Undefined symbol error in libpinocchio_pywrap.so
Closed this issue · 6 comments
Hello,
Whenever I source my devel/setup.bash, I can't import Pinocchio. I get the following error
In [1]: import pinocchio
---------------------------------------------------------------------------
ImportError Traceback (most recent call last)
<ipython-input-1-d3c451d02715> in <module>()
----> 1 import pinocchio
/opt/openrobots/lib/python2.7/site-packages/pinocchio/__init__.py in <module>()
4
5 import numpy
----> 6 from .libpinocchio_pywrap import *
7
8 from .libpinocchio_pywrap import __version__, __raw_version__
ImportError: /opt/openrobots/lib/python2.7/site-packages/pinocchio/libpinocchio_pywrap.so: undefined symbol: _ZN7eigenpy9NumpyType7getTypeEv
However I can import Pinocchio sucessfully when I don't source my workspace, which is confusing... Besides, I have exactly the same problem with Crocoddyl. I'm using Ubuntu 18.04 and python 2.7.17, and here's my robotpkg config
ii robotpkg-ddp-actuator-solver 1.2.0 amd64 Differentiable Dynamics Programming for one actuator
ii robotpkg-dynamic-graph-v3 4.1.0 amd64 Efficient data-flow C++ library for robotics
ii robotpkg-eigen-quadprog 1.0.0r1 amd64 QuadProg QP solver using the Eigen3 library
ii robotpkg-eiquadprog 1.2.0 amd64 algorithm of Goldfarb and Idnani for the solution of a QP
ii robotpkg-example-robot-data 3.3.0 amd64 Robot URDFs for benchmarking and developed examples
ii robotpkg-hpp-fcl 1.4.4 amd64 Humanoid Path Planner (collision detection and distance computations)
ii robotpkg-libccd 2.0 amd64 collision detection
ii robotpkg-octomap 1.9.0 amd64 An Efficient Probabilistic 3D Mapping Framework Based on Octrees
ii robotpkg-parametric-curves 1.3.1 amd64 This packages provides a template based parametric curves library.
ii robotpkg-pinocchio 2.4.5 amd64 Efficient rigid body dynamics
ii robotpkg-py27-crocoddyl 1.3.0 amd64 Optimal control library for robot under contact sequence
ii robotpkg-py27-dynamic-graph-v3 3.5.2 amd64 Efficient data-flow C++ library for robotics (python bindings)
ii robotpkg-py27-eigenpy 2.4.0 amd64 Eigen bindings to python
ii robotpkg-py27-example-robot-data 3.3.0 amd64 Robot URDFs for benchmarking and developed examples
ii robotpkg-py27-hpp-fcl 1.4.4r1 amd64 Humanoid Path Planner (collision detection and distance computations (python bindingss))
ii robotpkg-py27-parametric-curves 1.3.1 amd64 This packages provides a template based parametric curves library (python bindings)
ii robotpkg-py27-pinocchio 2.4.5 amd64 Efficient rigid body dynamics (python bindings)
ii robotpkg-py27-qt4-gepetto-viewer+doc 4.9.0 amd64 Humanoid Path Planner (Simple viewer based on OpenSceneGraph)
ii robotpkg-py27-qt4-gepetto-viewer-corba 5.4.0 amd64 Humanoid Path Planner (Corba server for gepetto-viewer)
ii robotpkg-py27-quadprog 0.1.6r1 amd64 Solve a strictly convex quadratic program
ii robotpkg-py27-sot-core-v3 4.9.0 amd64 Hierarchical task based control of humanoid robots (python bindings)
ii robotpkg-py27-sot-dynamic-pinocchio-v3 3.5.1 amd64 Create entities to compute a robot dynamics properties using pinocchio (python bindings)
ii robotpkg-py27-sot-tools-v3 2.2.2 amd64 tools for the sot-core package (python bindings)
ii robotpkg-py27-sot-torque-control 1.5.2 amd64 Torque control for humanoids (python bindings)
ii robotpkg-py27-tsid 1.3.0 amd64 Efficient Inverse Dynamics (python bindings)
ii robotpkg-qpoases+doc 3.2.1r1 amd64 C++ implementation of the online active set strategy
ii robotpkg-simple-humanoid-description 1.0.2 amd64 Rigid body model of a simple humanoid robot.
