I was thinking about a solution in development to have node_modules in my Docker containers and on my host machine to have Intelisense on VSCode for Typescript.

Is this solution (which works) a good one? With this way of doing things, I can install my dependencies directly in Docker.

The advantage of Docker remains the same, despite the fact that it’s less isolated because of the volumes. I don’t have to install NodeJS locally, and my production environment remains the same (same Docker images).

—

Dockerfile:

FROM arm64v8/node:20.17-alpine

WORKDIR /app

ARG NODE_ENV=development

ENV PATH=/app/node_modules/.bin:$PATH

EXPOSE 4321

ENTRYPOINT ["yarn", "workspace", "web", "dev"]

docker-compose.yml:

services:
  install-deps:
    image: arm64v8/node:20.17-alpine
    entrypoint: ./docker-entrypoint.sh
    profiles: ["install-deps"]
    build:
      context: .
    working_dir: /app
    volumes:
      - ./:/app

  web:
    build:
      context: ./
      dockerfile: apps/web/Dockerfile
    restart: unless-stopped
    volumes:
      - ./apps/web:/app/apps/web
      - ./node_modules:/app/node_modules
      - ./turbo.json:/app/turbo.json
      - ./package.json:/app/package.json
      - ./yarn.lock:/app/yarn.lock

The command to install dependencies:

docker compose run --rm install-deps yarn