Add Docker build system and containerized development environment

.actrc

@@ -0,0 +1,2 @@
+-P ubuntu-latest=node:22-bookworm
+--container-options=--memory=32g --network=host

.dockerignore

@@ -0,0 +1,17 @@
+# Exclude large directories that aren't needed for build
+vscode/node_modules
+
+# Exclude build artifacts
+*.tar.gz
+*.log
+
+# Exclude development files
+.vscode/
+.idea/
+*.swp
+*.swo
+*~
+
+# Exclude OS files
+.DS_Store
+Thumbs.db

.github/workflows/build.yml

@@ -9,6 +9,7 @@ concurrency:
   cancel-in-progress: ${{ github.event_name == 'pull_request' }}
 
 jobs:
+  # The main job for building the application
   build:
     name: Build sagemaker-code-editor
     runs-on: ubuntu-latest
@@ -37,7 +38,13 @@ jobs:
       - name: Apply patches (if any)
         run: |
           if [ -d patches ] && [ "$(ls -A patches)" ]; then
-            quilt push -a || true
+            {
+              quilt push -a --leave-rejects --color=auto 
+            } || {
+              printf "\nPatching error, review logs!\n"
+              find ./vscode -name "*.rej"
+              exit 1
+            }
           fi
 
       - name: Set Development Version
@@ -53,18 +60,17 @@ jobs:
           cd vscode
           export DISABLE_V8_COMPILE_CACHE=1
           export UV_THREADPOOL_SIZE=4
+
+          npm install -g node-gyp
 
+          # Install dependencies using npm, skip optional and native modules
           npm install
 
-          VSCODE_RIPGREP_VERSION=$(jq -r '.dependencies."@vscode/ripgrep"' package.json)
-          mv package.json package.json.orig
-          jq 'del(.dependencies."@vscode/ripgrep")' package.json.orig > package.json
-
-          npm install
-          npm install --ignore-scripts "@vscode/ripgrep@${VSCODE_RIPGREP_VERSION}"
-
+          # Run the gulp build task with memory optimizations
           ARCH_ALIAS=linux-x64
-          npx gulp vscode-reh-web-${ARCH_ALIAS}-min
+          node --max-old-space-size=32768 --optimize-for-size \
+              ./node_modules/gulp/bin/gulp.js \
+              "vscode-reh-web-${ARCH_ALIAS}-min"
 
       - name: Find build output
         id: find_output
@@ -96,6 +102,7 @@ jobs:
           tar czf $TARBALL -C "$PARENT_DIR" "$BUILD_DIR_NAME"
 
       - name: Upload build artifact
+        if: env.ACT == ''
         uses: actions/upload-artifact@v4
         with:
           name: npm-package

.github/workflows/e2e.yml

@@ -16,23 +16,20 @@ jobs:
     runs-on: ubuntu-latest
     if: github.event.workflow_run.conclusion == 'success' || github.event_name == 'pull_request'
     steps:
-      - uses: actions/checkout@v4
-      - uses: actions/setup-node@v4
+      - name: Checkout repo with submodules
+        uses: actions/checkout@v4
         with:
-          node-version: 20
-          cache: npm
-      - run: npm ci
-      - name: Download build artifact
-        if: github.event_name != 'pull_request'
-        uses: actions/download-artifact@v4
+          submodules: recursive
+      - name: Set up Node.js
+        uses: actions/setup-node@v4
         with:
-          name: build-package
-          github-token: ${{ secrets.GITHUB_TOKEN }}
-          run-id: ${{ github.event.workflow_run.id }}
+          node-version: 20
+          cache: 'npm'
+          cache-dependency-path: '**/package-lock.json'
       - name: Run E2E tests
         run: |
           # Add your actual E2E test commands here
-          echo "Running E2E tests..."
+          echo "[PLACEHOLDER] Running E2E tests..."
           # Example: npm run test:e2e
       - uses: actions/upload-artifact@v4
         if: failure()