ii robotpkg-sot-core-v3 4.9.0 amd64 Hierarchical task based control of humanoid robots
ii robotpkg-sot-dynamic-pinocchio-v3 3.5.1 amd64 Create entities to compute a robot dynamics properties using pinocchio
ii robotpkg-sot-tools-v3 2.2.2 amd64 tools for the sot-core package
ii robotpkg-sot-torque-control 1.5.2 amd64 Torque control for humanoids
ii robotpkg-tsid 1.3.0 amd64 Efficient Inverse Dynamics based on Pinocchio
I checked echo $LD_LIBRARY_PATH, which returns
/home/sebastien/devel_test/workspace/devel/lib:/opt/ros/melodic/lib:/opt/openrobots/lib:/opt/openrobots/lib/dynamic-graph-plugins
And the dependencies with ldd /opt/openrobots/lib/python2.7/site-packages/pinocchio/libpinocchio_pywrap.so, which returns
linux-vdso.so.1 (0x00007fff5a967000)
libeigenpy.so => /opt/ros/melodic/lib/libeigenpy.so (0x00007f51d86fb000)
libboost_python-py27.so.1.65.1 => /usr/lib/x86_64-linux-gnu/libboost_python-py27.so.1.65.1 (0x00007f51d84bb000)
libpinocchio.so.2.4.5 => /opt/openrobots/lib/libpinocchio.so.2.4.5 (0x00007f51d828c000)
libhpp-fcl.so => /opt/openrobots/lib/libhpp-fcl.so (0x00007f51d7e41000)
libboost_system.so.1.65.1 => /usr/lib/x86_64-linux-gnu/libboost_system.so.1.65.1 (0x00007f51d7c3c000)
libboost_serialization.so.1.65.1 => /usr/lib/x86_64-linux-gnu/libboost_serialization.so.1.65.1 (0x00007f51d79fe000)
liburdfdom_world.so.1.0 => /usr/lib/x86_64-linux-gnu/liburdfdom_world.so.1.0 (0x00007f51d77dc000)
libstdc++.so.6 => /usr/lib/x86_64-linux-gnu/libstdc++.so.6 (0x00007f51d7453000)
libm.so.6 => /lib/x86_64-linux-gnu/libm.so.6 (0x00007f51d70b5000)
libgcc_s.so.1 => /lib/x86_64-linux-gnu/libgcc_s.so.1 (0x00007f51d6e9d000)
libc.so.6 => /lib/x86_64-linux-gnu/libc.so.6 (0x00007f51d6aac000)
libpthread.so.0 => /lib/x86_64-linux-gnu/libpthread.so.0 (0x00007f51d688d000)
libboost_filesystem.so.1.65.1 => /usr/lib/x86_64-linux-gnu/libboost_filesystem.so.1.65.1 (0x00007f51d6673000)
liburdfdom_model.so.1.0 => /usr/lib/x86_64-linux-gnu/liburdfdom_model.so.1.0 (0x00007f51d6451000)
libassimp.so.4 => /usr/lib/x86_64-linux-gnu/libassimp.so.4 (0x00007f51d5a86000)
liboctomap.so.1.9 => /opt/ros/melodic/lib/liboctomap.so.1.9 (0x00007f51d5840000)
liboctomath.so.1.9 => /opt/ros/melodic/lib/liboctomath.so.1.9 (0x00007f51d5639000)
libtinyxml.so.2.6.2 => /usr/lib/x86_64-linux-gnu/libtinyxml.so.2.6.2 (0x00007f51d5424000)
libconsole_bridge.so.0.4 => /usr/lib/x86_64-linux-gnu/libconsole_bridge.so.0.4 (0x00007f51d521f000)
/lib64/ld-linux-x86-64.so.2 (0x00007f51d9fea000)
libz.so.1 => /lib/x86_64-linux-gnu/libz.so.1 (0x00007f51d5002000)
libminizip.so.1 => /usr/lib/x86_64-linux-gnu/libminizip.so.1 (0x00007f51d4df7000)
librt.so.1 => /lib/x86_64-linux-gnu/librt.so.1 (0x00007f51d4bef000)
But I am still unable to understand where this conflict might come from. Do you know how to solve this issue ?
You have a conflict between Robotpkg and ROS.
ibeigenpy.so => /opt/ros/melodic/lib/libeigenpy.so (0x00007f51d86fb000)
is pointing to Melodic. While you are using Pinocchio from robotpkg. Please choose one of them by adjusting your environment variables.
Thank you , now it works !
libeigenpy.so => /opt/openrobots/lib/libeigenpy.so (0x00007fcaa2694000)
Hello.
I think I am having the same problem as described here, but still don't understand how to resolve the problem. I have installed pinocchio by following the steps described here: https://stack-of-tasks.github.io/pinocchio/download.html, including saving the path variables at the end of my bashrc config. As I understand it, when I source my devel/setup.bash, it makes changes / adds further alternative paths to some of these variables, but I don't know which variable affects what, exactly, and how to get the system to find this particular package in its correct place without breaking others.
Hi @akosv95, what operating system are you on? And is it possible that you have multiple versions installed?
- Can you confirm if you have pinocchio/eigenpy from robotpkg?
- Do you happen to have a ROS version of eigenpy or pinocchio installed? I.e. is "ros-$ROS_DISTRO-eigenpy" installed?
You can circumvent this by adding the devel/setup.bash source before the robotpkg PATH modifications. Note that if you use MoveIt that also uses eigenpy from ROS it could break MoveIt (the robotpkg and ROS versions of eigenpy may differ)
I am using Ubuntu 18.04 and ROS Melodic.
I have the following relevant packages installed:
robotpkg-pinocchio/bionic,now 2.4.6 amd64 [installed,automatic]
robotpkg-py27-eigenpy/bionic,now 2.4.0 amd64 [installed,automatic]
ros-melodic-eigenpy/bionic,now 2.3.1-1bionic.20200410.133723 amd64 [installed,automatic]
No ROS version of pinocchio as far as I can tell.
/edit: Okay I tried sourcing the setup file and then applying the robotpkg paths and it seems to work now. Thank you very much!