Skins/.gitea/workflows/test-skins.yml

name: Test Skins

on:
  push:
  pull_request:

jobs:
  link-check:
    runs-on: ubuntu-latest

    steps:
      - uses: actions/checkout@v4

      - name: Validate links and assets
        shell: bash
        run: |
          set -uo pipefail

          RED="\033[31m"
          GREEN="\033[32m"
          RESET="\033[0m"

          ERRORS=()

          ############################################################
          # Remove all NUL/control characters BEFORE any decoding
          ############################################################
          sanitize_url() {
              printf '%s' "$1" \
                | tr -d '\000' \
                | tr -d '\r' \
                | sed 's/[[:cntrl:]]//g'
          }

          ############################################################
          # Safe unicode URL decode (no null bytes)
          ############################################################
          urldecode() {
              local url="$1"
              url="${url//+/ }"
              printf '%s' "$url" | perl -pe 's/%([A-Fa-f0-9]{2})/chr(hex($1))/eg'
          }

          ############################################################
          # Detect Gitea "soft 404" even when status=200 OK
          ############################################################
          check_http() {
              local url="$1"
              echo "      → Checking external: $url"

              local status body

              # Try HEAD first
              status=$(curl -Is --max-time 10 "$url" | head -n1 | awk '{print $2}')
              if [[ "$status" =~ ^2|3 ]]; then
                  body=$(curl -Ls --max-time 10 "$url")
                  if echo "$body" | grep -qiE "404 Not Found|Path .* doesn't exist|File not found|does not exist|Not Found"; then
                      return 1
                  fi
                  return 0
              fi

              # Try GET fallback
              status=$(curl -Is --max-time 10 -X GET "$url" | head -n1 | awk '{print $2}')
              if [[ "$status" =~ ^2|3 ]]; then
                  body=$(curl -Ls --max-time 10 "$url")
                  if echo "$body" | grep -qiE "404 Not Found|Path .* doesn't exist|File not found|does not exist|Not Found"; then
                      return 1
                  fi
                  return 0
              fi

              return 1
          }

          ############################################################
          # Local file existence check
          ############################################################
          check_local() {
              local path="$1"
              path="${path#/}"  # strip leading slash
              local decoded
              decoded=$(urldecode "$path")

              echo "      → Checking local: $decoded"

              [[ -e "$decoded" ]]
          }

          ############################################################
          # Extract URLs from Markdown + HTML
          ############################################################
          extract_links() {
              local f="$1"

              # Markdown links
              grep -oE '\[[^]]*\]\([^)]*\)' "$f" \
                  | sed -E 's/.*\((.*)\).*/\1/'

              # Markdown images
              grep -oE '!\[[^]]*\]\([^)]*\)' "$f" \
                  | sed -E 's/.*\((.*)\).*/\1/'

              # Raw URLs
              grep -oE 'https?://[^ )"]+' "$f"

              # <img src="">
              grep -oE '<img[^>]*src="[^"]+"' "$f" \
                  | sed -E 's/.*src="([^"]*)".*/\1/'

              # <video src="">
              grep -oE '<video[^>]*src="[^"]+"' "$f" \
                  | sed -E 's/.*src="([^"]*)".*/\1/'
          }

          echo "🔍 Scanning Markdown files..."
          echo

          find . -type f -name '*.md' | while IFS= read -r mdfile; do
              echo "📄 Checking: $mdfile"

              while IFS= read -r url; do
                  [[ -z "$url" ]] && continue
                  [[ "$url" == mailto:* ]] && continue

                  # Clean input BEFORE use
                  cleaned=$(sanitize_url "$url")

                  # Skip tag-version links in docs
                  if [[ "$mdfile" == ./docs/* ]] && [[ "$cleaned" == *"/src/tag/"* ]]; then
                      continue
                  fi

                  # external
                  if [[ "$cleaned" == http* ]]; then
                      if ! check_http "$cleaned"; then
                          ERRORS+=("❌ Broken external link: $cleaned (in $mdfile)")
                      fi

                  # local file
                  else
                      if ! check_local "$cleaned"; then
                          ERRORS+=("❌ Missing local file: $cleaned (in $mdfile)")
                      fi
                  fi

              done < <(extract_links "$mdfile")

              echo
          done

          ############################################################
          # Final result
          ############################################################
          echo
          if (( ${#ERRORS[@]} > 0 )); then
              echo -e "${RED}✖ Errors found:${RESET}"
              printf "%s\n" "${ERRORS[@]}"

              echo
              echo -e "${RED}❌ Failing job because broken links were found.${RESET}"
              exit 1
          else
              echo -e "${GREEN}✔ All links OK!${RESET}"
          fi