pr-check-lint_content.yml

name: Lint and review content files

on:
  pull_request:
    paths:
      - .github/workflows/pr-check-lint_content.yml
  pull_request_target:
    paths:
      - .nvmrc
      - "*.md"
      - "files/**/*.md"

permissions:
  # Compare commits and add reviewdog comments.
  pull-requests: write

concurrency:
  group: ci-${{ github.workflow }}-${{ github.event_name }}-${{ github.event.pull_request.number }}
  cancel-in-progress: true

jobs:
  lint-and-review-docs:
    runs-on: ubuntu-latest

    steps:
      - name: Check if PR reviews can be posted
        id: permissions
        env:
          EVENT_NAME: ${{ github.event_name }}
          HEAD_REPO: ${{ github.event.pull_request.head.repo.full_name }}
          BASE_REPO: ${{ github.repository }}
        run: |
          if [[ "${EVENT_NAME}" == "pull_request_target" || "${HEAD_REPO}" == "${BASE_REPO}" ]]; then
            echo "CAN_POST_REVIEWS=true" >> "$GITHUB_OUTPUT"
          fi

      - name: Checkout BASE
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          # explicitly specify base ref to ensure `pull_request` runs
          # behave the same as `pull_request_target` runs
          ref: ${{ github.event.pull_request.base.sha }}
          persist-credentials: false

      - name: Get changed files
        id: check
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
          BASE_SHA: ${{ github.event.pull_request.base.sha }}
          HEAD_SHA: ${{ github.event.pull_request.head.sha }}
        run: |
          # Use the GitHub API to get the list of changed files
          # documentation: https://docs.github.com/rest/commits/commits#compare-two-commits

          # Get files as newline-separated list
          FILTERED_FILES=$(gh api repos/{owner}/{repo}/compare/${BASE_SHA}...${HEAD_SHA} \
            --jq '.files | .[] | select(.status|IN("added", "modified", "renamed", "copied", "changed")) | .filename' | \
            egrep -i "^.*\.md$" || true)

          # Store as multiline output
          EOF="$(openssl rand -hex 8)"
          echo "DIFF_DOCUMENTS<<${EOF}" >> "$GITHUB_OUTPUT"
          echo "${FILTERED_FILES}" >> "$GITHUB_OUTPUT"
          echo "${EOF}" >> "$GITHUB_OUTPUT"

          # Also set a simple flag for whether we have files
          if [ -n "${FILTERED_FILES// /}" ]; then  # Remove all spaces and check if anything remains
            echo "HAS_FILES=true" >> "$GITHUB_OUTPUT"
          else
            echo "HAS_FILES=false" >> "$GITHUB_OUTPUT"
          fi

      - name: Checkout HEAD
        if: steps.check.outputs.HAS_FILES == 'true'
        uses: actions/checkout@de0fac2e4500dabe0009e67214ff5f5447ce83dd # v6.0.2
        with:
          ref: ${{ github.event.pull_request.head.sha }}
          path: pr_head
          persist-credentials: false

      - name: Get changed content from HEAD
        if: steps.check.outputs.HAS_FILES == 'true'
        run: |
          git config --global user.email "108879845+mdn-bot@users.noreply.github.com"
          git config --global user.name "mdn-bot"

          # Remove non-Markdown files from BASE.
          find files -type f ! -name '*.md' -delete
          git commit --quiet -m "Remove non-Markdown files from PR base" files

          # Remove files from BASE.
          rm -r files *.md

          # Remove non-Markdown files from HEAD.
          find pr_head -type f ! -name '*.md' -delete

          # Move Markdown files from HEAD into BASE.
          mv pr_head/files pr_head/*.md .

          # Remove HEAD checkout.
          rm -r pr_head

          # To avoid contents of PR getting into the diff that we are going to generate
          # after running the linters, here we make a dummy commit.
          # Note, this commit is not getting pushed.
          git add .
          git commit -m "Use Markdown files from PR head" .

      - name: Setup Node.js environment
        if: steps.check.outputs.HAS_FILES == 'true'
        uses: actions/setup-node@53b83947a5a98c8d113130e565377fae1a50d02f # v6.3.0
        with:
          node-version-file: ".nvmrc"
          cache: npm

      - name: Install
        if: steps.check.outputs.HAS_FILES == 'true'
        run: npm ci
        env:
          GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}

      - name: Check CRLF line endings
        id: crlf
        if: steps.check.outputs.HAS_FILES == 'true'
        env:
          DIFF_DOCUMENTS: ${{ steps.check.outputs.DIFF_DOCUMENTS }}
        run: |
          EOF="$(openssl rand -hex 8)"
          readarray -t files_to_lint <<< "$DIFF_DOCUMENTS"
          printf "Files to process (%d files):\n" "${#files_to_lint[@]}"
          printf "'%s'\n" "${files_to_lint[@]}"

