Create a functionnal package of SOFA Framework to a Conda environnement

[Namejess]

Create a functionnal package of SOFA Framework to a Conda environnement

• Overview & objectives
• Solutions & tools
• PoC

Overview

@hugtalbot gave me a mission : I have 2 weeks to purpose a PoC of a Sofa packaging to a Conda environment.

There are some difficulties for some people to implement Sofa in a Python ecosystem (Anaconda) and there is no functionnal package at the moment. My goal is to determine what it is possible to achieve, to understand the needs of the end-users, and with which tools.

Solutions & tools

After some discussions, there are several tools that can be used to make the packaging :

• Pypi
• Conda Forge / Build

Some differences against thses two tools :

Pypi

Pip is a package manager to install python packages from Pypi. Generally, you download source code that is installed according to instructions in setup.py, using setuptools.

Most appropriate for pure Python packages, and Pip depends on a python interpreter.

Conda

Packages are downloaded from Anaconda Cloud. With conda, we download and install a pre-compiled platform dependent binaries. Conda is a package manager that handles not only Python packages but also packages in other languages, and can create virtual environments, that do not need to contain Python.

In this case, Conda seems to be more appropriate, because of pre-compilated environments.

PoC

Here all ressources i used to build the solution :

• Conda Build Documentation
• Anaconda.org
• Anaconda Documentation
• SOFA Documentation
• Conda Forge Documentation
• Stack Overflow Thread

At first I was on a Windows system but after several builds attempts (which I will come back to) I decided to go on an Archlinux system, which solved many of my problems.

To create a Conda package, you have to install the whole Anaconda ecosystem. To use GUI packages with Linux, you will need to install the following extended dependencies for Qt:

Arch :

sudo pacman -Sy libxau libxi libxss libxtst libxcursor libxcomposite libxdamage libxfixes libxrandr libxrender mesa-libgl  alsa-lib libglvnd

Debian / Ubuntu :

sudo apt-get install libgl1-mesa-glx libegl1-mesa libxrandr2 libxrandr2 libxss1 libxcursor1 libxcomposite1 libasound2 libxi6 libxtst6

Download and launch the .sh installation (follow the Installation Guide if got a problem).

Then, install conda build :

conda install conda-build

To Build a conda package from scratch you have to create a file a root directory named meta.yaml. It is with this file that all the configuration of the package is done. Setup a directory like this :

git clone -b v22.12 https://github.com/sofa-framework/sofa.git sofa/src

Create a folder that will contain the meta.yaml file (recipe is a good practice name) :

mkdir recipe
cd recipe/
touch meta.yaml

The template by default looks like this :

package:
  name:
  version:

source:
  git_rev:
  git_url:

requirements:
  build:
    - python
    - setuptools

  run:
    - python

test:
  imports:
    -

about:
  home:

Here is the meta.yaml file I used after long trials and iterations (with commentary):

package:
  name: sofa
  version: 22.11.00

source:
  path : ..
requirements:
  build:
    - {{ compiler('cxx') }}
    - "cmake"
    - "ninja"
    - "clang"
    - setuptools
    - "boost >=1.65.1"
    - "python >=3.9"

  host:
    - "boost-cpp>=1.67"
    - "boost >=1.65.1"

build:
  number: 0

  script: |
    cmake -S . -B build \
    -DCMAKE_GENERATOR=Ninja \
    -DCMAKE_INSTALL_PREFIX=$PREFIX \
    -DCMAKE_CXX_COMPILER=clang++ \
    -DCMAKE_C_COMPILER=clang \

    cd $SRC_DIR/build 
    ninja

about:
  home: https://github.com/sofa-framework/sofa
  license: MIT
  license_file: LICENSE

extra:
  channels:
    - conda-forge
    - defaults

To launch build :

sudo conda build .