.gitignore

@@ -1,2 +1,4 @@
 .DS_Store
-.pc
+.pc
+.artifacts
+bin

Makefile

@@ -0,0 +1,82 @@
+.PHONY: build-cache build install-act run-github run-local clean-vscode clean
+
+build-cache:
+	@echo "Building SageMaker Code Editor (multi-stage npm cache)..."
+	docker buildx build \
+		--platform linux/amd64 \
+		--progress=plain \
+		--memory=32g \
+		-t npm-cache:latest \
+		-f scripts/Dockerfile.build.cache .
+
+build:
+	@echo "Building SageMaker Code Editor (original)..."
+	docker buildx build \
+		--platform linux/amd64 \
+		--progress=plain \
+		--memory=32g \
+		--output type=local,dest=./.artifacts \
+		-t localbuild:latest \
+		-f scripts/Dockerfile.build .
+
+install-act:
+	@echo "Installing act (GitHub Actions runner)..."
+	@if ! command -v act >/dev/null 2>&1 && [ ! -f ./bin/act ]; then \
+		curl -s https://raw.githubusercontent.com/nektos/act/master/install.sh | bash; \
+		echo "act installed successfully"; \
+	else \
+		echo "act is already available"; \
+	fi
+
+run-github: install-act
+	@echo "Running complete GitHub Actions workflow locally..."
+	@echo "Available workflows:"
+	@ls -la .github/workflows/
+	@echo ""
+	@echo "Running full build.yml workflow..."
+	@if command -v act >/dev/null 2>&1; then \
+		act push -W .github/workflows/build.yml --platform ubuntu-22.04=catthehacker/ubuntu:act-22.04 --container-options "--memory=32g --memory-swap=32g"; \
+	else \
+		./bin/act push -W .github/workflows/build.yml --platform ubuntu-22.04=catthehacker/ubuntu:act-22.04 --container-options "--memory=32g --memory-swap=32g"; \
+	fi
+
+run-local:
+	@if [ -z "$(TARBALL)" ]; then \
+		echo "Building and running SageMaker Code Editor locally on port 8888..."; \
+		docker build -f scripts/Dockerfile.dev -t local-code-editor-dev . || exit 1; \
+		echo "Stopping container..."; \
+		docker stop local-code-editor-dev; \
+		sleep 2; \
+		echo "Starting container on http://localhost:8888"; \
+		docker run --rm -d -p 8888:8000 -v .:/workspace -u $(id -u):$(id -g) --entrypoint /workspace/scripts/run-code-editor-dev.sh --name local-code-editor-dev local-code-editor-dev || exit 1; \
+		docker logs -f local-code-editor-dev; \
+	else \
+		echo "Building and running SageMaker Code Editor locally on port 8888..."; \
+		docker build -f scripts/Dockerfile.run --build-arg TARBALL=$(TARBALL) -t local-code-editor . || exit 1; \
+		echo "Stopping container..."; \
+		docker stop local-code-editor; \
+		sleep 2; \
+		echo "Starting container on http://localhost:8888"; \
+		docker run --rm -d -p 8888:8000 --name local-code-editor local-code-editor || exit 1; \
+		docker logs -f local-code-editor; \
+	fi
+
+clean-vscode:
+	@echo "Cleaning VSCode node_modules..."
+	@find . -type d -name "node_modules" -exec rm -rf {} + 2>/dev/null || true
+	@rm -rf vscode/out/* 2>/dev/null || true
+	@echo "VSCode cleanup completed"
+
+clean: clean-vscode
+	@echo "Cleaning act temporary files and Docker images..."
+	@echo "Removing act cache..."
+	@rm -rf ~/.cache/act 2>/dev/null || true
+	@echo "Act cleanup completed"
+
+reset-vscode:
+	@echo "Resetting vscode submodule..."
+	git submodule update --init --recursive
+	git submodule foreach --recursive "git reset --hard HEAD && sudo git clean -fd"
+	@echo "Resetting patches..."
+	sudo rm -rf .pc/*
+

README.md

@@ -28,6 +28,25 @@ This script will:
 - Run `yarn watch` from within the `vscode` folder
 - Open a new terminal and run `./vscode/scripts/code-server.sh --launch`
 
+## Make Commands
+
+Available make targets for building and testing:
+
+### 1. When making local changes to iterate faster where tarball generation is not required [each run takes 10-20 mins]
+- `make run-local` - Build and run SageMaker Code Editor locally from source and does not require a TARBALL; this process runs a watcher so changes are automatically picked from local workspace
+- `make clean-vscode` - Cleans node_modules and out files
+
+### 2. Once local changes are tested; follow this process to generate minified tarball [each run takes ~40 mins to build] 
+- `make build-cache` - Build SageMaker Code Editor with multi-stage npm cache; Run once and layer gets cached with node_modules
+- `make build` - Build SageMaker Code Editor and output artifacts (tarball) to ./artifacts
+- `make run-local TARBALL=<tarball-name>` - Build and run SageMaker Code Editor locally on port 8888 using specified tarball from previos step. Example: `make run-local TARBALL=sagemaker-code-editor-1.101.2.tar.gz`
+
+### 3. This process is used to test and simulate github workflows locally [each run takes ~60 mins] 
+- `make run-github` - Run complete GitHub Actions workflow locally using act
+
+### 4. Cleanup
+- `make clean` - Cleans node_modules, out files, and act temporary files
+
 ## Troubleshooting and Feedback
 
