[site] Optimize Hugo CI workflow with caching and updates

.github/workflows/hugo.yml

@@ -3,16 +3,62 @@ name: test hugo
 on: [pull_request]
 
 env:
-  NODE_VERSION: 16
+  NODE_VERSION: 22
   HUGO_VERSION: 0.126.1
 
 jobs:
   lint:
+    outputs:
+      source_content_hash: ${{ steps.calc-hash.outputs.result }}
+      changed_files_key_part: ${{ steps.changed-files-key.outputs.key_part }}
+      debug_changed_files_list: ${{ steps.changed-files-key.outputs.debug_changed_files_list }}
     runs-on: ubuntu-latest
 
     steps:
       - name: Checkout
-        uses: actions/checkout@v3
+        uses: actions/checkout@v4
+        with:
+          submodules: true # Fetch Hugo themes (true OR recursive)
+          fetch-depth: 2 # Fetch enough history for HEAD~1
+
+      - name: Get changed files key
+        id: changed-files-key
+        run: |
+          # Get list of changed files, sort them, then hash the list.
+          git diff --name-only HEAD~1 HEAD | LC_ALL=C sort | md5sum | head -c 32 > changed_files_key_part.txt
+          echo "key_part=$(cat changed_files_key_part.txt)" >> $GITHUB_OUTPUT
+          ACTUAL_FILES=$(git diff --name-only HEAD~1 HEAD | LC_ALL=C sort | tr '\n' ' ' | sed 's/ $//')
+          echo "debug_changed_files_list=${ACTUAL_FILES}" >> $GITHUB_OUTPUT
+
+      - name: Calc hash
+        id: calc-hash
+        run: |
+          # Calculate md5sum for each file, sort the output (for consistent ordering),
+          # then calculate md5sum of that list of hashes. head -c 32 takes the md5 hash string.
+          find . -path ./.git -prune -o -type f -print0 | xargs -0 md5sum | LC_ALL=C sort -k 2 | md5sum | head -c 32 > hash.txt
+          hash_value=$(cat hash.txt)
+          echo "result=$hash_value" >> $GITHUB_OUTPUT
+
+      - name: Debug Cache Key
+        run: |
+          echo "calc-hash output: ${{ steps.calc-hash.outputs.result }}"
+          echo "changed-files key part: ${{ steps.changed-files-key.outputs.key_part }}"
+          echo "changed-files actual list: ${{ steps.changed-files-key.outputs.debug_changed_files_list }}"
+
+      - name: Upload Hugo source artifact
+        uses: actions/upload-artifact@v4
+        with:
+          # Cache only the specific directories and files needed for the Hugo build, and not capturing the full .git folder.
+          path: |
+            archetypes
+            assets
+            config
+            content
+            data
+            layouts
+            static
+            themes
+          name: hugo-source-files
 
       - name: Lint Filenames
         uses: julie-ng/lowercase-linter@v1
@@ -22,15 +68,35 @@ jobs:
           path: "."
           pr-comment: true
           repo-token: ${{ secrets.GITHUB_TOKEN }}
+
   test-linux:
     needs: lint
     name: Build hugo on Linux
     runs-on: ubuntu-latest
     steps:
-      - uses: actions/checkout@v3
+      - name: Download Hugo source artifact
+        uses: actions/download-artifact@v4
         with:
-          submodules: true # Fetch Hugo themes (true OR recursive)
-          fetch-depth: 0 # Fetch all history for .GitInfo and .Lastmod
+          name: hugo-source-files
+          path: . # Download to the root of the workspace
+
+      - name: Debug Cache Key
+        run: |
+          echo "Using source_content_hash from lint job: ${{ needs.lint.outputs.source_content_hash }}"
+          echo "Using changed_files_key_part from lint job: ${{ needs.lint.outputs.changed_files_key_part }}"
+          echo "Actual changed files list from lint job: ${{ needs.lint.outputs.debug_changed_files_list }}"
+
+      - name: Cache Hugo build outputs
+        id: cache-hugo-build
+        uses: actions/cache@v4
+        with:
+          # Cache Hugo's generated 'public' directory and 'resources' directory
+          path: |
+            public
+            resources
+          key: ${{ runner.os }}-hugo-build-${{ needs.lint.outputs.source_content_hash }}-${{ needs.lint.outputs.changed_files_key_part }}
+          restore-keys: |
+            ${{ runner.os }}-hugo-build-${{ needs.lint.outputs.source_content_hash }}-
 
       - name: Setup Hugo
         uses: peaceiris/actions-hugo@v3
@@ -39,7 +105,11 @@ jobs:
           extended: true
 
       - name: Build Hugo
-        run: hugo --environment=production --minify --templateMetrics --logLevel info
+        # If cache hit for Hugo build outputs, this step might still run but Hugo might determine nothing needs to be rebuilt.
+        # For a more explicit skip, you could check `steps.cache-hugo-build.outputs.cache-hit`
+        # if: steps.cache-hugo-build.outputs.cache-hit != 'true'
+        run: |
+          hugo --environment=production --minify --templateMetrics --logLevel info
 
   # test-windows:
   #   needs: lint