD    (revision 347468)
+++ PKGBUILD    (working copy)
@@ -4,7 +4,6 @@
 pkgbase=shiboken2
 pkgname=(shiboken2 python-shiboken2)
 _qtver=5.12.1
-_clangver=7.0.1
 pkgver=${_qtver/-/}
 pkgrel=1
 arch=(x86_64)
@@ -30,7 +29,7 @@
 }

 package_shiboken2() {
-  depends=(clang=$_clangver llvm libxslt qt5-xmlpatterns)
+  depends=(clang llvm libxslt qt5-xmlpatterns)

   cd build
   make DESTDIR="$pkgdir" install

And then it compiled, but the web browser didn't work! Gave an error along the line of WebGui module not found, or something like that. So I installed a couple packages which I thought might fix the issue (alas to no avail).

Then I just did this and it worked:

diff --git a/PKGBUILD b/PKGBUILD
index a8726ce..2c56996 100644
--- a/PKGBUILD
+++ b/PKGBUILD
@@ -2,7 +2,7 @@

 _appname=freecad
 pkgname="${_appname}-git"
-pkgver=0.17.r1484.g1e6818b6f
+pkgver=0.17.r2570.g6363c90a2
 pkgrel=1
 epoch=1
 pkgdesc='A general purpose 3D CAD modeler - git checkout'
@@ -38,9 +38,9 @@ build() {
          -DFREECAD_USE_OCC_VARIANT="Official Version" \
           -DFREECAD_USE_EXTERNAL_PIVY=ON \
           -DBUILD_QT5=ON \
-          -DBUILD_QT5_WEBKIT=OFF \
+          -DBUILD_QT5_WEBKIT=ON \
           -DPYTHON_EXECUTABLE=/usr/bin/python2 \
-          -DBUILD_WEB=OFF
+          -DBUILD_WEB=ON

     make
 }

And it worked! Luckly I don't have to care about deps, cause I installed everything with web{kit,gui,engine} in it's name.

The package depends on python2-netcdf4. Please add this to the list of dependencies.

Error:

Cloning freecad build files...
Checking freecad dependencies...
Cloning python2-pivy build files...
Checking python2-pivy dependencies...
Cloning python2-pyside2 build files...
Checking python2-pyside2 dependencies...
Cloning coin build files...
Checking coin dependencies...
Cloning python2-shiboken2 build files...
Checking python2-shiboken2 dependencies...
Cloning soqt build files...
Checking soqt dependencies...
Warning: cannot resolve "qt5-xmlpatterns>=5.12.1", a dependency of "python2-shiboken2"
Warning: cannot resolve "python2-shiboken2>=5.12.1", a dependency of "python2-pyside2"
Warning: cannot resolve "qt5-base>=5.12.1", a dependency of "python2-pyside2"
Warning: cannot resolve "python2-pyside2", a dependency of "freecad"
Warning: cannot resolve "qt5-xmlpatterns>=5.12.1", a dependency of "python2-shiboken2"
Warning: cannot resolve "python2-shiboken2", a dependency of "freecad"
Failed to prepare transaction.
Failed to prepare transaction:
could not satisfy dependencies:
unable to satisfy dependency 'qt5-xmlpatterns>=5.12.1' required by python2-shiboken2
unable to satisfy dependency 'python2-shiboken2>=5.12.1' required by python2-pyside2
unable to satisfy dependency 'qt5-base>=5.12.1' required by python2-pyside2
unable to satisfy dependency 'python2-pyside2' required by freecad
unable to satisfy dependency 'python2-shiboken2' required by freecad