 For any issues that customers would like to report, please route to the `amazon-sagemaker-feedback` repository: https://github.com/aws/amazon-sagemaker-feedback

scripts/Dockerfile.build

@@ -0,0 +1,27 @@
+FROM npm-cache:latest AS builder
+
+WORKDIR /workspace
+
+# Declare a build argument to control the minify flag
+ARG ARGS
+
+# Copy source files except vscode so new patches are picked up without cache busting
+COPY patches ./patches
+COPY resources ./resources
+COPY .git ./.git
+COPY scripts/ ./scripts/
+COPY .pc ./.pc
+COPY LICENSE .
+COPY LICENSE-THIRD-PARTY .
+
+# Apply patches and build
+RUN chmod +x scripts/docker-install.sh && \
+    ./scripts/docker-install.sh -t "$(cat vscode/package.json | grep '"version"' | cut -d'"' -f4)"
+
+RUN echo "Build arguments: $ARGS"
+RUN chmod +x scripts/create-code-editor-tarball.sh && \
+    ./scripts/create-code-editor-tarball.sh -v "$(cat vscode/package.json | grep '"version"' | cut -d'"' -f4)"
+
+# Final stage: Minimal image with only the build artifacts to copy from
+FROM scratch
+COPY --from=builder /workspace/artifacts .

scripts/Dockerfile.build.cache

@@ -0,0 +1,43 @@
+FROM ubuntu:24.04 AS base-image
+
+# Install required tools and Node.js
+RUN apt-get update && apt-get install -y \
+    git \
+    python3 \
+    python3-pip \
+    build-essential \
+    curl \
+    quilt \
+    pkg-config \
+    libx11-dev \
+    libxkbfile-dev \
+    libsecret-1-dev \
+    libkrb5-dev \
+    libgssapi-krb5-2 \
+    time \
+    && rm -rf /var/lib/apt/lists/*
+
+# Install Node.js 22 (latest LTS)
+RUN curl -fsSL https://deb.nodesource.com/setup_22.x | bash - && \
+    apt-get install -y nodejs
+
+# Stage 2: NPM Dependencies Cache
+# This stage uses the pre-configured base-image to build the cache.
+FROM base-image AS npm-cache
+
+WORKDIR /workspace
+
+COPY vscode ./vscode
+COPY patches ./patches
+COPY resources ./resources
+COPY .git ./.git
+COPY scripts/postinstall.sh ./scripts/postinstall.sh
+COPY scripts/docker-install.sh ./scripts/docker-install.sh
+COPY scripts/copy-resources.sh ./scripts/copy-resources.sh
+COPY .pc ./.pc
+COPY LICENSE .
+COPY LICENSE-THIRD-PARTY .
+
+# Apply patches and build
+RUN chmod +x scripts/docker-install.sh && \
+    ./scripts/docker-install.sh -t "$(cat vscode/package.json | grep '"version"' | cut -d'"' -f4)"

scripts/Dockerfile.dev

@@ -0,0 +1,36 @@
+FROM ubuntu:24.04 AS base-image
+
+# Install required tools and Node.js
+RUN apt-get update && apt-get install -y \
+    git \
+    python3 \
+    python3-pip \
+    build-essential \
+    curl \
+    quilt \
+    pkg-config \
+    libx11-dev \
+    libxkbfile-dev \
+    libsecret-1-dev \
+    libkrb5-dev \
+    libgssapi-krb5-2 \
+    time \
+    && rm -rf /var/lib/apt/lists/*
+
+# Install Node.js 22 (latest LTS)
+RUN curl -fsSL https://deb.nodesource.com/setup_22.x | bash - && \
+    apt-get install -y nodejs
+
+# Stage 2: NPM Dependencies Cache
+# This stage uses the pre-configured base-image to build the cache.
+FROM base-image AS npm-cache
+
+# Install supervisord
+RUN apt-get update && \
+    apt-get install -y supervisor
+
+WORKDIR /workspace
+
+EXPOSE 8080
+
+# Following is the entrypoint script for local runs: /workspace/scripts/run-code-editor-dev.sh

scripts/Dockerfile.run

@@ -0,0 +1,24 @@
+FROM public.ecr.aws/sagemaker/sagemaker-distribution:latest-cpu
+
+# Accept tarball name as build argument
+ARG TARBALL=code-editor1.8.0b5.tar.gz
+
+# Switch to root to install
+USER root
+
+# Copy and extract the code editor to /tmp
+COPY .artifacts/${TARBALL} /tmp/
+RUN cd /tmp && tar -xzf ${TARBALL}
+
+# Move to final location and set permissions
+RUN mv /tmp/sagemaker-code-editor /opt/ && \
+    chmod +x /opt/sagemaker-code-editor/bin/code-server-oss
+
+# Add to PATH
+ENV PATH="/opt/sagemaker-code-editor/bin:$PATH"
+
+# Expose port
+EXPOSE 8000
+
+# Run code-server-oss with host binding to all interfaces
+CMD ["code-server-oss", "--host", "0.0.0.0", "--without-connection-token"]