Errors

If you get this far, normally the build will break around 860/1970, with this error :

Possible Solutions

We tried to skip the script in meta.yaml and create a file named build.sh (in the same folder recipe). This script looks like this :

#!/bin/bash

  

# Define the path to the directory where the script will be created

SCRIPT_PATH="$CONDA_PREFIX/include/sofa/plugins/CMakeLists.txt"

  

# Define the content of the script

SCRIPT_CONTENT='cmake_minimum_required(VERSION 3.22)

  

# REQUIRED DEPENDENCIES

set(SOFA_BUILD_METIS "ON" CACHE BOOL "" FORCE)

set(PLUGIN_SOFAPYTHON "OFF" CACHE BOOL "" FORCE)

set(PLUGIN_SOFAPYTHON3 "ON" CACHE BOOL "" FORCE)

  

#EXTERNAL DIRECTORY'"'"'s plugins

list(APPEND CMAKE_MODULE_PATH "${CMAKE_CURRENT_LIST_DIR}/cmake")

find_package(SofaFramework REQUIRED)

sofa_add_plugins(STLIB SoftRobots ModelOrderReduction SofaPython3 BeamAdapter CosseratPlugin SoftRobotsInverse)

  

set(PLUGIN_MPC_PLUGIN "OFF" CACHE BOOL "" FORCE)

set(PLUGIN_STLIB "ON" CACHE BOOL "" FORCE)

set(PLUGIN_SOFTROBOTS "ON" CACHE BOOL "" FORCE)

set(PLUGIN_SOFTROBOTSINVERSE "ON" CACHE BOOL "" FORCE)

set(PLUGIN_COSSERATPLUGIN "ON" CACHE BOOL "" FORCE)

set(PLUGIN_BEAMADAPTER "ON" CACHE BOOL "" FORCE)

set(PLUGIN_MODELORDERREDUCTION "ON" CACHE BOOL "" FORCE)

set(PLUGIN_SOFAOFFSCREENCAMERA "OFF" CACHE BOOL "" FORCE)

'

# Create the script

mkdir -p "$CONDA_PREFIX/include/sofa/plugins/"

echo "$SCRIPT_CONTENT" > "$SCRIPT_PATH"

  

# Build SOFA

  

cmake -S $CONDA_PREFIX/include/sofa/plugins/ -B build \

-DCMAKE_GENERATOR=Ninja \

-DCMAKE_INSTALL_PREFIX="$PREFIX" \

-DCMAKE_CXX_COMPILER=clang++ \

-DCMAKE_C_COMPILER=clang \

-DPLUGIN_SOFAMATRIX=ON

  

cd $SCRIPT_PATH

ninja

We need you ! If you have any ideas, do not hesitate 👍

[alxbilger]

Thanks for sharing your experiments

[EtienneMenager]

Hi guys,

I also did some testing on the packaging via conda. This would allow to have the runSofa and the python interface directly accessible for users (for developers it is necessary to have access to the source code). I managed to get a version of runSofa to run, even though I'm currently having problems with Qt. Here is my configuration:

package:
  name: sofa
  version: 22.11.00

source:
  - git_url: https://github.com/sofa-framework/sofa.git
    folder: ./sofa/sofa/src
    git_tag: v22.12.00

requirements:
  build:
    - {{compiler('cxx')}}
    - python>=3.9
    - boost-cpp>=1.67
    - cmake
    - ninja
    - clang
    - setuptools
    - boost>=1.65.1
    - python>=3.7
    - qt>=5.12.9
    - Scipy>=1.2.0
    - numpy>=1.16.0
    - pybind11

  run:
    - python>=3.9
    - eigen>=3.2.10
    - qt>=5.12.9
    - qt-webengine>=5.15
    - qt-main>=5.15
    - python>=3.7
    - boost>=1.65.1
    - boost-cpp>=1.67

build:
  number: 0
  binary_relocation: false

about:
  home: https://github.com/sofa-framework/sofa
  license: MIT
  license_file: LICENSE.LGPL.txt

extra:
  channels:
    - conda-forge
    - defaults

and my build:

#!/bin/bash
CONDA_OVERRIDE_OSX=11.5.1
CXXFLAGS="${CXXFLAGS} -D_LIBCPP_DISABLE_AVAILABILITY"

mkdir ./sofa/sofa/build
cmake -S ./sofa/sofa/src -B ./sofa/sofa/build \
  -DCMAKE_GENERATOR=Ninja \
  -DCMAKE_INSTALL_PREFIX="$PREFIX" \
  -DCMAKE_CXX_COMPILER=clang++ \
  -DCMAKE_C_COMPILER=clang \
  -DPLUGIN_SOFAMATRIX=ON

cd ./sofa/sofa/build
ninja -j4
#ninja install

To improve the installation it would be necessary to do:

1. Fix the path to Qt.
2. Integrate the plugins and in particular SofaPython3.
3. Organize the elements at installation time (for the moment it is stored anywhere in the PREFIX_PATH).
4. Adapt the installation for cross platform.

Cheers,

Etienne

[jcarpent]

If not, you should remove any reference to python and numpy.

[jcarpent]

For QT, you can use the command: export Qt5_DIR=$PREFIX/lib/cmake/Qt5

[EtienneMenager]

Not yet. For the moment I only did some rapid modification to override the error in the Jessy tests. But I will add SofaPython3 and other plugins soon.

Thanks for the QT path!

[jcarpent]

My recommendation is also to not force the use of CLANG and let conda use its own internal compiler adapted to the OS.

[hugtalbot]

wow, @jcarpent is watching 👓
👍 👍 👍

thanks a lot for all these feedback!

[Namejess]

Thanks everyone for feedbacks, I'll try these this week 👍

[EtienneMenager]

Hi guys,

I've made some progress on this subject during these few days but I'm now stuck with a question that is outside my knowledge. Maybe someone will have an idea to solve it.

The idea is to create two tools:

1. An installation script. This is a script that allows to install Sofa from source in a conda environment, by updating all environment variables, paths to libraries, etc. This allows to install Sofa easily and quickly in a conda environment by having access to its source code to be able to modify it.
2. A conda package allowing to use Sofa and SofaPython3 in a conda environment. Here we don't have access to the source code but it is possible to use the tools of Sofa and SofaPython3 to create new scenes, to use easily SofaGym, etc.

For point 1, I propose something like this:

#!/bin/bash
CONDA_OVERRIDE_OSX=11.5.1

#The idea here is to build sofa against a Conda environment.
#Don't forget to activate the conda environment before to launch this
#script

mamba install 'qt>=5.12.0' 'boost>=1.65.1' glew 'eigen>=3.2.10' 'pybind11>=2.3' conda-build

export CMAKE_PREFIX_PATH=$CONDA_PREFIX/pkgs
export VERSION=22.12


mkdir -p $CMAKE_PREFIX_PATH/dev_sofa/sofa/build
mkdir -p $CMAKE_PREFIX_PATH/dev_sofa/plugins

