This repository contains a set of **dev container definitions** to help get you up and running with a containerized environment. The definitions describe the appropriate container image, runtime arguments for starting the container, and VS Code extensions that should be installed. They are intended to be dropped into an existing project or folder rather than acting as sample projects. (See the [vscode-remote-try-*](https://github.com/search?q=org%3Amicrosoft+vscode-remote-try-&type=Repositories) repositories if you are looking for sample projects.)
Have a container set up you're proud of and would like to share? Want to see some changes made to an existing definition? We love contributions! Read on to learn how.
Since devcontainer.json can be commited to a source code repository, the definitions here are not intended to cover every possible permutation. When thinking about contributing a new dev container definition, consider the following questions:
- Are the requirements cumbersome to install or are they typically things added by a package manager (pip, npm, etc) using a project's manifest file (`package.json`, `requirements.txt`, etc)?
2. How likely are other developers to find the definition useful on its own? Could the scenario be broadened to help more people?
If the definition is too similar to others, consider contributing a PR to improve an existing one instead. If the scenario is too specific consider generalizing it and making it more broadly applicable.
The contents of the folders in the `containers` directory ultimately populate the available definitions list shown in the **Remote-Containers: Add Development Container Configuration Files...** command. To make this work, each folder consists of up to three elements:
1.**The container definition itself** - These are the files and folders that will be added to a user's existing project / folder if they select the definition. Typically these files are stored in a `.devcontainer` folder.
2.**Test assets** - While you are creating your definition, you may need to use a test project to make sure it works as expected. Contributing these files back will also help others that want to contribute to your definition in the future. These files are typically located in a `test-project` folder.
3.**A `.npmignore` file** - This tells VS Code which files in the folder should be ignored when a user selects it for their project / folder. The file typically lists test assets or folders.
### Creating a new definition
Most contributions require you to agree to a Contributor License Agreement (CLA) declaring that you have the right to, and actually do, grant us the rights to use your contribution. For details, visit https://cla.microsoft.com.
When you submit a pull request, a CLA-bot will automatically determine whether you need to provide a CLA and decorate the PR appropriately (e.g., label, comment). Simply follow the instructions provided by the bot. You will only need to do this once across all repos using our CLA.
2. Create a new folder in the `containers` directory. The name of the folder is effectively the **definition ID** and should follow the following format:
````text
<language>-<optional:version>-<descriptor>
````
You'll find many examples in the current `containers` folder.
3. You can grab one of the templates from the `container-templates` folder to help you get an idea of what to contribute for different scenarios, but here's a quick summary of what you should include:
```text
📁 <language>-<optional:version>-<descriptor>
📁 .devcontainer
📄 devcontainer.json
📄 Dockerfile (optional)
📄 docker-compose.yml (optional)
📁 test-project (optional)
📄 .npmignore
📄 README.md
```
See the [VS Code Remote Development documentation](https://aka.ms/vscode-remote/docker) for information on the expected contents of `devcontainer.json` and how it relates to other files listed above.
Note that any additional assets can be included as needed, but keep in mind that these will overlay on top of an existing project. Keeping these files in the `.devcontainer` folder should reduce the chances of something conflicting but note that any commands that are run are relative to the root of the project, so you'll need to include `.devcontainer` in any path references.
Finally, create a `README.md` in the folder with a brief description of the purpose of the container definition and any manual steps required to use it.
4. Update [`.npmignore`](https://docs.npmjs.com/misc/developers#keeping-files-out-of-your-package) if you've added new folders that should be excluded if used. Add anything you don't want copied in to a user's existing project / folder into this file in [glob](https://facelessuser.github.io/wcmatch/glob/) form.
Each `RUN` statement creates a Docker image "layer". If one `RUN` statement adds in temporary contents, these contents remain in this layer in the image even if they are deleted in a subsequent `RUN`. This means the image takes more storage locally and results in slower image download times if you publish the image to a registry.
... or across multiple lines (note the `\` at the end escaping the newline):
```Dockerfile
RUN apt-get update \
&& apt-get -y install git \
&& apt-get autoremove -y \
&& apt-get clean -y \
&& rm -rf /var/lib/apt/lists/*
```
2. Put the commands in a script, temporarily copy it into the container, then remove it. e.g.:
```Dockerfile
COPY ./my-script.sh /tmp/my-script.sh
RUN bash /tmp/my-script.sh \
&& rm -f /tmp/my-script.sh
```
Some other tips:
1. You'd be suprised [how big package lists](https://askubuntu.com/questions/179955/var-lib-apt-lists-is-huge) can get, so be sure to clean these up too. Most Docker images that use Debian / Ubuntu use the following command to clean these up:
The only downside of doing this is that `apt-get update` has to be executed before you install a packages. However, in most cases adding this package to a Dockerfile is a better choice anyway since this will survive a "rebuild" of the image and the creation of an updated container.
2. Use the scripts in the [script library](./script-library) in this repository where appropriate. You do not even need to copy the script into your `.devcontainer` folder to use it. See the [README](./script-library) for details. Most existing definitions use the "common" script to ensure things like `git`, a non-root user, and useful command line utilities like `ps`, `ip`, `jq` are present.
