use github api to get diff file and changed files

.github/workflows/pr-test.yml

@@ -14,56 +14,52 @@ on:
 jobs:
   tests:
     runs-on: ubuntu-latest
+    env:
+      CHANGED_FILES: /tmp/changed-files.txt
+      BASE_SHA: ${{ github.event.pull_request.base.sha }}
+      HEAD_SHA: ${{ github.event.pull_request.head.sha }}
 
     steps:
-      - name: Diffset
-        id: diffset
-        uses: softprops/diffset@v1
-        with:
-          base: ${{ github.base_ref }}
-        env:
-          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
-
-      - name: Check diffset
-        run: |
-          # document changes
-          GIT_DIFF_CONTENT=$(echo "${{ steps.diffset.outputs.diffset }}" | tr -s " " "\n" | egrep -i "^files/.*/.*\.(html|md)$" | tr "\n" " ")
-          echo "GIT_DIFF_CONTENT=$GIT_DIFF" >> $GITHUB_ENV
-          echo "changed documents: $GIT_DIFF_CONTENT"
-          # image changes
-          GIT_DIFF_IMAGE=$(echo "${{ steps.diffset.outputs.diffset }}" | tr -s " " "\n" | egrep -i "^files/.*/.*\.(png|jpeg|jpg|gif|svg|webp)$" | tr "\n" " ")
-          echo "GIT_DIFF_IMAGE=$GIT_DIFF_IMAGE" >> $GITHUB_ENV
-          echo "changed images: $GIT_DIFF_IMAGE"
-          # set global flag if there are any changes
-          if [ ! -z "$GIT_DIFF_CONTENT" ] || [ ! -z "$GIT_DIFF_IMAGE" ]; then
-            echo "FILE_CHANGE=1" >> $GITHUB_ENV
-          fi
-
       - uses: actions/checkout@v3
-        if: env.FILE_CHANGE
+        # Needed otherswise it will check out a merge commit which means
+        # you can't get a `git diff` for this PR compared to its base.
+        # with:
+        #   ref: ${{ github.event.pull_request.head.sha }}
 
       - uses: actions/checkout@v3
-        if: env.FILE_CHANGE
         with:
           repository: mdn/content
           path: mdn/content
 
       - name: Setup Node.js environment
-        if: env.FILE_CHANGE
         uses: actions/setup-node@v3
         with:
           node-version: "16"
           cache: "yarn"
           cache-dependency-path: mdn/content/yarn.lock
 
       - name: Install all yarn packages
-        if: env.FILE_CHANGE
         working-directory: ${{ github.workspace }}/mdn/content
         run: |
           yarn --frozen-lockfile
 
+      - name: Get Changed files
+        run: |
+          # Use the GitHub API to get the list of changed files
+          # documenation: https://docs.github.com/rest/commits/commits#compare-two-commits
+          gh api repos/{owner}/{repo}/compare/${{ env.BASE_SHA }}...${{ env.HEAD_SHA }} \
+            --paginate \
+            --jq '.files | .[] | select(.status|IN("added", "modified", "renamed", "copied", "changed")) | .filename' > ${{ env.CHANGED_FILES }}
+        env:
+          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
+
+      - name: Get changed documents
+        run: |
+          DIFF_DOCUMENTS=$(cat ${{ env.CHANGED_FILES }} | egrep -i "^files/.*/.*\.(html|md)$" | tr "\n" " ")
+          echo "GIT_DIFF_CONTENT=$DIFF_DOCUMENTS" >> $GITHUB_ENV
+
       - name: Build changed content
-        if: env.GIT_DIFF_CONTENT
+        if: ${{ env.GIT_DIFF_CONTENT }}
         env:
           CONTENT_ROOT: ${{ github.workspace }}/mdn/content/files
           CONTENT_TRANSLATED_ROOT: ${{ github.workspace }}/files
@@ -128,7 +124,7 @@ jobs:
           # directory.
           # The purpose of this is for the PR Review Companion to later
           # be able to use this raw diff file for the benefit of analyzing.
-          gh pr diff ${{ github.event.number }} > $BUILD_OUT_ROOT/DIFF
+          wget https://github.com/${{ github.action_repository }}/compare/${{ env.BASE_SHA }}...${{ env.HEAD_SHA }}.diff -O $BUILD_OUT_ROOT/DIFF
 
       - name: Merge static assets with built documents
         if: ${{ env.GIT_DIFF_CONTENT }}
@@ -159,9 +155,9 @@ jobs:
 
       - name: Get changed images
         if: ${{ env.GIT_DIFF_CONTENT }}
-        # note that we should get the absolute path of the changed images
         run: |
-          DIFF_IMAGES=$(echo ${{ env.GIT_DIFF_IMAGES }} | | tr -s " " "\n" | xargs readlink -e | tr "\n" " ")
+          # note that we should get the absolute path of the changed images
+          DIFF_IMAGES=$(cat ${{ env.CHANGED_FILES }} | egrep -i "^files/.*/.*\.(png|jpeg|jpg|gif|svg|webp)$" | xargs readlink -e | tr "\n" " ")
           echo "GIT_DIFF_IMAGES=$DIFF_IMAGES" >> $GITHUB_ENV
 
       - name: Check changed files