lmontaut / collision-detection-benchmark Goto Github PK

I come from the RSS 2022 paper, and I have been trying to run the benchmark on a freshly installed Ubuntu20.04. I am strictly following the installation guidelines. However, during the cmake stage, the following dependencies are missing:

• CMake can't find eigenpy, which I fixed by running conda install eigenpy -c conda-forge.
• CMake can't find qhull, which I fixed by installing qhull from source using CMake (see https://github.com/qhull/qhull). Note that sudo apt install libqhull-dev won't work, because the package does not contain CMakeConfig files (See https://packages.ubuntu.com/focal/amd64/libqhull-dev/filelist). We might be able to fix the apt installed problem by providing a FindQhull.cmake, but that's another story.

I hope the missing dependencies mentioned above could be added readme.

After successfully building and installing the code (on commit 2546b89) on a freshly installed Ubuntu20.04, I tried to run the benchmark using the following command:

However, I got the following error:

It seems that GJKVariant enum type was not imported to the hppfcl namespace. So I tried to import everything from hppfcl using the following command in the Python console:
from hppfcl import *
Then I printed out all the things in the Python environment using the following command:
locals()
Here is the result:

It seems that even though GJKVariant was binded in https://github.com/lmontaut/hpp-fcl/blob/340fe6a16b547901c260008824dc65ad6e96ab73/python/gjk.cc#L93, it was not imported somehow.

I am not very familiar with boost Python, so I am asking for help here.

Hi, thanks for open-sourcing the benchmark. After successfully installing all dependencies, I started to build the repo using CMake. The commit is 2546b89. CMake failed and its error message points to Line 63 in hpp/fcl/include/hpp/fcl/internal/intersect.h.
Error gave by the g++ complier was:
error: ‘Vector’ in namespace ‘Eigen’ does not name a template type; did you mean ‘VectorXd’?
My fix was to use Eigen::Matrix<double, 4, 1> instead of the Eigen::Vector<double ,4>.
My Eigen version is 3.3.7
After the modification, the code complied successfully.

I am writing this issue to make sure my modification is correct.