git clone -b v$VERSION https://github.com/sofa-framework/sofa.git $CMAKE_PREFIX_PATH/dev_sofa/sofa/src 2> /dev/null || (git -C $CMAKE_PREFIX_PATH/dev_sofa/sofa/src pull)
git clone -b v$VERSION https://github.com/SofaDefrost/SoftRobots.git $CMAKE_PREFIX_PATH/dev_sofa/plugins/SoftRobots 2> /dev/null || (git -C $CMAKE_PREFIX_PATH/dev_sofa/plugins/SoftRobots pull)
git clone -b v$VERSION https://github.com/SofaDefrost/plugin.Cosserat.git $CMAKE_PREFIX_PATH/dev_sofa/plugins/plugin.Cosserat 2> /dev/null || (git -C $CMAKE_PREFIX_PATH/dev_sofa/plugins/plugin.Cosserat pull)
git clone -b v$VERSION https://github.com/sofa-framework/BeamAdapter.git $CMAKE_PREFIX_PATH/dev_sofa/plugins/BeamAdapter 2> /dev/null || (git -C $CMAKE_PREFIX_PATH/dev_sofa/plugins/BeamAdapter pull)
git clone -b v$VERSION https://github.com/SofaDefrost/STLIB.git $CMAKE_PREFIX_PATH/dev_sofa/plugins/STLIB 2> /dev/null || (git -C $CMAKE_PREFIX_PATH/dev_sofa/plugins/STLIB pull)
git clone -b v$VERSION https://github.com/SofaDefrost/ModelOrderReduction.git $CMAKE_PREFIX_PATH/dev_sofa/plugins/ModelOrderReduction 2> /dev/null || (git -C $CMAKE_PREFIX_PATH/dev_sofa/plugins/ModelOrderReduction pull)
git clone -b v$VERSION https://github.com/sofa-framework/SofaPython3.git $CMAKE_PREFIX_PATH/dev_sofa/plugins/SofaPython3 2> /dev/null || (git -C $CMAKE_PREFIX_PATH/dev_sofa/plugins/SofaPython3 pull)

touch $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt
echo 'set(CMAKE_BUILD_TYPE RelWithDebInfo)' > $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt
#EXTERNAL DIRECTORY's plugins
echo 'sofa_add_subdirectory(plugin STLIB STLIB)' >> $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt
echo 'sofa_add_subdirectory(plugin SoftRobots SoftRobots)' >> $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt
echo 'sofa_add_subdirectory(plugin ModelOrderReduction ModelOrderReduction)' >> $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt
echo 'sofa_add_subdirectory(plugin SofaPython3 SofaPython3)' >> $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt
echo 'sofa_add_subdirectory(plugin BeamAdapter BeamAdapter)' >> $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt
#echo 'sofa_add_subdirectory(plugin plugin.Cosserat CosseratPlugin)' >> $CMAKE_PREFIX_PATH/dev_sofa/plugins/CMakeLists.txt


export QT5PATH="$(find $CONDA_PREFIX -name Qt5)"
export EIGENPATH="$(find $CONDA_PREFIX -name signature_of_eigen3_matrix_library)"
export EIGENPATH=$(dirname $EIGENPATH)
export SITEPACKAGESPATH="$(find $CONDA_PREFIX/lib -name site-packages)"
export PATH=$CONDA_PREFIX:$PATH:$SITEPACKAGESPATH

cmake -S $CMAKE_PREFIX_PATH/dev_sofa/sofa/src -B $CMAKE_PREFIX_PATH/dev_sofa/sofa/build \
  -DCMAKE_GENERATOR=Ninja \
  -DCMAKE_INSTALL_PREFIX=$CMAKE_PREFIX_PATH \
  -DCMAKE_CXX_COMPILER=clang++ \
  -DCMAKE_C_COMPILER=clang \
  -DEIGEN3_INCLUDE_DIR=$EIGENPATH \
  -DSOFA_EXTERNAL_DIRECTORIES=$CMAKE_PREFIX_PATH/dev_sofa/plugins \
  -DCMAKE_PREFIX_PATH="$QT5PATH;$CMAKE_PREFIX_PATH;$CONDA_PREFIX/bin" \
  -DPython_EXECUTABLE=$CONDA_PREFIX/bin/python \
  -DPython_ROOT_DIR=$CONDA_PREFIX/bin \
  -DSP3_PYTHON_PACKAGES_LINK_DIRECTORY=$SITEPACKAGESPATH \
  -DPython_FIND_STRATEGY=LOCATION \
  -DPython_FIND_FRAMEWORK=NEVER \
  -DPLUGIN_SOFAPYTHON3=ON \
  -DPLUGIN_BEAMADAPTER=ON \
  -DPLUGIN_STLIB=ON \
  -DPLUGIN_SOFTROBOTS=ON \
  -DPLUGIN_MODELORDERREDUCTION=ON
#-DPLUGIN_COSSERATPLUGIN=ON

