sync-with-cpython.yml

name: Sync with CPython

on:
  schedule:
    - cron: '0 4 * * *'
  workflow_dispatch:
    inputs:
      cpython_tag:
        description: 'CPython tag/branch to sync against (e.g. 3.14)'
        required: false
        default: '3.14'

permissions:
  contents: write
  issues: write

# Shared with maintenance.yml so the two never push at the same time.
concurrency:
  group: repo-writes
  cancel-in-progress: false

jobs:
  sync:
    runs-on: ubuntu-latest
    timeout-minutes: 45
    env:
      CPYTHON_TAG: ${{ github.event.inputs.cpython_tag || '3.14' }}
    steps:
      - name: Checkout translation repo
        uses: actions/checkout@v4
        with:
          fetch-depth: 0

      # --- Skip the expensive build if upstream hasn't moved -------------
      - name: Get upstream head
        id: upstream
        run: |
          sha=$(git ls-remote https://github.com/python/cpython \
                  "refs/heads/$CPYTHON_TAG" "refs/tags/$CPYTHON_TAG" \
                | head -1 | cut -f1)
          echo "sha=$sha" >> "$GITHUB_OUTPUT"
          # Marker lives outside the repo so `git add --all` never sees it.
          echo "$sha" > "$RUNNER_TEMP/sync-marker"

      - name: Check whether this upstream head was already synced
        id: seen
        # Fail open: no SHA -> no lookup -> full run.
        if: steps.upstream.outputs.sha != ''
        uses: actions/cache/restore@v4
        with:
          path: ${{ runner.temp }}/sync-marker
          key: synced-${{ env.CPYTHON_TAG }}-${{ steps.upstream.outputs.sha }}
          lookup-only: true

      - name: Decide whether to run
        run: |
          if [ "${{ steps.seen.outputs.cache-hit }}" = "true" ] \
             && [ "${{ github.event_name }}" != "workflow_dispatch" ]; then
            echo "Upstream ${{ steps.upstream.outputs.sha }} already synced; skipping."
            echo "SKIP=true" >> "$GITHUB_ENV"
          fi

      # --- Sync ------------------------------------------------------------
      - name: Set up Python
        if: env.SKIP != 'true'
        uses: actions/setup-python@v4
        with:
          python-version: '3.12'

      - name: Install gettext
        if: env.SKIP != 'true'
        run: sudo apt-get install -y gettext

      - name: Install Python dependencies
        if: env.SKIP != 'true'
        # Consider installing from CPython's Doc/requirements.txt instead of
        # unpinned sphinx, so Sphinx upgrades can't change msgids on their own.
        run: pip install sphinx sphinx-intl

      - name: Sync msgids (fetch, build gettext, merge, validate)
        if: env.SKIP != 'true'
        run: python scripts/po_sync.py sync-only "$CPYTHON_TAG" --create-new

      - name: Stage changes
        if: env.SKIP != 'true'
        run: git add --all

      - name: Unstage POT-Creation-Date-only changes
        if: env.SKIP != 'true'
        run: python scripts/unstage_cosmetic.py

      # --- Build the report from what is actually staged -------------------
      - name: Build review report (new files + newly fuzzy strings)
        if: env.SKIP != 'true'
        run: |
          report="$RUNNER_TEMP/report.md"
          : > "$report"
          work=$(mktemp -d)

          # New PO files
          new_files=$(git diff --staged --name-only --diff-filter=A -- '*.po')
          if [ -n "$new_files" ]; then
            {
              echo "## 🆕 New PO files ($(echo "$new_files" | wc -l))"
              echo "$new_files" | sed 's/^/- `/; s/$/`/'
              echo ""
            } >> "$report"
          fi

          # Fuzzy entries that are new in this run (present now, absent at HEAD)
          fuzzy_out="$work/fuzzy.md"
          : > "$fuzzy_out"
          git diff --staged --name-only --diff-filter=M -- '*.po' |
          while IFS= read -r f; do
            git show "HEAD:$f" > "$work/old_full.po"
            msgattrib --only-fuzzy --no-obsolete --no-wrap "$work/old_full.po" > "$work/old.po" 2>/dev/null || : > "$work/old.po"
            msgattrib --only-fuzzy --no-obsolete --no-wrap "$f"                > "$work/new.po" 2>/dev/null || : > "$work/new.po"
            # entries in exactly one of old/new, then keep those also in new
            msgcomm --less-than=2 --no-wrap "$work/old.po" "$work/new.po" > "$work/sym.po"   2>/dev/null || : > "$work/sym.po"
            msgcomm --no-wrap "$work/new.po" "$work/sym.po"                > "$work/added.po" 2>/dev/null || : > "$work/added.po"
            count=$(grep -c '^#, .*fuzzy' "$work/added.po" || true)
            if [ "${count:-0}" -gt 0 ]; then
              {
                echo "### \`$f\` ($count new fuzzy)"
                grep -E '^msgid ".+"' "$work/added.po" | sed 's/^msgid /- /' | head -10
                [ "$count" -gt 10 ] && echo "- … and $((count - 10)) more"
                echo ""
              } >> "$fuzzy_out"
            fi
          done

          if [ -s "$fuzzy_out" ]; then
            { echo "## 🔍 Newly fuzzy strings"; cat "$fuzzy_out"; } >> "$report"
          fi

      # --- Commit + push ---------------------------------------------------
      - name: Commit and push if changed
        id: commit
        if: env.SKIP != 'true'
        run: |
          git config user.name "github-actions[bot]"
          git config user.email "github-actions[bot]@users.noreply.github.com"
          if git diff --staged --quiet; then
            echo "Nothing to commit."
            echo "committed=false" >> "$GITHUB_OUTPUT"
            exit 0
          fi
          git commit -m "chore: sync msgids with CPython $CPYTHON_TAG"
          git pull --rebase
          git push
          echo "committed=true" >> "$GITHUB_OUTPUT"

      # Only after a successful push, so a failed run can't create duplicates.
      - name: Open issue for new files / newly fuzzy strings
        if: env.SKIP != 'true' && steps.commit.outputs.committed == 'true'
        env:
          GH_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          report="$RUNNER_TEMP/report.md"
          if [ -s "$report" ]; then
            gh label create "translation" \
              --color "0075ca" \
              --description "Translation review needed" \
              --repo "${{ github.repository }}" 2>/dev/null || true
            echo "Triggered by sync with \`$CPYTHON_TAG\` (commit $(git rev-parse --short HEAD))." >> "$report"
            gh issue create \
              --title "🔍 Translation review needed after CPython sync ($(date +%Y-%m-%d))" \
              --body-file "$report" \
              --label "translation" \
              --repo "${{ github.repository }}"
          fi

      # Record the upstream head as synced (only reached if everything above passed).
      - name: Remember synced upstream head
        if: env.SKIP != 'true' && steps.upstream.outputs.sha != ''
        uses: actions/cache/save@v4
        with:
          path: ${{ runner.temp }}/sync-marker
          key: synced-${{ env.CPYTHON_TAG }}-${{ steps.upstream.outputs.sha }}