Tired of trying to compile ROS1 on Ubuntu 22.04? Tired of breaking your systems due to conflicts with the ton of dependencies ROS1 has? Here we propose a solution. A completely isolated environment, with ROS1 support by default, extensible, and with all the tools included for developing ROS1 applications without having to install them locally.
ros_in_doccker.mp4
- Ubuntu 22.04
- tmux
- VSCode
- Click on "Use this template", or fork if you like.
- Git clone the repo locally and
cd
into it. - Create an
src/
directory inside the repository:mkdir -p src/
- Clone inside the
src/
directory the ROS1 code you want to develop/test. I will be using the ros_tutorials as an example, but it can be as complex as you wish, so,git clone git@github.com:ros/ros_tutorials.git src/
- Launch
code .
, and then go to the "Remote Explorer" tab and hit "reopen the current folder in a container", this should launch a full dev environemnt with some extensions to develop your ROS application in the dockerize environment - Launch the
Build Task
, Ctrl+Shift+p and type "Tasks: Run Build Task"
- Click on "Use this template", or fork if you like.
- Git clone the repo locally and
cd
into it. - Launch
make docker
to create a local docker container image with all the nice stuff installed. If you need any extra dependency, now is the time to add it to the Dockerfile. - Create an
src/
directory inside the repository:mkdir -p src/
- Clone inside the
src/
directory the ROS1 code you want to develop/test. I will be using the ros_tutorials as an example, but it can be as complex as you wish, so,git clone git@github.com:ros/ros_tutorials.git src/
- Launch
make
, this will build your project and open atmux
session with all the batteries included.
Got you covered, just use the environment variable export ROS_BAGS=/path/to/data/in/host
and
launch make
. Your rosbag files will be mounted in the dev container in ~/ros_ws/bags
and are
read-only accessible to the ROS1 applications.
Could not be more simple, just add all your user-space command on the Dockerfile.
apt install <your-libs>
and enjoy the setup.
Of course, you can run rviz
and friends inside the dev container, it looks a bit ugly but at least
it works.
For now, you only need, I expect this repo to be used by "intermediate" developers, so I guess you can figure out how to do that.
- docker
- docker-compose
- (VSCode) dev-containers extension
The entire setup is NOT safe at all, so, use it at your own risk. I'm mounting directories from the
host machine to the docker container where the user has sudo
access without a password. So you can
literally delete some stuff with root
permissions without even typing a password. So, you are
warned!
I built this project on top of my dotfiles, so it's completely overfitted to my own needs and I don't intend to provide support for extra stuff. If there is something you don't like or don't need, feel free to modify your copy of the Dockerfile. You are in full control of what you want and not.