Add Docker build system and containerized development environment

Arkaprava De · Arkaprava De · commit 31df0b7a74fc · 2025-09-11T18:53:47.000Z
- Add Docker build configurations for development and production
- Add Makefile for simplified build commands
- Add GitHub Actions cache support for Docker builds
- Add .actrc for local GitHub Actions testing
- Add .dockerignore for optimized Docker builds
- Add scripts for Docker-based installation and tarball creation
- Remove old tarball creation script
diff --git a/.actrc b/.actrc
@@ -0,0 +1,2 @@
+-P ubuntu-latest=node:22-bookworm
+--container-options=--memory=32g --network=host
diff --git a/.dockerignore b/.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
diff --git a/.github/workflows/build.yml b/.github/workflows/build.yml
@@ -21,14 +21,15 @@ jobs:
 
   run-unit-tests:
     name: Run unit tests
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     steps:
       # Checkout repository code
       - name: Checkout code
         uses: actions/checkout@v4
 
       # Verify CSP line exists in target TypeScript file
       - name: Check CSP configuration in webClientServer.ts
+        shell: sh
         run: |
           TARGET_FILE="patched-vscode/src/vs/server/node/webClientServer.ts"
           REQUIRED_TEXT="'connect-src \'self\' ws: wss: https://main.vscode-cdn.net http://localhost:* https://localhost:* https://login.microsoftonline.com/ https://update.code.visualstudio.com https://*.vscode-unpkg.net/ https://default.exp-tas.com/vscode/ab https://vscode-sync.trafficmanager.net https://vscode-sync-insiders.trafficmanager.net https://*.gallerycdn.vsassets.io https://marketplace.visualstudio.com https://openvsxorg.blob.core.windows.net https://az764295.vo.msecnd.net  https://code.visualstudio.com https://*.gallery.vsassets.io https://*.rel.tunnels.api.visualstudio.com wss://*.rel.tunnels.api.visualstudio.com https://*.servicebus.windows.net/ https://vscode.blob.core.windows.net https://vscode.search.windows.net https://vsmarketplacebadges.dev https://vscode.download.prss.microsoft.com https://download.visualstudio.microsoft.com https://*.vscode-unpkg.net https://open-vsx.org;'"
@@ -50,10 +51,10 @@ jobs:
   # The main job for building the application
   build:
     name: Build sagemaker-code-editor
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     # Ensure unit tests pass before building
     needs: run-unit-tests 
-    timeout-minutes: 180
+    timeout-minutes: 60
     env:
       # Environment variable to optimize the build process
       DISABLE_V8_COMPILE_CACHE: 1
@@ -104,24 +105,17 @@ jobs:
           cd vscode
           export DISABLE_V8_COMPILE_CACHE=1
           export UV_THREADPOOL_SIZE=4
+
+          npm install -g node-gyp
           
-          # Install dependencies using npm
-          npm install
-          
-          # The logic for temporarily removing and re-adding ripgrep remains
-          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
-          
-          # Re-run install to remove ripgrep
+          # Install dependencies using npm, skip optional and native modules
           npm install
           
-          # Add ripgrep back using npm
-          npm install --ignore-scripts "@vscode/ripgrep@${VSCODE_RIPGREP_VERSION}"
-          
+          # Run the gulp build task with memory optimizations
           ARCH_ALIAS=linux-x64
-          # Run the gulp build task using npx
-          npx gulp vscode-reh-web-${ARCH_ALIAS}-min
+          node --max-old-space-size=32768 --optimize-for-size \
+              ./node_modules/gulp/bin/gulp.js \
+              "vscode-reh-web-${ARCH_ALIAS}-min"
 
       # Step 7: Find the exact path of the original build output directory.
       - name: Find build output
@@ -157,14 +151,15 @@ jobs:
       
