Single build task for each OS for CI (#2026)

.github/workflows/ci.yml

@@ -102,7 +102,7 @@ jobs:
             echo "$NON_CLAUDE_FILES"
           fi
 
-  test:
+  build:
     needs: [check-changes]
     if: needs.check-changes.outputs.should_run_tests == 'true'
     # Why Mac and Windows?
@@ -118,10 +118,6 @@ jobs:
             { name: "windows", image: "windows-latest" },
             { name: "macos", image: "macos-latest" },
           ]
-        # If you update this, you need to update
-        # scripts/generate-playwright-summary.js
-        shard: [1, 2, 3, 4]
-        shardTotal: [4]
     runs-on: ${{ matrix.os.image }}
     steps:
       - name: Checkout code
@@ -142,19 +138,75 @@ jobs:
           GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
       - name: Presubmit check (e.g. lint, format)
         # do not run this on Windows (it fails and not necessary)
-        # Only run on shard 1 to avoid redundant execution
-        if: contains(matrix.os.name, 'macos') && matrix.shard == 1
+        if: contains(matrix.os.name, 'macos')
         run: npm run presubmit
       - name: Type-checking
-        # do not run this on windows (it's redunant)
-        # Only run on shard 1 to avoid redundant execution
-        if: contains(matrix.os.name, 'macos') && matrix.shard == 1
+        # do not run this on windows (it's redundant)
+        if: contains(matrix.os.name, 'macos')
         run: npm run ts
       - name: Unit tests
-        # do not run this on windows (it's redunant)
-        # Only run on shard 1 to avoid redundant execution
-        if: contains(matrix.os.name, 'macos') && matrix.shard == 1
+        # do not run this on windows (it's redundant)
+        if: contains(matrix.os.name, 'macos')
         run: npm run test
+      - name: Build
+        env:
+          NODE_OPTIONS: "--max-old-space-size=4096"
+        run: npm run pre:e2e
+      - name: Prep test server
+        run: cd testing/fake-llm-server && npm install && npm run build && cd -
+      # macOS: Archive app bundle with tar to preserve metadata (code signatures, extended attributes)
+      # This is necessary because GitHub artifact upload/download loses these attributes
+      - name: Archive macOS app bundle
+        if: contains(matrix.os.name, 'macos')
+        run: tar -cvf out-macos.tar out/
+      - name: Upload build artifacts (macOS)
+        if: contains(matrix.os.name, 'macos')
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        with:
+          name: build-${{ matrix.os.name }}
+          path: |
+            out-macos.tar
+            scaffold/
+            testing/fake-llm-server/
+          retention-days: 1
+      - name: Upload build artifacts (Windows)
+        if: contains(matrix.os.name, 'windows')
+        uses: actions/upload-artifact@ea165f8d65b6e75b540449e92b4886f43607fa02 # v4.6.2
+        with:
+          name: build-${{ matrix.os.name }}
+          path: |
+            out/
+            scaffold/
+            testing/fake-llm-server/
+          retention-days: 1
+
+  e2e-tests:
+    needs: [build]
+    strategy:
+      fail-fast: false
+      matrix:
+        os:
+          [
+            { name: "windows", image: "windows-latest" },
+            { name: "macos", image: "macos-latest" },
+          ]
+        shard: [1, 2, 3, 4]
+        shardTotal: [4]
+    runs-on: ${{ matrix.os.image }}
+    steps:
+      - name: Checkout code
+        uses: actions/checkout@v4
+      - name: Initialize environment
+        uses: actions/setup-node@v4
+        with:
+          node-version-file: package.json
+          cache: npm
+          cache-dependency-path: package-lock.json
+      - name: Install node modules
+        run: npm ci --no-audit --no-fund --progress=false
+        env:
+          # Required for @vscode/ripgrep to download binaries without hitting GitHub API rate limits
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
       - name: Setup pnpm
         uses: pnpm/action-setup@a7487c7e89a18df4991f7f222e4898a00d66ddda # v4.1.0
         with:
@@ -178,14 +230,16 @@ jobs:
         run: cd scaffold && pnpm install
       - name: Install nextjs-template dependencies
         run: cd nextjs-template && pnpm install
+      - name: Download build artifacts
+        uses: actions/download-artifact@v4
+        with:
+          name: build-${{ matrix.os.name }}
+      # macOS: Extract tar archive to restore app bundle with preserved metadata
+      - name: Extract macOS app bundle
+        if: contains(matrix.os.name, 'macos')
+        run: tar -xvf out-macos.tar
       - name: Install Chromium browser for Playwright
         run: npx playwright install chromium --with-deps
-      - name: Build
-        env:
-          NODE_OPTIONS: "--max-old-space-size=4096"
-        run: npm run pre:e2e
-      - name: Prep test server
-        run: cd testing/fake-llm-server && npm install && npm run build && cd -
       - name: E2E tests (Shard ${{ matrix.shard }}/4)
         # You can add debug logging to make it easier to see what's failing
         # by adding "DEBUG=pw:browser" in front.
@@ -203,7 +257,7 @@ jobs:
     # Merge reports after playwright-tests, even if some shards have failed
     # Skip if tests were skipped (only .claude files changed)
     if: ${{ !cancelled() && needs.check-changes.outputs.should_run_tests == 'true' }}
-    needs: [check-changes, test]
+    needs: [check-changes, e2e-tests]
     permissions:
       contents: read
       pull-requests: write