cd $CMAKE_PREFIX_PATH/dev_sofa/sofa/build
ninja -j4

#Set path to SOFAROOT and SOFAPATH - add alias to runSofa - set PYTHONPATH
mkdir -p $CONDA_PREFIX/etc/conda/activate.d
mkdir -p $CONDA_PREFIX/etc/conda/deactivate.d

touch $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
touch $CONDA_PREFIX/etc/conda/deactivate.d/env_vars.sh

echo '#!/bin/sh' > $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
echo 'alias my_runSofa='$CMAKE_PREFIX_PATH'/dev_sofa/sofa/build/bin/runSofa' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
echo 'export SOFA_ROOT='$CMAKE_PREFIX_PATH'/dev_sofa/sofa/build' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
echo 'export SOFA_PATH='$CMAKE_PREFIX_PATH'/dev_sofa/sofa/build/bin' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh
echo 'unset PYTHONPATH' >> $CONDA_PREFIX/etc/conda/activate.d/env_vars.sh

echo '#!/bin/sh' > $CONDA_PREFIX/etc/conda/deactivate.d/env_vars.sh
echo 'unalias my_runSofa' >> $CONDA_PREFIX/etc/conda/deactivate.d/env_vars.sh
echo 'unset SOFA_ROOT' >> $CONDA_PREFIX/etc/conda/deactivate.d/env_vars.sh
echo 'unset SOFA_PATH' >> $CONDA_PREFIX/etc/conda/deactivate.d/env_vars.sh

mamba develop $CMAKE_PREFIX_PATH/dev_sofa/plugins/STLIB/python3/src
mamba develop $CMAKE_PREFIX_PATH/dev_sofa/sofa/build/lib/python3/site-packages/

I tested this script on my computer (MacOs). It allows to compile Sofa with the different plugins by searching the different libraries (qt, Eigen, python, ...) in the conda environment. It also allows to create an alias to the executable and to update the environment variables automatically. After running this script, runSofa runs perfectly and it is possible to load any scene written in python. However there is a problem when I run the python interpreter:

The PYTHONPATH is correctly set to the installed sofapython3 site-packages, and the files are well found. But when the init.py of Sofa calls the "import Sofa.Helper", we get a segmentation error (from the .so file handling this import). I don't know how to fix this problem.

For point 2, I want to use what I did for point 1 to generate Sofa+Sofapython3. However for the moment, the $Python_EXECUTABLE variable I use in my script does not point to a recognized python executable during the conda build. Maybe here someone will have a solution too.

Cheers,

Etienne

[hugtalbot]

Thanks a lot for sharing @EtienneMenager
That's already a lot 👏

@damienmarchal do you see any explanation on the SP3 side?

[damienmarchal]

I really know nothing about conda... but to investigate the failure of the loading of binary python module my first check would be to be sure that the dynamic loading of a shared library (because the .so is such a thing) is not messed with other system-wide installed library. So I would first be sure there is not doubt on who is loading what.

From your build directory for sofa, can you try something like:
ldd lib/python3/site-packages/Sofa/Helper.cpython-310-x86_64-linux-gnu.so

On my system it returns:

