PCL version

[an99990]

Hi, I am trying to build rtabmap and rtabmap_ros inside a Docker and am having PCL version conflicts. Is there a PCL installation insde one of these packages ? Is there a way to control the installed PCL version ?

# create conda env to install ros
RUN conda install mamba -c conda-forge
SHELL ["conda", "run", "-n", "base", "/bin/bash", "-c"]
RUN conda  config --env --add channels conda-forge
# and the robostack channels
RUN conda  config --env --add channels robostack
RUN conda  config --env --add channels robostack-experimental

RUN mamba install ros-galactic-desktop
RUN mamba install compilers cmake pkg-config make ninja colcon-common-extensions ros-galactic-rosbridge-suite ros-galactic-rmw-fastrtps-cpp ros-galactic-octomap-msgs ros-galactic-octomap
RUN conda install dynaconf mesalib mesa-libgl-devel-cos6-x86_64 pyopengl ros-galactic-navigation2

thank you

[Tobias-Fischer]

Hi, I would strongly recommend opening a PR here that adds rtabmap and rtabmap_ros to our build pipeline (it can be as simple as adding one line, see https://github.com/RoboStack/ros-galactic/pull/94/files). Building it there will be much easier than inside docker, in particular for the case of rtabmap (there are a few patches needed for these packages, see https://github.com/RoboStack/ros-galactic/blob/main/patch/ros-galactic-rtabmap.patch).

PCL is pinned and needs to be installed via mamba: https://github.com/RoboStack/ros-galactic/blob/dc6ae958429267cf93ebd05393c8c4813807c81c/conda_build_config.yaml#L14-L15

I also see that you still don't create a new environment for RoboStack - please follow our readme, you will run into issues if you don't create a new environment that won't be easy to debug.

[Tobias-Fischer]

Closing due to inactivity - feel free to reopen if needed