Preparing...
Cloning freecad-python3-git build files...
Checking freecad-python3-git dependencies...
Cloning python-pivy-git build files...
Checking python-pivy-git dependencies...
Cloning coin-hg build files...
Checking coin-hg dependencies...
Resolving dependencies...
Checking inter-conflicts...
Downloading xerces-c (3.2.2-1-x86_64)...
Checking keyring...
Checking integrity...
Loading packages files...
Checking file conflicts...
Checking available disk space...
Installing xerces-c (3.2.2-1)...
Building coin-hg...
==> Making package: coin-hg r11844.cf2a467adeca-1 (lun 11 feb 2019 02:18:53 CET)
==> Checking runtime dependencies...
==> Checking buildtime dependencies...
==> Retrieving sources...
  -> Cloning coin hg repo...
applying clone bundle from https://api.media.atlassian.com/file/ca8ac694-8b46-45b5-affb-6bbba948b9cd/binary?client=64098d31-1676-4941-8a09-30b447252903&token=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJhY2Nlc3MiOnsidXJuOmZpbGVzdG9yZTpmaWxlOmNhOGFjNjk0LThiNDYtNDViNS1hZmZiLTZiYmJhOTQ4YjljZCI6WyJyZWFkIl19LCJleHAiOjE1NDk4NDg0OTIsImlzcyI6IjY0MDk4ZDMxLTE2NzYtNDk0MS04YTA5LTMwYjQ0NzI1MjkwMyIsIm5iZiI6MTU0OTg0ODA3Mn0._rx2QAazHxbUMvyweZuovZtKKI0HIUHEMKTzQYpVsi8
adding changesets
adding manifests
adding file changes
added 11707 changesets with 71756 changes to 6487 files (+1 heads)
finished applying clone bundle
searching for changes
adding changesets
adding manifests
adding file changes
added 260 changesets with 1364 changes to 911 files
new changesets de2e5a9fee0a:cbbeac5f7984
11707 local changesets published
  -> Cloning generalmsvcgeneration hg repo...
requesting all changes
adding changesets
adding manifests
adding file changes
added 3 changesets with 7 changes to 4 files
new changesets 17bc4850b39e:d12bf6cbb77c
  -> Cloning boost-header-libs-full hg repo...
applying clone bundle from https://api.media.atlassian.com/file/8cd0ceee-fcba-42a0-96ad-68c5b6c4067b/binary?client=64098d31-1676-4941-8a09-30b447252903&token=eyJhbGciOiJIUzI1NiIsInR5cCI6IkpXVCJ9.eyJhY2Nlc3MiOnsidXJuOmZpbGVzdG9yZTpmaWxlOjhjZDBjZWVlLWZjYmEtNDJhMC05NmFkLTY4YzViNmM0MDY3YiI6WyJyZWFkIl19LCJleHAiOjE1NDk4NDg2NjMsImlzcyI6IjY0MDk4ZDMxLTE2NzYtNDk0MS04YTA5LTMwYjQ0NzI1MjkwMyIsIm5iZiI6MTU0OTg0ODI0M30.M9szyU2jfG8ptEi6ujXpXiIUKrvE_ofUFQuWTp1y0pw
adding changesets
adding manifests
adding file changes
added 5 changesets with 17776 changes to 13201 files
finished applying clone bundle
searching for changes
no changes found
5 local changesets published
==> Validating source files with md5sums...
    coin ... Skipped
    generalmsvcgeneration ... Skipped
    boost-header-libs-full ... Skipped
==> Removing existing $srcdir/ directory...
==> Extracting sources...
  -> Creating working copy of coin hg repo...
updating to branch default
cloning subrepo build/general from /tmp/pamac-build/coin-hg/generalmsvcgeneration
cloning subrepo cpack.d from /tmp/pamac-build/coin-hg/cpack.d
abort: repository /tmp/pamac-build/coin-hg/cpack.d not found!
==> ERROR: Failure while creating working copy of coin hg repo
    Aborting...

The current PKGBUILD seems to be broken. I get: CMake Error: No source or binary directory provided. This seems to be triggered by the cmake command in the PKGBUILD missing a directory as argument (i.e. ".").

(On some systems I also have problems with cmake not being able to identify the correct opencascade include directory. I'm still trying to figure out why it happens only occasionally, and why only on some systems.)

Btw. I've removed the shiboken (shitbroken) dependency and it builds fine in clean chroot also Part and PartDesign workbenches seem to be functioning OK.

@miffi: Can you elaborate please? I just rebuild freecad and I get an error about shiboken2:

/opt/freecad-python3/bin/FreeCAD: error while loading shared libraries: libshiboken2.cpython-37m-x86_64-linux-gnu.so.5.12: cannot open shared object file: No such file or directory

which packages have you installed, have you rebuilt freecad? Thanks

My install appeared to not be happy about python-shiboken2 and pyside2 updating to 5.12. I downgraded the two packages and it's working fine for now.