          CRLF_FAILED=true
          CRLF_LOG=$(git ls-files --eol "${files_to_lint[@]}" | grep -E 'w/(mixed|crlf)') || CRLF_FAILED=false
          echo "CRLF_LOG<<${EOF}" >> "$GITHUB_OUTPUT"
          echo "${CRLF_LOG}" >> "$GITHUB_OUTPUT"
          echo "${EOF}" >> "$GITHUB_OUTPUT"
          echo "CRLF_FAILED=${CRLF_FAILED}" >> "$GITHUB_OUTPUT"

      - name: Diff
        if: steps.check.outputs.HAS_FILES == 'true'
        run: |
          git status
          git diff HEAD --color=always --word-diff --word-diff-regex='[[:alnum:]_]+'

      - name: Run markdownlint
        id: markdownlint
        if: steps.check.outputs.HAS_FILES == 'true'
        env:
          DIFF_DOCUMENTS: ${{ steps.check.outputs.DIFF_DOCUMENTS }}
        run: |
          EOF="$(openssl rand -hex 8)"
          readarray -t files_to_lint <<< "$DIFF_DOCUMENTS"

          MD_LINT_FAILED=false
          MD_LINT_LOG=$(npx markdownlint-cli2 --fix "${files_to_lint[@]}" 2>&1) || MD_LINT_FAILED=true
          echo "MD_LINT_LOG<<${EOF}" >> "$GITHUB_OUTPUT"
          echo "${MD_LINT_LOG}" >> "$GITHUB_OUTPUT"
          echo "${EOF}" >> "$GITHUB_OUTPUT"
          echo "MD_LINT_FAILED=${MD_LINT_FAILED}" >> "$GITHUB_OUTPUT"

      - name: Diff
        if: steps.check.outputs.HAS_FILES == 'true'
        run: |
          git status
          git diff HEAD --color=always --word-diff --word-diff-regex='[[:alnum:]_]+'

      - name: Lint front-matter
        id: frontmatter
        if: steps.check.outputs.HAS_FILES == 'true'
        env:
          DIFF_DOCUMENTS: ${{ steps.check.outputs.DIFF_DOCUMENTS }}
        run: |
          EOF="$(openssl rand -hex 8)"
          readarray -t files_to_lint <<< "$DIFF_DOCUMENTS"

          FM_LINT_FAILED=false
          FM_LINT_LOG=$(node scripts/front-matter_linter.js --fix true "${files_to_lint[@]}" 2>&1) || FM_LINT_FAILED=true
          echo "FM_LINT_LOG<<${EOF}" >> "$GITHUB_OUTPUT"
          echo "${FM_LINT_LOG}" >> "$GITHUB_OUTPUT"
          echo "${EOF}" >> "$GITHUB_OUTPUT"
          echo "FM_LINT_FAILED=${FM_LINT_FAILED}" >> "$GITHUB_OUTPUT"

      - name: Diff
        if: steps.check.outputs.HAS_FILES == 'true'
        run: |
          git status
          git diff HEAD --color=always --word-diff --word-diff-regex='[[:alnum:]_]+'

      - name: Run Prettier
        id: prettier
        if: steps.check.outputs.HAS_FILES == 'true'
        env:
          DIFF_DOCUMENTS: ${{ steps.check.outputs.DIFF_DOCUMENTS }}
        run: |
          EOF="$(openssl rand -hex 8)"
          readarray -t files_to_lint <<< "$DIFF_DOCUMENTS"

          PRETTIER_FAILED=false
          PRETTIER_LOG=$(npx prettier --check "${files_to_lint[@]}" 2>&1) || PRETTIER_FAILED=true
          echo "PRETTIER_LOG<<${EOF}" >> "$GITHUB_OUTPUT"
          echo "${PRETTIER_LOG}" >> "$GITHUB_OUTPUT"
          echo "${EOF}" >> "$GITHUB_OUTPUT"
          echo "PRETTIER_FAILED=${PRETTIER_FAILED}" >> "$GITHUB_OUTPUT"
          npx prettier -w "${files_to_lint[@]}"

      - name: Diff
        id: lint-diff
        if: always() && steps.check.outputs.HAS_FILES == 'true'
        run: |
          git status
          git diff HEAD --color=always --word-diff --word-diff-regex='[[:alnum:]_]+'

          if [[ -n $(git diff) ]]; then
            echo "FILES_MODIFIED=true" >> "$GITHUB_OUTPUT"
          fi

      - name: Setup reviewdog
        if: (steps.lint-diff.outputs.FILES_MODIFIED == 'true' || steps.markdownlint.outputs.MD_LINT_FAILED == 'true') && steps.permissions.outputs.CAN_POST_REVIEWS == 'true'
        uses: reviewdog/action-setup@d8a7baabd7f3e8544ee4dbde3ee41d0011c3a93f # v1.5.0
        with:
          reviewdog_version: latest

