Function getGitVisibleFiles

src/extraction/index.ts:1018–1065

* Get all files visible to git (tracked + untracked but not ignored). * Respects .gitignore at all levels (root, subdirectories) and descends into * embedded (nested, non-submodule) git repos. Returns null on failure * (non-git project) so callers can fall back to a filesystem walk.

(rootDir: string)

 * (non-git project) so callers can fall back to a filesystem walk.
 */
function getGitVisibleFiles(rootDir: string): Set<string> | null {
  try {
    // Check if the project directory is gitignored by a parent repo.
    // When rootDir lives inside a parent git repo that ignores it,
    // `git ls-files` returns nothing — fall back to filesystem walk.
    const gitRoot = execFileSync(
      'git',
      ['rev-parse', '--show-toplevel'],
      { cwd: rootDir, encoding: 'utf-8', timeout: 5000, stdio: ['pipe', 'pipe', 'pipe'], windowsHide: true }
    ).trim();

    if (path.resolve(gitRoot) !== path.resolve(rootDir)) {
      try {
        // git check-ignore exits 0 if the path IS ignored, 1 if not
        execFileSync(
          'git',
          ['check-ignore', '-q', path.resolve(rootDir)],
          { cwd: rootDir, encoding: 'utf-8', timeout: 5000, stdio: ['pipe', 'pipe', 'pipe'], windowsHide: true }
        );
        // Directory is gitignored by parent repo — fall back to filesystem walk
        return null;
      } catch {
        // Not ignored — safe to use git ls-files
      }
    }

    const files = new Set<string>();
    const embeddedRoots = new Set<string>();
    collectGitFiles(rootDir, '', files, embeddedRoots, loadIncludeIgnoredMatcher(rootDir));
    // Apply built-in default ignores uniformly — to tracked files too, since
    // committing a dependency/build dir doesn't make it project code. A
    // `.gitignore` negation (e.g. `!vendor/`) is the explicit opt-in. (issue #407)
    // Files inside an EMBEDDED repo are matched against that repo's own rules,
    // not the parent's: the parent's .gitignore hides the child repo from git,
    // not from the index. (#514)
    const ig = buildScopeIgnore(rootDir, embeddedRoots);
    const visible = new Set([...files].filter((f) => !ig.ignores(f)));
    // Force-include first-party source the project whitelisted in
    // `codegraph.json` `include`. These are gitignored, so `git ls-files` never
    // listed them above — discover them directly off disk and add them. (The
    // common SVN+Git dual-VCS case: source committed to SVN, gitignored out of
    // Git, but still wanted in the graph.)
    for (const f of collectIncludedFilesForRoot(rootDir)) visible.add(f);
    return visible;
  } catch {
    return null;
  }
}

/**
 * Result of git-based change detection.