It's much more concerning that qtwebkit is for qt4, as is python2-pyside, so this PKGBUILD was switched to use -DBUILD_QT5=ON without switching any of the dependencies.

Please make this depend on qt5-webkit instead of qtwebkit (like freecad-qt5-git does). The former is maintained and available from the official repositories.

opencascade7 is now opencascade. PKGBUILD needs update All aur packages related to freecad are broken ...

I ran into the same problem with the 'lptheads' linker directive.

Is there any difference between this and freecad-git?

It would be good to merge the py3 git aur package and make sure we're using only py3/qt5

Hi,I just tried to build the package but it fails with a cmake error trying to determine if pthread_create exists. Has anyone had or even solved this problem? Here's my CMakeErr.log:

Determining if the pthread_create exist failed with the following output:
Change Dir: /build/freecad-git/src/freecad-git/CMakeFiles/CMakeTmp

Run Build Command:"/usr/bin/make" "cmTC_d9d4f/fast"
/usr/bin/make -f CMakeFiles/cmTC_d9d4f.dir/build.make CMakeFiles/cmTC_d9d4f.dir/build
make[1]: Entering directory '/build/freecad-git/src/freecad-git/CMakeFiles/CMakeTmp'
Building C object CMakeFiles/cmTC_d9d4f.dir/CheckSymbolExists.c.o
/usr/bin/cc   -march=x86-64 -mtune=generic -O2 -pipe -fstack-protector-strong -fno-plt    -o CMakeFiles/cmTC_d9d4f.dir/CheckSymbolExists.c.o   -c /build/freecad-git/src/freecad-git/CMakeFile
s/CMakeTmp/CheckSymbolExists.c
Linking C executable cmTC_d9d4f
/usr/bin/cmake -E cmake_link_script CMakeFiles/cmTC_d9d4f.dir/link.txt --verbose=1
/usr/bin/cc -march=x86-64 -mtune=generic -O2 -pipe -fstack-protector-strong -fno-plt   -Wl,-O1,--sort-common,--as-needed,-z,relro,-z,now  -rdynamic CMakeFiles/cmTC_d9d4f.dir/CheckSymbolExist
s.c.o  -o cmTC_d9d4f 
/usr/bin/ld: CMakeFiles/cmTC_d9d4f.dir/CheckSymbolExists.c.o: in function `main':
CheckSymbolExists.c:(.text.startup+0x3): undefined reference to `pthread_create'
collect2: error: ld returned 1 exit status
make[1]: *** [CMakeFiles/cmTC_d9d4f.dir/build.make:87: cmTC_d9d4f] Error 1
make[1]: Leaving directory '/build/freecad-git/src/freecad-git/CMakeFiles/CMakeTmp'
make: *** [Makefile:121: cmTC_d9d4f/fast] Error 2

File /build/freecad-git/src/freecad-git/CMakeFiles/CMakeTmp/CheckSymbolExists.c:
/* */
#include <pthread.h>

int main(int argc, char** argv)
{
  (void)argv;
#ifndef pthread_create
  return ((int*)(&pthread_create))[argc];
#else
  (void)argc;
  return 0;
#endif
}

Determining if the function pthread_create exists in the pthreads failed with the following output:
Change Dir: /build/freecad-git/src/freecad-git/CMakeFiles/CMakeTmp