       # Step 10: Upload the tarball as a build artifact.
       - name: Upload build artifact
+        if: env.ACT == ''
         uses: actions/upload-artifact@v4
         with:
           name: npm-package
           path: sagemaker-code-editor-${{ env.VERSION }}.tar.gz
   # Run end-to-end tests after the build is complete
   run-e2e-tests:
     name: Run e2e tests
-    runs-on: ubuntu-latest
+    runs-on: ubuntu-22.04
     needs: build  # Ensure e2e tests run after build
     steps:
       # Checkout repository code
diff --git a/Makefile b/Makefile
@@ -0,0 +1,62 @@
+.PHONY: build-cache build install-act run-github run-local clean
+
+build-cache:
+	@echo "Building SageMaker Code Editor (multi-stage npm cache)..."
+	docker buildx build \
+		--progress=plain \
+		--memory=32g \
+		-t npm-cache:latest \
+		-f scripts/Dockerfile.build.cache .
+
+build:
+	@echo "Building SageMaker Code Editor (original)..."
+	docker buildx build \
+		--progress=plain \
+		--memory=32g \
+		--output type=local,dest=./artifacts \
+		-t localbuild:latest \
+		-f scripts/Dockerfile.build .
+
+install-act:
+	@echo "Installing act (GitHub Actions runner)..."
+	@if ! command -v act >/dev/null 2>&1 && [ ! -f ./bin/act ]; then \
+		curl -s https://raw.githubusercontent.com/nektos/act/master/install.sh | bash; \
+		echo "act installed successfully"; \
+	else \
+		echo "act is already available"; \
+	fi
+
+run-github: install-act
+	@echo "Running complete GitHub Actions workflow locally..."
+	@echo "Available workflows:"
+	@ls -la .github/workflows/
+	@echo ""
+	@echo "Running full build.yml workflow..."
+	@if command -v act >/dev/null 2>&1; then \
+		act push -W .github/workflows/build.yml --platform ubuntu-22.04=catthehacker/ubuntu:act-22.04 --container-options "--memory=32g --memory-swap=32g"; \
+	else \
+		./bin/act push -W .github/workflows/build.yml --platform ubuntu-22.04=catthehacker/ubuntu:act-22.04 --container-options "--memory=32g --memory-swap=32g"; \
+	fi
+
+run-local:
+	@if [ -z "$(TARBALL)" ]; then \
+		echo "Usage: make run-local TARBALL=<tarball-name>"; \
+		echo "Example: make run-local TARBALL=code-editor1.8.0b5.tar.gz"; \
+		exit 1; \
+	fi
+	@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 .
+	docker stop local-code-editor 2>/dev/null || true
+	@echo "Starting container on http://localhost:8888"
+	docker run --rm -d -p 8888:8000 --name local-code-editor local-code-editor
+	docker logs -f local-code-editor
+
+clean:
+	@echo "Cleaning build artifacts..."
+	rm -f artifacts
+	@echo "Clean build artifacts completed"
+	@echo "Cleaning act temporary files and Docker images..."
+	@echo "Removing act cache..."
+	@rm -rf ~/.cache/act 2>/dev/null || true
+	@echo "Act cleanup completed"
+
diff --git a/README.md b/README.md
@@ -28,6 +28,19 @@ 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:
+
+- `make build-cache` - Build SageMaker Code Editor with multi-stage npm cache
+- `make build` - Build SageMaker Code Editor and output artifacts to ./artifacts
+- `make install-act` - Install act (GitHub Actions runner) for local testing
+- `make run-github` - Run complete GitHub Actions workflow locally using act
+- `make run-local TARBALL=<tarball-name>` - Build and run SageMaker Code Editor locally on port 8888 using specified tarball
+- `make clean` - Clean build artifacts and act temporary files
