zvec/.github/workflows/clang_tidy.yml

name: Clang-Tidy

on:
  workflow_call:
  workflow_dispatch:

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

permissions:
  contents: read

jobs:
  clang_tidy:
    name: Clang-Tidy Checks
    runs-on: ubuntu-24.04
    steps:
      - name: Checkout code
        uses: actions/checkout@v6
        with:
          submodules: recursive
          # fetch-depth: 0 is required for tj-actions/changed-files to correctly
          # compute the diff against the base branch on pull_request events.
          fetch-depth: 0

      - name: Install dependencies
        run: |
          sudo apt-get update
          sudo apt-get install -y clang-tidy=1:18.0-59~exp2 cmake ninja-build libomp-dev          

      - name: Configure CMake and export compile commands
        run: |
          cmake -S . -B build -G Ninja \
            -DCMAKE_EXPORT_COMPILE_COMMANDS=ON \
            -DBUILD_TOOLS=ON          

      - name: Collect changed C/C++ files
        id: changed_files
        uses: tj-actions/changed-files@v47
        with:
          files: |
            **/*.c
            **/*.cc
            **/*.cpp
            **/*.cxx            
          files_ignore: |
            thirdparty/**
            build/**            

      - name: Filter changed files against compile_commands.json
        id: tidy_files
        if: steps.changed_files.outputs.any_changed == 'true'
        run: |
          python3 - <<'PY'
          import json
          import os
          from pathlib import Path

          # all_changed_files is space-separated when output_format is not set;
          # tj-actions v46 defaults to space-separated for the env var form.
          # We read from the file written by the action instead via GITHUB_OUTPUT,
          # but the safest approach is to use the JSON output format.
          raw = os.environ.get("ALL_CHANGED_FILES", "")
          changed = [f for f in raw.split() if f]

          compile_db_path = Path("build/compile_commands.json")
          with compile_db_path.open("r", encoding="utf-8") as fh:
              compile_db = json.load(fh)

          # Build a set of absolute, normalised paths from compile_commands.json.
          compile_entries = set()
          for entry in compile_db:
              file_field = entry.get("file", "")
              if file_field:
                  compile_entries.add(os.path.normpath(file_field))

          cwd = Path.cwd().resolve()
          selected = []
          skipped = []

          for rel_path in changed:
              abs_path = os.path.normpath(str((cwd / rel_path).resolve()))
              if abs_path in compile_entries:
                  selected.append(rel_path)
              elif not Path(rel_path).is_file():
                  skipped.append(f"{rel_path} (file not found)")
              else:
                  skipped.append(f"{rel_path} (not in compile_commands.json)")

          github_output = os.environ["GITHUB_OUTPUT"]
          with open(github_output, "a", encoding="utf-8") as out:
              out.write(f"any_tidy_files={'true' if selected else 'false'}\n")
              out.write("all_tidy_files<<TIDY_EOF\n")
              out.write("\n".join(selected) + "\n")
              out.write("TIDY_EOF\n")
              out.write("skipped_files<<SKIP_EOF\n")
              out.write("\n".join(skipped) + "\n")
              out.write("SKIP_EOF\n")
          PY          
        env:
          ALL_CHANGED_FILES: ${{ steps.changed_files.outputs.all_changed_files }}

      - name: Show skipped files
        if: steps.changed_files.outputs.any_changed == 'true' && steps.tidy_files.outputs.skipped_files != ''
        run: |
          echo "=== Files skipped by clang-tidy ==="
          printf '%s\n' "${{ steps.tidy_files.outputs.skipped_files }}"          

      - name: Compute submodule commits hash
        id: submodule_hash
        run: |
          # Collect each submodule's current commit SHA, sort for a stable order,
          # then SHA-256 the result so the cache key stays a fixed length.
          hash=$(git submodule status --recursive | awk '{print $1}' | sort | sha256sum | awk '{print $1}')
          echo "value=${hash}" >> "$GITHUB_OUTPUT"          

      - name: Cache third-party build artifacts
        id: cache_thirdparty
        uses: actions/cache@v5
        with:
          path: build/external
          # Cache key combines:
          #   - thirdparty CMakeLists / cmake / patch file contents
          #   - all submodule commit SHAs (upgrading a submodule busts the cache)
          key: thirdparty-${{ runner.os }}-${{ hashFiles('thirdparty/**/*.cmake', 'thirdparty/**/CMakeLists.txt', 'thirdparty/**/*.patch') }}-${{ steps.submodule_hash.outputs.value }}

      - name: Build
        if: steps.tidy_files.outputs.any_tidy_files == 'true'
        run: |
          # build so that clang-tidy can find the headers
          ninja -C build zvec_db          

      - name: Run clang-tidy on changed files
        if: steps.tidy_files.outputs.any_tidy_files == 'true'
        run: |
          # Read the newline-delimited file list into an array.
          mapfile -t files_to_check <<'TIDY_EOF'
          ${{ steps.tidy_files.outputs.all_tidy_files }}
          TIDY_EOF

          failed=0
          for file in "${files_to_check[@]}"; do
            # Skip blank lines that mapfile may produce.
            [[ -z "${file// }" ]] && continue
            if [ -f "$file" ]; then
              echo "=== clang-tidy: $file ==="
              clang-tidy -p build --warnings-as-errors='*' "$file" || failed=1
            fi
          done

          exit $failed          

      - name: No C/C++ files changed
        if: steps.changed_files.outputs.any_changed != 'true' || steps.tidy_files.outputs.any_tidy_files != 'true'
        run: echo "No changed source files with compile_commands entries to analyse."