Run Build Command:"/usr/bin/make" "cmTC_b3bbb/fast"
/usr/bin/make -f CMakeFiles/cmTC_b3bbb.dir/build.make CMakeFiles/cmTC_b3bbb.dir/build
make[1]: Entering directory '/build/freecad-git/src/freecad-git/CMakeFiles/CMakeTmp'
Building C object CMakeFiles/cmTC_b3bbb.dir/CheckFunctionExists.c.o
/usr/bin/cc   -march=x86-64 -mtune=generic -O2 -pipe -fstack-protector-strong -fno-plt -DCHECK_FUNCTION_EXISTS=pthread_create   -o CMakeFiles/cmTC_b3bbb.dir/CheckFunctionExists.c.o   -c /usr/share/cmake-3.13/Modules/CheckFunctionExists.c
Linking C executable cmTC_b3bbb
/usr/bin/cmake -E cmake_link_script CMakeFiles/cmTC_b3bbb.dir/link.txt --verbose=1
/usr/bin/cc -march=x86-64 -mtune=generic -O2 -pipe -fstack-protector-strong -fno-plt -DCHECK_FUNCTION_EXISTS=pthread_create  -Wl,-O1,--sort-common,--as-needed,-z,relro,-z,now  -rdynamic CMakeFiles/cmTC_b3bbb.dir/CheckFunctionExists.c.o  -o cmTC_b3bbb -lpthreads 
/usr/bin/ld: cannot find -lpthreads
collect2: error: ld returned 1 exit status
make[1]: *** [CMakeFiles/cmTC_b3bbb.dir/build.make:87: cmTC_b3bbb] Error 1
make[1]: Leaving directory '/build/freecad-git/src/freecad-git/CMakeFiles/CMakeTmp'
make: *** [Makefile:121: cmTC_b3bbb/fast] Error 2

I think it might be caused by the "-lpthreads" ld-flag. As far as I know, it is a singular "-lpthread", or am I wrong here?

Hi @ILF. Originally I'd been using the 'opencascade7' AUR package and then replaced it with the 'oce' AUR package because it seemed to be better maintained then. At that time there was no 'opencascade' package in the standard repositories. It was added in September and I haven't noticed it till your post now.

I'll make a transition to it sometime next week. If you know how to build in clean chroot and do basic check for unnecessary dependencies I can add you as a maintainer and you can make the change sooner if you have time.

@p90 "pyside2-tools" is in the 'Extra' standard arch repo for a long time. Just click the link in the dependency listing on this page.

Hi,

I think the dependency "pyside2-tools" went the way of the Dodos or at least cannot be found in the aur anymore.

@Grawp: quick question, why not use opencascade instead of oce. I suppose there's a significant amount of people who use both kicad and freecad, and kicad-git package moved to opencascade, because freecad used it, but your package (which is probable the best freecad package in aur) still uses OCE. Can you switch it to opencascade?

@drhooves would you be interested in helping kickstart a FreeCAD-Assembly3-git port? Further discussion: https://github.com/ceremcem/build-freecad-asm3/issues/2

Python 3.7 fix is now committed in upstream FreeCAD so I have deleted the https://github.com/Grawp/FreeCAD.git repo.

For python3.7 y have change the source in PKGBUILD by: /// source=("${_appname}::git+https://github.com/Grawp/FreeCAD.git" ///

https://forum.freecadweb.org/viewtopic.php?t=12534&p=250293

thx Grawp !

@jorges I don't use FEM but you certainly need to add -DBUILD_FEM_NETGEN=1 to cmake arguments.

Is anyone using the FEM workbench? I installed netgen-git, a requirement for creating meshes, and rebuilt freecad-python3-git but it seems as it wasn't picked up. There's still the possibility to use an external tool for this, gmsh, which I am using at the moment, but I was curious about netgen

@drhooves Hi, do you plan updating this package? I've created and am maintaining a fairly well working freecad-python3-git. In case you don't wish to continue maintaining this packge maybe we could merge it with mine?

In case anyone is using this just for Qt5, the freecad package (https://aur.archlinux.org/packages/freecad/) just switched to using Qt5.

Maybe it needs salome-smesh package??

CMake Error: The following variables are used in this project, but they are set to NOTFOUND.
Please set them or make sure they are set and tested correctly in the CMake files:
NGLIB_INCLUDE_DIR
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
   used as include directory in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
NGLIB_LIBNGLIB
    linked by target "NETGENPlugin" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "MEFISTO2" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "SMESH" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "Driver" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "SMDS" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "StdMeshers" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "DriverSTL" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "DriverDAT" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "DriverUNV" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh
    linked by target "SMESHDS" in directory /tmp/yaourt-tmp-harvie/aur-freecad-git/src/freecad-git/src/3rdParty/salomesmesh

About my issue: FreeCAD was complaining about missing libshiboken2.cpython-36m-x86_64-linux-gnu.so.2.0. I have libshiboken2.cpython-36m-x86_64-linux-gnu.so.5.9 instead. A simple soft link has solved the problem.