+
+Example: `make run-local TARBALL=code-editor1.8.0b5.tar.gz`
+
 ## 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
diff --git a/artifacts/code-editor1.8.0b5.tar.gz b/artifacts/code-editor1.8.0b5.tar.gz
diff --git a/artifacts/sagemaker-code-editor-1.101.2.tar.gz b/artifacts/sagemaker-code-editor-1.101.2.tar.gz
diff --git a/bin/act b/bin/act
diff --git a/scripts/Dockerfile.build b/scripts/Dockerfile.build
@@ -0,0 +1,17 @@
+FROM npm-cache:latest AS builder
+
+WORKDIR /workspace
+
+# Declare a build argument to control the minify flag
+ARG ARGS
+
+# Apply patches and build
+COPY scripts/create-code-editor-tarball.sh ./scripts/create-code-editor-tarball.sh
+
+RUN echo "ARKA 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 .
diff --git a/scripts/Dockerfile.build.cache b/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 all package.json and package-lock.json using BuildKit glob
+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 LICENSE .
+COPY LICENSE-THIRD-PARTY .
+
+# Apply patches and build
+RUN chmod +x scripts/docker-install.sh && \
+    ./scripts/docker-install.sh -t "$(cat vscode/package.json | grep '"version"' | cut -d'"' -f4)"
diff --git a/scripts/Dockerfile.dev b/scripts/Dockerfile.dev
@@ -0,0 +1,25 @@
+# Stage 3: Main Build
+# This stage also uses the base-image, but only copies the necessary artifacts from Stage 2.
+FROM npm-cache
+
+WORKDIR /workspace
+
+# Copy source code
+COPY scripts ./scripts
+COPY .git ./.git
+COPY LICENSE .
+COPY LICENSE-THIRD-PARTY .
+
+# Apply patches and build
+# RUN chmod +x scripts/docker-install.sh && \
+#     ./scripts/docker-install.sh -t "$(cat vscode/package.json | grep '"version"' | cut -d'"' -f4)"
+
+WORKDIR /workspace/vscode
+
+# RUN npm i
+
+# RUN npm run watch
+
+ENTRYPOINT [ "sh", "-c", "npm run watch & ./scripts/code-web.sh --host 0.0.0.0 --port 8080" ]
+
+EXPOSE 8080
diff --git a/scripts/Dockerfile.run b/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"]
diff --git a/scripts/create-code-editor-tarball.sh b/scripts/create-code-editor-tarball.sh
@@ -0,0 +1,41 @@
+#!/bin/bash
+
+while getopts "v:m" opt; do
+  case $opt in
+    v) version="$OPTARG"
+    ;;
+    ?) usage; exit 1 ;;
+  esac
+done
+
+if [[ -z $version ]]; then
+ echo "Please provide version using '-v'";
+ exit 1
+fi
+
+VERSION=$version
+
+# Set current project root
+PROJ_ROOT=$(pwd)
+
+pushd ${PROJ_ROOT}/vscode
+
+printf "\n======== Running gulp build task ========\n"
+export DISABLE_V8_COMPILE_CACHE=1
+export UV_THREADPOOL_SIZE=4
+export ARCH_ALIAS=linux-x64
+export NODE_ENV=production
+export MINIFY=true
+node --max-old-space-size=16384 --optimize-for-size \
+    ./node_modules/gulp/bin/gulp.js \
+    "vscode-reh-web-${ARCH_ALIAS}${MINIFY:+-min}"
+
+popd
+
+TARBALL="sagemaker-code-editor-${VERSION}.tar.gz"
+BUILD_DIR_PATH=artifacts
+
+mv vscode-reh-web-linux-x64 sagemaker-code-editor
+mkdir artifacts
+tar -czf artifacts/${TARBALL} sagemaker-code-editor 
+sha256sum artifacts/${TARBALL}
diff --git a/scripts/create_code_editor_tarball.sh b/scripts/create_code_editor_tarball.sh
diff --git a/scripts/docker-install.sh b/scripts/docker-install.sh

Original file line number	Diff line number	Diff line change
`@@ -0,0 +1,2 @@`
	`1`	`+-P ubuntu-latest=node:22-bookworm`
	`2`	`+--container-options=--memory=32g --network=host`