feat: Optimize Docker caching and harden CI cache behavior

.dockerignore

@@ -0,0 +1,11 @@
+.git
+**/node_modules
+**/obj
+**/bin
+**/dist
+TestResults
+artifacts
+*.tar
+.copilot
+.buildkit-cache
+scratch

.github/workflows/Build-Test-And-Deploy.yaml

@@ -15,6 +15,9 @@ jobs:
   build-and-test:
     runs-on: ubuntu-latest
     environment: "BuildAndUploadImage"
+    permissions:
+      actions: write  # required for docker/build-push-action GHA cache writes
+      contents: read
 
     steps:
       - uses: actions/checkout@v6
@@ -49,6 +52,35 @@ jobs:
 
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@v4
+        id: setup-buildx
+
+      - name: Cache Docker build mounts (main)
+        if: github.ref == 'refs/heads/main'
+        uses: actions/cache@v5
+        id: cache-mounts-main
+        with:
+          path: .buildkit-cache
+          key: buildkit-cache-${{ hashFiles('Dockerfile', 'Directory.Packages.props', 'NuGet.config', 'src/**/package-lock.json') }}
+          restore-keys: |
+            buildkit-cache-
+
+      - name: Restore Docker build mounts (PR)
+        if: github.event_name == 'pull_request' || github.event_name == 'merge_group'
+        uses: actions/cache/restore@v5
+        id: cache-mounts-pr
+        with:
+          path: .buildkit-cache
+          key: buildkit-cache-${{ hashFiles('Dockerfile', 'Directory.Packages.props', 'NuGet.config', 'src/**/package-lock.json') }}
+          restore-keys: |
+            buildkit-cache-
+
+      - name: Inject Docker cache mounts
+        uses: reproducible-containers/buildkit-cache-dance@v3
+        with:
+          builder: ${{ steps.setup-buildx.outputs.name }}
+          cache-dir: .buildkit-cache
+          dockerfile: Dockerfile
+          skip-extraction: ${{ github.ref != 'refs/heads/main' || steps.cache-mounts-main.outputs.cache-hit == 'true' }}
 
         # Build but no push with a PR
       - name: Docker build (no push)
@@ -58,21 +90,22 @@ jobs:
           push: false
           tags: temp-pr-validation
           file: ./Dockerfile
+          cache-from: type=gha,scope=try-main
 
-      # Only build for dev registry — prod gets the image via az acr import in deploy-production
+        # Only build for dev registry — prod gets the image via az acr import in deploy-production
       - name: Build Container Image
-        if: github.event_name != 'pull_request_target' && github.event_name != 'pull_request'
+        if: github.ref == 'refs/heads/main'
         uses: docker/build-push-action@v7
         with:
           tags: ${{ vars.DEVCONTAINER_REGISTRY }}/try:${{ github.sha }},${{ vars.DEVCONTAINER_REGISTRY }}/try:latest
           file: ./Dockerfile
           context: .
           outputs: type=docker,dest=${{ github.workspace }}/tryimage.tar
-          cache-from: type=gha
-          cache-to: type=gha,mode=max
+          cache-from: type=gha,scope=try-main
+          cache-to: type=gha,mode=max,scope=try-main
 
       - name: Upload artifact
-        if: github.event_name != 'pull_request_target' && github.event_name != 'pull_request'
+        if: github.ref == 'refs/heads/main'
         uses: actions/upload-artifact@v7.0.1
         with:
           name: tryimage
@@ -138,7 +171,7 @@ jobs:
           artifact-name: 'integration-test-playlists'
 
   deploy-development:
-    if: github.event_name != 'pull_request_target' && github.event_name != 'pull_request'
+    if: github.ref == 'refs/heads/main'
     runs-on: ubuntu-latest
     needs: [build-and-test, build-and-test-windows, integration-tests]
     concurrency:
@@ -189,7 +222,7 @@ jobs:
             --image "${{ vars.DEVCONTAINER_REGISTRY }}/try:${{ github.sha }}"
 
   deploy-production:
-    if: github.event_name != 'pull_request_target' && github.event_name != 'pull_request'
+    if: github.ref == 'refs/heads/main'
     runs-on: ubuntu-latest
     needs: deploy-development
     concurrency:

Dockerfile

@@ -1,27 +1,50 @@
+# syntax=docker/dockerfile:1
 FROM mcr.microsoft.com/dotnet/sdk:10.0-azurelinux3.0 AS build-env
 WORKDIR /App
 