@DanielH Good catch. I've added python-numpy as a non-optional dependency. Regarding the issue you've described, if I've understood it correctly, it seems that I'm not affected https://ibb.co/kVd3xH https://ibb.co/h8FixH (Tested on Arch with Plasma 5 and on Manjaro with Xfce4. In Arch I'm on revision 0.17_pre.r6798.g85b42a16d-1 ) Btw. I mostly use FreeCAD via python scripting via CadQuery.

@sl1pkn07 with [KDE-Unstable] you are on your own.

fail build for me

https://pastebin.com/5wcTT2JP

with Qt from [KDE-Unstable] 5.110beta2-2

I've installed freecad-python3-git, at present the only FC version in AUR that compiles without issues (thanks, Grawp!); hours and hours and hours compiling its dependencies and FC-p3-g itself.

But running FC some issues have arised...

1. FC is working, but in all Workbenches, under Two side Lighting, shape colours are displayed too much dark, despite colour or changes configured in Property panel > View or in View > Appeareance menu. Default shape colour RGB 204,204,204 is showed as a very dark grey, near black. Other selected palette colours are showed as the same but very dark colour. However, line colours can be changed without issue.

Setting Lighting = One side in Property view panel solves this problem. Are you experiencing the same issue?

1. If I'm not wrong, python-numpy seems to be a obligatory dependency for many (internal and external) modules. In fact, Test framework WB Self-test TestApp.All shows less failures (and 0 errors) installing python-numpy. I only see: [code] Traceback (most recent call last): File "/usr/lib/python3.6/unittest/case.py", line 59, in testPartExecutor yield File "/usr/lib/python3.6/unittest/case.py", line 605, in run testMethod() File "/opt/freecad-python3/Mod/Test/UnitTests.py", line 73, in testSchemes self.assertAlmostEqual(1, v.Value, msg="Failed with \"{0}\" scheme: {1} != 1".format(schemes[i], v.Value), delta=0.00001) File "/usr/lib/python3.6/unittest/case.py", line 878, in assertAlmostEqual raise self.failureException(msg) AssertionError: 1 != 0.9993118200000013 within 1e-05 delta : Failed with "Building Euro (cm/m²/m³)" scheme: 0.9993118200000013 != 1 [/code] that appears to be a conversion unit failure.

@Grawp: IMHO, adding python-numpy as [optional] dependency of FC 0.17 will help users experiency.

1. Assembly2 WB is not working at all in this version of FC 0.17: [code] invalid syntax (angleConstraint.py, line 18) Traceback (most recent call last): <br> File "<string>", line 6, in Initialize [/code]</string>

Built and installed it with trizen -S freecad-python3-git on Manjaro and it works, at least the Part Design workbench.

You have probably some old version of shiboken or other packages installed. Try rebuilding and reinstalling every AUR package that freecad-python3-git depends on (recursively).

@ElektroQuark Doing full rebuild in chroot on Arch. I also deployed Manjaro VM and am trying to build (albeit not in chroot) and install it there. We'll see shortly.

@ElektroQuark Manjaro is not Arch. This is a good reason to install Arch... :)

I can't run it. Terminal drops the error: "libshiboken2.cpython-36m-x86_64-linux-gnu.so.2.0"

I'n on Manjaro Linux.

@MicroJoe The message you posted only indicates that the build was aborted due to an error. Since the build might have continued for some time before it was aborted, you probably need to scroll up to see the error that caused this.

Current build status is ok on FreeCAD's page, but I have following error during build:

[ 81%] Linking CXX shared library ../../lin64/gcc/lib/libTKIGES.so<br> [ 81%] Built target TKIGES<br> make: *** [Makefile:130: all] Error 2<br> ==> ERROR: A failure occurred in build().<br> Aborting...

Adding netcdf as suggested by manuel_v did not solve the build issue.

netcdf is required as make dependency.

I've added python-pivy-git and coin-hg packages. Draft workbench should probably work now. Let me know if it doesn't.

Pivy for Python 3 https://github.com/FreeCAD/pivy is broken right now https://travis-ci.org/FreeCAD/pivy

When they get it in order, I'll create an AUR package and add it as a dependency here.