      - name: Suggest changes using diff
        if: steps.lint-diff.outputs.FILES_MODIFIED == 'true' && steps.permissions.outputs.CAN_POST_REVIEWS == 'true'
        env:
          REVIEWDOG_GITHUB_API_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          TMPFILE=$(mktemp)
          git diff >"${TMPFILE}"
          git stash -u && git stash drop
          reviewdog \
            -name="mdn-linter" \
            -f=diff \
            -f.diff.strip=1 \
            -filter-mode=diff_context \
            -reporter=github-pr-review < "${TMPFILE}"

      - name: Add reviews for markdownlint errors
        if: steps.markdownlint.outputs.MD_LINT_FAILED == 'true' && steps.permissions.outputs.CAN_POST_REVIEWS == 'true'
        env:
          MD_LINT_LOG: ${{ steps.markdownlint.outputs.MD_LINT_LOG }}
          REVIEWDOG_GITHUB_API_TOKEN: ${{ secrets.GITHUB_TOKEN }}
        run: |
          echo "${MD_LINT_LOG}" | \
            reviewdog \
            -efm="%f:%l:%c %m" \
            -efm="%f:%l %m" \
            -name="markdownlint" \
            -diff="git diff" \
            -reporter="github-pr-review"

      - name: Fail if any issues pending
        if: steps.lint-diff.outputs.FILES_MODIFIED == 'true' || steps.crlf.outputs.CRLF_FAILED == 'true' || steps.markdownlint.outputs.MD_LINT_FAILED == 'true' || steps.frontmatter.outputs.FM_LINT_FAILED == 'true'
        env:
          CRLF_FAILED: ${{ steps.crlf.outputs.CRLF_FAILED }}
          MD_LINT_FAILED: ${{ steps.markdownlint.outputs.MD_LINT_FAILED }}
          FM_LINT_FAILED: ${{ steps.frontmatter.outputs.FM_LINT_FAILED }}
          PRETTIER_FAILED: ${{ steps.prettier.outputs.PRETTIER_FAILED }}
          CRLF_LOG: ${{ steps.crlf.outputs.CRLF_LOG }}
          MD_LINT_LOG: ${{ steps.markdownlint.outputs.MD_LINT_LOG }}
          FM_LINT_LOG: ${{ steps.frontmatter.outputs.FM_LINT_LOG }}
          PRETTIER_LOG: ${{ steps.prettier.outputs.PRETTIER_LOG }}
        run: |
          echo -e "\nPlease fix all the linting issues mentioned in the following logs and in the PR review comments."

          if [[ ${CRLF_FAILED} == 'true' ]]; then
            echo -e "\n\n🪵 In the following files make sure all the lines end with only Line Feed (LF) character and not with Carriage Return Line Feed (CRLF) characters:"
            echo "${CRLF_LOG}"
            echo "For more information refer https://gist.github.com/LunarLambda/3df0840b336a5e314e4ffdac03cbf619 ."
            echo "You may use https://app.execeratics.com/LFandCRLFonline/?l=en online tool to convert line endings from CRLF to LF."
          fi

          if [[ ${MD_LINT_FAILED} == 'true' ]]; then
            echo -e "\n\n🪵 Logs from markdownlint:"
            echo "${MD_LINT_LOG}"
          fi

          if [[ ${FM_LINT_FAILED} == 'true' ]]; then
            echo -e "\n\n🪵 Logs from front-matter linter:"
            echo "${FM_LINT_LOG}"
          fi

          if [[ ${PRETTIER_FAILED} == 'true' ]]; then
            echo -e "\n\n🪵 Logs from Prettier formatter:"
            echo "${PRETTIER_LOG}"
            echo -e "\nYou can use Prettier playground to format the files online (configuration pre-filled): https://prettier.io/playground/#N4Igxg9gdgLgprEAuEBiABABwIYGd7owAWc6CMAlgE6kBmFANqSTSADQgSaXS7KjYqVCAHcACoIR8U2BiOwBPPhwBGVbGADWcGAGVsAWzgAZClDjIYVAK5xV6rTt04wZgOaWbdkLjgGKnrYccAAemHBUFEawsgAqEVCCFHDStLK+HLjuTACK1hDwyGkMGSAAVrghutlweQUWSMWlAI758GLCmNIgeAC05nAAJkPsIFbYjO4AwhAGBtjIPQwMo1lQbkwAgjBWFCrW7RGm5kXp3kQwBgwA6kQU8LgucLpS9xQAbvcKi2C4yiDvWwASSgw1gujAkW4m1BuhgCiYpxK3kwwl813UmEWqJSEXeFg4Zl8VBgHWwbnmSNKOCoxMW8yomkGoigo1RZhg1wog2IyAAHAAGDg0VrUOBkikLRpnDgwbAqLk8ojIABMHGsvli8tSMpAfhUQ2Gg2M2HW1nJcAAYhAqPMdu5FtgDhAQABfV1AA \n"
          fi

          exit 1