-# Copy everything
-COPY . ./
-
 # Make sure we run bash
 CMD ["bash"]
 
-# Make sure we get all the updates and tools we need to build
-RUN tdnf install gawk -y
-# This is Node v16.  For 18, use nodejs18.
-RUN tdnf install nodejs -y
-RUN tdnf install npm -y
-RUN tdnf clean all
+# Install all required build tools in a single layer (rarely changes — stays cached)
+# This is Node v16. For 18, use nodejs18.
+RUN --mount=type=cache,id=try-tdnf,target=/var/cache/tdnf,sharing=locked \
+    tdnf install -y gawk nodejs npm
+
+# Copy only the files needed to restore dependencies.
+# These layers are cached until a manifest file changes, so routine source edits
+# don't re-run the expensive restore steps below.
+COPY NuGet.config global.json Directory.Build.props Directory.Build.targets Directory.Packages.props TryDotNet.sln ./
+COPY eng/ ./eng/
+
+# .csproj files — one COPY per project to preserve directory structure
+COPY src/Microsoft.TryDotNet/Microsoft.TryDotNet.csproj src/Microsoft.TryDotNet/
+COPY src/Microsoft.TryDotNet.FileIntegration.Tests/Microsoft.TryDotNet.FileIntegration.Tests.csproj src/Microsoft.TryDotNet.FileIntegration.Tests/
+COPY src/Microsoft.TryDotNet.IntegrationTests/Microsoft.TryDotNet.IntegrationTests.csproj src/Microsoft.TryDotNet.IntegrationTests/
+COPY src/Microsoft.TryDotNet.SimulatorGenerator/Microsoft.TryDotNet.SimulatorGenerator.csproj src/Microsoft.TryDotNet.SimulatorGenerator/
+COPY src/Microsoft.TryDotNet.Tests/Microsoft.TryDotNet.Tests.csproj src/Microsoft.TryDotNet.Tests/
+COPY src/Microsoft.TryDotNet.WasmRunner/Microsoft.TryDotNet.WasmRunner.csproj src/Microsoft.TryDotNet.WasmRunner/
 
-# Build javascript library
-RUN /App/build-js.sh
+# npm manifests
+COPY src/microsoft-trydotnet/package.json src/microsoft-trydotnet/package-lock.json src/microsoft-trydotnet/
+COPY src/microsoft-trydotnet-editor/package.json src/microsoft-trydotnet-editor/package-lock.json src/microsoft-trydotnet-editor/
+COPY src/microsoft-trydotnet-styles/package.json src/microsoft-trydotnet-styles/package-lock.json src/microsoft-trydotnet-styles/
+COPY src/microsoft-learn-mock/package.json src/microsoft-learn-mock/package-lock.json src/microsoft-learn-mock/
+
+# Restore NuGet packages. The cache mount persists the package cache across local
+# builds; in CI the layer itself is cached by the GHA cache backend.
+RUN --mount=type=cache,id=try-nuget,target=/root/.nuget/packages \
+    dotnet restore --configfile /App/NuGet.config /App/TryDotNet.sln
+
+# Copy all remaining source (changes frequently — only layers below rebuild on edits)
+COPY . ./
 
-# Restore
-RUN dotnet restore --configfile /App/NuGet.config /App/TryDotNet.sln
+# Build javascript library. The npm cache mount speeds up repeated local builds.
+RUN --mount=type=cache,id=try-npm,target=/root/.npm \
+    /App/build-js.sh
 
-# Build and publish a release
-RUN dotnet publish -c Release -o out /App/src/Microsoft.TryDotNet
+# Publish only what we deploy
+RUN --mount=type=cache,id=try-nuget,target=/root/.nuget/packages \
+    dotnet publish -c Release --no-restore -o out /App/src/Microsoft.TryDotNet
 
 # Build runtime image
 FROM mcr.microsoft.com/dotnet/sdk:10.0-azurelinux3.0
@@ -31,9 +54,9 @@ WORKDIR /App
 # Make sure we run bash
 CMD ["bash"]
 
-# Make sure we get all the tools we need
-RUN tdnf install procps -y
-RUN tdnf clean all
+# Install runtime tools in a single layer
+RUN --mount=type=cache,id=try-tdnf,target=/var/cache/tdnf,sharing=locked \
+    tdnf install -y procps
 
 # Copy from build image
 COPY --from=build-env /App/out .