Compiling PCL from source using Docker
This tutorial explains how to build the Point Cloud Library from source using docker.
The walkthrough was done using Ubuntu as hosting OS of docker. The reason is that docker can be much easier set up in linux OSs compared to Windows. Other possible alternatives in case your main host is Windows could be WSL and or using virtual machine where some linux OS is installed.
Advantages - Disadvantages
Selecting to use docker to build PCL from source offers the following benefits: * Docker container provides some sort of isolated development environment. For someone familiar to python it is quite similar concept to virtual environment. * There is no need to install pcl dependencies standalone. * Installing, updating and maintaining different compilers (clang, g++) or version of other related programs is easier in docker container. * Once setting up docker the setup is pretty stable and there is no need to spend time for troubleshooting issues. Therefore the focus can be only in programming.
Only disadvantage that i would think is the need to have a basic knowledge of linux OS and commands since it is much easier to setup docker in linux OSs compared to Windows.
Open a terminal in Ubuntu and run the corresponding commands from each installation section
Check if curl is already installed:
$ curl --version
If it is not already installed, run in terminal the relative command for your OS:: https://www.tecmint.com/install-curl-in-linux
Check if git is already installed:
$ git --version
If it is not already installed, run in terminal the relative command for your OS:: https://git-scm.com/download/linux
Check if docker is already installed:
$ docker --version
If it is not already installed, follow the instructions from https://github.com/docker/docker-install and run in terminal:
$ curl -fsSL https://get.docker.com -o get-docker.sh $ sh get-docker.sh
Other useful commands are:
$ docker ps $ service docker status
The first one shows the running containers while the latter shows the docker status. If everything is fine it will be active (running). You can start/stop docker if needed by running:
$ service docker start/stop
It might be required to add sudo in docker commands if permissions are not set properly. See part run docker commands without sudo on how to set them correctly so the sudo command is not required.
Downloading PCL source code
Download the pcl source code in Ubuntu:
$ git clone https://github.com/PointCloudLibrary/pcl.git
Docker container configuration
To run docker commands without sudo:
$ sudo groupadd docker $ sudo usermod -aG docker $USER $ newgrp docker
Verify you can run docker without sudo:
$ docker run hello-world
Pull the docker image by running:
$ docker pull pointcloudlibrary/env:20.04
The docker image above will have OS Ubuntu 20.04. Other possible available images can be found under ‘https://hub.docker.com/r/pointcloudlibrary/env/tags’
It is also possible to use the Dockerfile under .dev folder to set up your docker image. The method of pulling the official docker image is considered more stable option though.
Run the container:
$ docker run --user $(id -u):$(id -g) -v $PWD/pcl:/home --rm -it pointcloudlibrary/env:20.04 bash
where $PWD:/pcl:/home represents the pcl source code in Ubuntu while home represents the pcl source code inside the docker container. Option –rm tells docker to remove the container when it exits automatically. By default, when the container exits, its file system persists on the host system. The -it option is used when dealing with the interactive processes like bash and tells Docker to keep the standard input attached to the terminal.
Using volumes, actions performed on a file in Ubuntu such as creating new files are directly mapped to the selected path location inside docker container.
To exit the container simply run in terminal exit.
After running the container, we need to navigate to pcl source code and create a build folder in that directory.
$ cd home && mkdir build && cd build
In case you prefer to use a specific compiler like clang instead of gcc run:
$ export CC=/usr/bin/clang $ export CXX=/usr/bin/clang++
Last step is the cmake configuration which is done by running this inside the build folder:
$ cmake ..
Other cmake variables can be passed in this step for example cmake -DCMAKE_BUILD_TYPE=Release .. which will change the build target to “Release”. More details about cmake variables can be found in Customizing the PCL build process.
Finally compile everything by running:
$ make -j2
Install the result on docker:
$ make -j2 install
To get root access for just install command:
$ docker exec -it <container_name>
All the steps mentioned in this tutorial should be performed at least once and after that just running the container command and building or installing is enough. Periodically though it is recommended to pull the latest image to have possible updates that are incorporated in the meantime.