3. In all cases, you'll want to pay attention to package caching since this can also take up image space. Typically there is an option for a package manager to not cache when installing that you can use to minimize the size of the image. For example, for Alpine Linux, there's `apk --no-cache`
4. Watch out for the installation of "recommended" packages you don't need. By default, Debian / Ubuntu's `apt-get` installs packages that are commonly used with the one you specified - which in many cases isn't required. You can use `apt-get -y install --no-install-recommends` to avoid this problem.
Note that if you make major changes, Docker may occasionally not pick up your edits. If this happens, you can delete the existing container and image, open the folder locally, and go to step 2 above. Install the [Docker extension](https://marketplace.visualstudio.com/items?itemName=PeterJausovec.vscode-docker) locally (when not in a container) to make this easy.
After you get your container up and running, you can test it by adding test assets / projects into the definition folder and then adding their locations to the `.npmignore` file in [glob](https://facelessuser.github.io/wcmatch/glob/) form relative to the root of the folder. By convention, most definitions place test assets in a `test-project` folder and this path is referenced in the template `.npmignore` files.
Finally, commit your changes and submit a PR - we'll take a look at it, provide any needed feedback, and then merge it in. We appreciate any and all feedback!
While using a `Dockerfile` is a convenient way to get going with a new container definition, this method can slow down the process of creating the dev container since it requires the image to be built by anyone using it. If your definition is stable, we strongly recommend building and publishing your image to [GitHub Container Registry](https://docs.github.com/en/free-pro-team@latest/packages/getting-started-with-github-container-registry/about-github-container-registry) instead.
Once you've published your container image, just update `devcontainer.json` to reference the image instead of the `Dockerfile`. See `container-templates/image` for an example.
### Release cadence for new containers or container updates
The vscode-dev-containers repo is currently bundled with the VS Code Remote - Containers extension to ensure version compatibility. (There is a feature request to update the container list [out-of-band](https://github.com/microsoft/vscode-remote-release/issues/425), but this is not currently in place.)
Therefore, the release schedule for vscode-dev-containers is the same as the extension. The extension follows the same release cadence and schedule for **VS Code (stable)** as VS Code itself.
* VS Code uses [four week iterations](https://github.com/microsoft/vscode/wiki/Development-Process#inside-an-iteration).
* You can find the dates of the current iteration in a pinned issue in the [vscode repository](https://github.com/microsoft/vscode) and in-flight features for the extension itself in the [vscode-remote-release](https://github.com/microsoft/vscode-remote-release) repository.
The Remote - Containers extension ships into **VS Code Insiders** as new features land during the iteration. This is not currently an automated daily release like VS Code itself, so the exact timing during the iteration can vary. However, it will happen during endgame week at a minimum (week 4).
If there is an urgent need for an update to the definitions outside of this release cycle, please raise an issue in this repository requesting an out-of-band release to fix a critical issue.
The majority of VS Code Remote's documentation can be found in the [VS Code docs repository](https://github.com/Microsoft/vscode-docs). This is usually the best place to contribute, but if you have a correction or suggestion for the content found here, we welcome your contributions.
Have you identified a reproducible problem in a dev container definition? We want to hear about it! Here's how you can make reporting your issue as effective as possible.
This repository is specifically for dev container definitions. If you are not looking to report an issue related to a container definition, you may want to report the issue in the feedback repository for [VS Code Remote Development extensions](https://github.com/Microsoft/vscode-remote-release). Or you may be looking for the [VS Code OSS](https://github.com/Microsoft/vscode) repository. However, note that, the VS Code project is distributed across multiple repositories. See the list of [Related Projects](https://github.com/Microsoft/vscode/wiki/Related-Projects) if you aren't sure which repo is correct.
Before you create a new issue, please do a search in [open issues](https://github.com/Microsoft/vscode-dev-containers/issues) to see if the issue or feature request has already been filed.
Be sure to scan through the [most popular](https://github.com/Microsoft/vscode-dev-containers/issues?q=is%3Aopen+is%3Aissue+label%3Afeature-request+sort%3Areactions-%2B1-desc) feature requests.
If you find your issue already exists, make relevant comments and add your [reaction](https://github.com/blog/2119-add-reactions-to-pull-requests-issues-and-comments). Use a reaction in place of a "+1" comment:
* 👍 - upvote
* 👎 - downvote
If you cannot find an existing issue that describes your bug or feature, create a new issue using the guidelines below.
### Writing Good Bug Reports and Feature Requests
File a single issue per problem and feature request. Do not enumerate multiple bugs or feature requests in the same issue.
Do not add your issue as a comment to an existing issue unless it's for the identical input. Many issues look similar, but have different causes.
The more information you can provide, the more likely someone will be successful at reproducing the issue and finding a fix.
Once submitted, your report will go into a similar [issue tracking](https://github.com/Microsoft/vscode/wiki/Issue-Tracking) workflow that is used for the VS Code project. Be sure to understand what will happen next, so you know what to expect, and how to continue to assist throughout the process.