I see there is a python 3 pivy at https://github.com/looooo/pivy but I haven't been able to install it so far. I'm not sure what it requires. If someone else could get this working with this package that would be great.

@D_Schwartfeger Reproduced the problem. Looking into it now. I haven't tried Draft workbench before now, just PartDesign, Part and Mesh.

Changing to the Draft workbench gives the following in the Report View - "Error: Pivy not found, Draft workbench will be disabled."

Also -

import PyQt4 from PyQt4 import QtCore

causes a segfault on my system (Manjaro kernel 4.14.23-1, KDE Plasma 5.12.2, Qt 5.10.1)

Yes I am trying that as well. Thanks

@D_Schwartfeger you could try https://aur.archlinux.org/packages/freecad-python3-git I've uploaded it after all.

I see this is a problem with aur-opencascade7 and the AUR page has comments about how to fix this

On Manjaro Kde it fails with-

/tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:83:18: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'?

On Manjaro under KDE package didn't build with error-

/tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx: In member function 'virtual void IVtkVTK_ShapeData::InsertVertex(IVtk_IdType, IVtk_PointId, IVtk_MeshType)': /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:83:18: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? mySubShapeIDs->InsertNextTupleValue (&aShapeIDVTK); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:85:16: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? myMeshTypes->InsertNextTupleValue (&aType); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx: In member function 'virtual void IVtkVTK_ShapeData::InsertLine(IVtk_IdType, IVtk_PointId, IVtk_PointId, IVtk_MeshType)': /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:100:18: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? mySubShapeIDs->InsertNextTupleValue (&aShapeIDVTK); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:102:16: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? myMeshTypes->InsertNextTupleValue (&aType); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx: In member function 'virtual void IVtkVTK_ShapeData::InsertLine(IVtk_IdType, const IVtk_PointIdList, IVtk_MeshType)': /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:127:20: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? mySubShapeIDs->InsertNextTupleValue (&aShapeIDVTK); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:129:18: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? myMeshTypes->InsertNextTupleValue (&aType); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx: In member function 'virtual void IVtkVTK_ShapeData::InsertTriangle(IVtk_IdType, IVtk_PointId, IVtk_PointId, IVtk_PointId, IVtk_MeshType)': /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:146:18: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? mySubShapeIDs->InsertNextTupleValue (&aShapeIDVTK); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple /tmp/yaourt-tmp-donald/aur-opencascade7/src/occt-V7_2_0/src/IVtkVTK/IVtkVTK_ShapeData.cxx:148:16: error: 'class vtkIdTypeArray' has no member named 'InsertNextTupleValue'; did you mean 'InsertNextTuple'? myMeshTypes->InsertNextTupleValue (&aType); ^~~~~~~~~~~~~~~~~~~~ InsertNextTuple make[2]: [src/TKIVtk/CMakeFiles/TKIVtk.dir/build.make:327: src/TKIVtk/CMakeFiles/TKIVtk.dir/__/IVtkVTK/IVtkVTK_ShapeData.cxx.o] Error 1 make[1]: [CMakeFiles/Makefile2:1461: src/TKIVtk/CMakeFiles/TKIVtk.dir/all] Error 2 make: ** [Makefile:130: all] Error 2 ==> ERROR: A failure occurred in build(). Aborting... ==> ERROR: Makepkg was unable to build opencascade7. ==> Restart building opencascade7 ? [y/N]

netgen-nogui-occ7 can be compiled using occ 7.1.0, after that freecad-qt5-git compiles just fine.

Would anybody be interested in freecad git package with python3 and qt5? I have working PKGBUILD but I'm not sure whether it's worth uploading it to AUR to have yet another freecad package.

You can remove -DBUILD_FEM_NETGEN=1 from PKGBUILD or use the Appimage pre_17 from freecad release page. https://github.com/FreeCAD/FreeCAD/releases

The AppImage runs for me, but Freecad do not yet have an AppImage of the latest version (0.17).

Is there any way round this problem (netgen-nogui-occ7 not compiling)? It would be great to have 0.17...

It can't be installed as the needed version of netgen-nogui-occ7 will not compile as some headers are missing.