Merge branch 'master' into feat_repo-healthcheck-whitelist-char

pkg/healthcheck/nonascii.go

@@ -10,6 +10,7 @@ import (
 	"unicode"
 	"unicode/utf8"
 
+	"github.com/go-git/go-git/v5"
 	"github.com/go-git/go-git/v5/plumbing/format/gitattributes"
 )
 
@@ -48,6 +49,39 @@ func parseWhitelistedChars(csv string) map[rune]struct{} {
 	return whitelist
 }
 
+// getSubmodulePathsFromGoGit uses the go-git library to open the repository
+// at the given root path and retrieve a list of all submodule paths.
+// It returns a set of submodule paths for efficient lookup.
+func getSubmodulePathsFromGoGit(root string) (map[string]struct{}, error) {
+	submodulePaths := make(map[string]struct{})
+
+	// Open the git repository at the given path.
+	repo, err := git.PlainOpen(root)
+	if err != nil {
+		if err == git.ErrRepositoryNotExists {
+			return submodulePaths, nil
+		}
+		return nil, fmt.Errorf("error opening git repository: %w", err)
+	}
+
+	worktree, err := repo.Worktree()
+	if err != nil {
+		return nil, fmt.Errorf("error getting worktree: %w", err)
+	}
+
+	// Get the list of submodules.
+	submodules, err := worktree.Submodules()
+	if err != nil {
+		return nil, fmt.Errorf("error getting submodules: %w", err)
+	}
+
+	for _, sm := range submodules {
+		submodulePaths[filepath.ToSlash(sm.Config().Path)] = struct{}{}
+	}
+
+	return submodulePaths, nil
+}
+
 // getNonASCII retrieves a list of files in the specified root directory that contain non-ASCII characters.
 // It searches for non-ASCII characters in each file's content and returns a list of paths to files containing non-ASCII characters.
 func getNonASCII(root string, whitelist map[rune]struct{}) ([]string, error) {
@@ -59,6 +93,11 @@ func getNonASCII(root string, whitelist map[rune]struct{}) ([]string, error) {
 		gitattrExist = false
 	}
 
+	submodules, err := getSubmodulePathsFromGoGit(root)
+	if err != nil {
+		return nil, err
+	}
+
 	if gitattrExist {
 		fs := os.DirFS(".")
 		f, err := fs.Open(".gitattributes")
@@ -78,18 +117,22 @@ func getNonASCII(root string, whitelist map[rune]struct{}) ([]string, error) {
 			return err
 		}
 
+		relPath, err := filepath.Rel(root, path)
+		if err != nil {
+			return err
+		}
+
 		if info.IsDir() {
 			if info.Name() == ".git" {
 				return filepath.SkipDir
 			}
+			if _, isSubmodule := submodules[relPath]; isSubmodule {
+				return filepath.SkipDir
+			}
 			return nil
 		}
 
 		if gitattrExist {
-			relPath, err := filepath.Rel(root, path)
-			if err != nil {
-				return err
-			}
 			ret, matched := matcher.Match(strings.Split(relPath, "/"), nil)
 			if matched && ret["text"].IsUnset() && !ret["text"].IsSet() {
 				return nil