armgits/astar-turtlebot

ROS Humble Workspace

For package development and testing.

Setting up

Requirements

Setting up the workspace at least requires Linux, Docker, Git and VSCode on your computer.

Refer the Docker documentation for installing Docker engine. Choose to install Docker Engine with CLI over Docker Desktop.

Note: Check the customization section to learn more about the ways to make the workspace tailor-made for your project.

Cloning the repository

Since this is a template repository, you can create your own repository with the current layout as a starting point if you wish to.

Otherwise, cloning the repository to your home directory is sufficient to get started.

Get started with development

Make sure to install Dev containers and Docker VSCode extensions on host.

Open the folder of workspace in VSCode and click on the remote icon at the bottom-left corner of the screen. Choose Reopen in Container to open the folder in the container. That option also could appear in the pop-up at the bottom right if VSCode detects the .devcontainer/ folder.

Note: Building the container takes some time for the first time.

If the container builds and runs cleanly, the development environment is successfully setup!! Use the terminal in VSCode to do stuff.

Note: Make sure the type of terminal is setup as bash and set the default profile in the drop down menu beside the + icon.

Customizing the workspace

GPU Access

Docker can provide GPU acceleration to applications running inside the container if setup correctly. The steps to modify depends on the specific GPU.

The repository by default is set up to support Intel iGPUs (not tested yet) or no GPU at all to maximize compatibility.

For NVIDIA GPUs

• Install the container toolkit on host by following the steps in documentation.

• In devcontainer.json, set the build target to nvidia and comment/uncomment the appropriate run flags by following the comments.

For AMD GPUs

• Install ROCm software on host by following the steps in documentation. It is preferable to use the Native package manager method.

• In devcontainer.json set the build target to amd, uncomment the build arguments and comment/uncomment the run arguments by following the comments.

Setting a custom workspace name and directory

This requires modifying the devcontainer.json file in the .devcontainer/ directory.

The workspaceMount, workspaceFolder and the WORKSPACE variable in the containerEnv section deal with the way Docker mounts the workspace directory from host onto the container.

VSCode Extensions

This requires modifying the devcontainer.json in the .devcontainer/ directory.

Extensions inside the container are seperate from host. Extensions from the host are accessible inside the container, same isn't true in reverse. So a desired set of add-ons specific to the workspace can be added in the container!

Modify the customizations field inside devcontainer.json.

Changing the username

This requires modifying both devcontainer.json and Dockerfile in the .devcontainer/ directory.

Make sure the remoteUser field (in devconatiner.json) and USERNAME argument (in both devcontainer.json and Dockerfile) match to your desired username.

In addition, paths mentioned across both the files that contain user directory might need to be updated.

My suggestion: Simply find and replace the username keyword in both the files.

Credits

This article by Allison Thackston.