 linux-vdso.so.1 (0x00007ffeb79ec000)
	libSofaPython3.so.1.0 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofaPython3.so.1.0 (0x00007faa9f98c000)
	libpython3.10.so.1.0 => /lib/x86_64-linux-gnu/libpython3.10.so.1.0 (0x00007faa9f200000)
	libSofa.Core.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.Core.so.22.12.99 (0x00007faa9e600000)
	libSofa.Helper.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.Helper.so.22.12.99 (0x00007faa9f04c000)
	libstdc++.so.6 => /lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007faa9e200000)
	libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007faa9f948000)
	libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007faa9de00000)
	libSofa.Testing.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.Testing.so.22.12.99 (0x00007faa9f936000)
	libSofa.Simulation.Graph.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.Simulation.Graph.so.22.12.99 (0x00007faa9f8f3000)
	libSofa.Simulation.Core.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.Simulation.Core.so.22.12.99 (0x00007faa9da00000)
	libSofa.DefaultType.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.DefaultType.so.22.12.99 (0x00007faa9f8cd000)
	libSofa.Type.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.Type.so.22.12.99 (0x00007faa9f86f000)
	libgtest.so.2.6.2 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libgtest.so.2.6.2 (0x00007faa9f805000)
	libexpat.so.1 => /lib/x86_64-linux-gnu/libexpat.so.1 (0x00007faa9f01b000)
	libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007faa9f7e7000)
	libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007faa9e519000)
	libSofa.LinearAlgebra.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.LinearAlgebra.so.22.12.99 (0x00007faa9d600000)
	/lib64/ld-linux-x86-64.so.2 (0x00007faa9fb85000)
	libSofa.Simulation.Common.so.22.12.99 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libSofa.Simulation.Common.so.22.12.99 (0x00007faa9e4e2000)
	libtinyxml.so.2.6.2 => /home/dmarchal/projects/DEFROST/dev/sofa1/builds/gcc/lib/libtinyxml.so.2.6.2 (0x00007faa9e4c9000)

So I can check that all loaded libraries are the one compiled at the same time as Helper.cpython-310-x86_64-linux-gnu.so

[EtienneMenager]

Hi @damienmarchal ,

Thanks for your answer. I tried what you suggested and I obtained:

>>> otool -L Helper.cpython-310-darwin.so
Helper.cpython-310-darwin.so:
	@rpath/Helper.cpython-310-darwin.so (compatibility version 0.0.0, current version 0.0.0)
	@rpath/libSofaPython3.1.0.dylib (compatibility version 0.0.0, current version 1.0.0)
	@rpath/libpython3.10.dylib (compatibility version 3.10.0, current version 3.10.0)
	@rpath/libSofa.Testing.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Simulation.Graph.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Simulation.Common.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Simulation.Core.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Core.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.DefaultType.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.LinearAlgebra.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Helper.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Topology.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Geometry.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Type.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Compat.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Component.Compat.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libSofa.Config.22.12.00.dylib (compatibility version 0.0.0, current version 22.12.0)
	@rpath/libgtest.2.6.2.dylib (compatibility version 0.0.0, current version 2.6.2)
	/usr/lib/libc++.1.dylib (compatibility version 1.0.0, current version 1200.3.0)
	/usr/lib/libSystem.B.dylib (compatibility version 1.0.0, current version 1311.0.0)

Seems that I have a problem with my rpath. Have you ever seen this before?

Etienne

[hugtalbot]

As an update, the PR #23085 opened on the conda forge!

🔥 🔥 🔥

👏 @olivier-roussel 👏

[github-actions[bot]]

⚠️ ⚠️ ⚠️
@Namejess
We are setting a new forum management policy: topics like this one, in which the last reply is more than 4 months old, will automatically be closed. Without further news in the coming weeks, our bot will proceed to automatic closure, thus keeping this forum clean and fresh 🌱 Thank you for your understanding

[github-actions[bot]]

⚠️ ⚠️ ⚠️
@Namejess
In accordance with our forum management policy, the last reply is more than 4 months old and is therefore closed. Our objective is to keep the forum up to date and offer the best support experience.

Please feel free to reopen it if the topic is still active by providing us with an update. Please feel free to open a new thread at any time - we'll be happy to help where and when we can.

[github-actions[bot]]

⚠️ ⚠️ ⚠️
@Namejess
In accordance with our forum management policy, the last reply is more than 4 months old and is therefore closed. Our objective is to keep the forum up to date and offer the best support experience.

Please feel free to reopen it if the topic is still active by providing us with an update. Please feel free to open a new thread at any time - we'll be happy to help where and when we can.