Devcontainers provide a way for developers with an IDE. Like regular containers, devcontainers can be packaged with the appropriate tool versions, extensions, build & test targets out of the box.
Currently, the devcontainer setup leverages docker compose to setup two containers:
- The main development container which contains:
- Node and Python
- Datastore Emulator
- Datastore Emulator viewer
An extra step is required for Devcontainer on Mac:
- Click on the Docker icon, go to
Preferences
- Go to the Resources tab and select FILE SHARING
- Add
/workspaces
to the mountable directories - Click on Apply & Restart
You can click on the appropriate badge to get the environment setup:
Service | Click The Badge | Requirements | Pros | Cons |
---|---|---|---|---|
Visual Studio Code Remote - Containers | Locally: Docker and Visual Studio Code | Runs locally. Have more resources | Need to have VS Code and Docker setup | |
GitHub Codespaces | Access to Codespaces | Nothing to install locally | Limited resources: 4 cores, 8GB RAM, 32GB Storage |
Upon creating the devcontainer by clicking the appropriate badge, npm run setup
is run automatically.
Note: Sometimes when you open the terminal, it won't automatically activate
the python environment. Either 1) Open a new terminal (usually it works the
second time) or 2) manually activate it by running source cs-env/bin/activate
Most commands from the root README.md work as-is. There are some exceptions due to the database container being outside the main dev container.
Local | DevContainer |
---|---|
npm test |
npm run do-tests |
npm start |
npm run start-app |
To access the database, database viewer or the app when it is running, go to the "Ports" section next to the "Terminal". Find the Port you are looking for and click on the corresponding Local Address.
There are other ports needed for the IDE itself but mainly, you only need to watch:
Port | Service |
---|---|
15606 | Datastore Emulator |
8888 | Datastore Viewer |
8080 | Application (when running) |
8000 | Webtest (when running) |
In the docker-compose.yml file:
- For Python: Change the
VARIANT
argument to the appropriate variant - For Node: Change the
NODE_VERSION
to the appropriate version used by nvm
When there are dependency changes and environment changes in requirements.txt or package.json, the Devcontainer needs to be rebuilt:
rm -rf node_modules cs-env
in terminal- View -> Command Palette -> Rebuild container without cache