1 files changed, 145 insertions, 0 deletions
diff --git a/SanitizeUTF8.go b/SanitizeUTF8.go
new file mode 100644
index 0000000..9ad4171
--- /dev/null
+++ b/SanitizeUTF8.go
@@ -0,0 +1,145 @@
+// Copyright 2017-2025 WIT.COM Inc. All rights reserved.
+// Use of this source code is governed by the GPL 3.0
+
+package bugpb
+
+import (
+	"bytes"
+	"fmt"
+	"io"
+	"reflect"
+	"unicode/utf8"
+
+	"golang.org/x/text/encoding/charmap"
+	"google.golang.org/protobuf/proto"
+)
+
+// ValidateProtoUTF8 checks all string fields in a proto.Message recursively.
+func ValidateProtoUTF8(msg proto.Message) error {
+	return validateValue(reflect.ValueOf(msg), "")
+}
+
+func validateValue(val reflect.Value, path string) error {
+	if !val.IsValid() {
+		return nil
+	}
+
+	if val.Kind() == reflect.Ptr {
+		if val.IsNil() {
+			return nil
+		}
+		return validateValue(val.Elem(), path)
+	}
+
+	switch val.Kind() {
+	case reflect.Struct:
+		for i := 0; i < val.NumField(); i++ {
+			field := val.Field(i)
+			fieldType := val.Type().Field(i)
+			fieldPath := fmt.Sprintf("%s.%s", path, fieldType.Name)
+			if err := validateValue(field, fieldPath); err != nil {
+				return err
+			}
+		}
+
+	case reflect.String:
+		s := val.String()
+		if !utf8.ValidString(s) {
+			return fmt.Errorf("invalid UTF-8 string at %s: %q", path, s)
+		}
+
+	case reflect.Slice:
+		if val.Type().Elem().Kind() == reflect.Uint8 {
+			return nil // skip []byte
+		}
+		for i := 0; i < val.Len(); i++ {
+			if err := validateValue(val.Index(i), fmt.Sprintf("%s[%d]", path, i)); err != nil {
+				return err
+			}
+		}
+
+	case reflect.Map:
+		for _, key := range val.MapKeys() {
+			valItem := val.MapIndex(key)
+			if err := validateValue(valItem, fmt.Sprintf("%s[%v]", path, key)); err != nil {
+				return err
+			}
+		}
+	}
+
+	return nil
+}
+
+// SanitizeProtoUTF8 fixes all invalid UTF-8 strings in a proto.Message recursively.
+func SanitizeProtoUTF8(msg proto.Message) error {
+	return sanitizeValue(reflect.ValueOf(msg), "")
+}
+
+func sanitizeValue(val reflect.Value, path string) error {
+	if !val.IsValid() {
+		return nil
+	}
+
+	if val.Kind() == reflect.Ptr {
+		if val.IsNil() {
+			return nil
+		}
+		return sanitizeValue(val.Elem(), path)
+	}
+
+	switch val.Kind() {
+	case reflect.Struct:
+		for i := 0; i < val.NumField(); i++ {
+			field := val.Field(i)
+			fieldType := val.Type().Field(i)
+			if !field.CanSet() {
+				continue
+			}
+			if err := sanitizeValue(field, fmt.Sprintf("%s.%s", path, fieldType.Name)); err != nil {
+				return err
+			}
+		}
+
+	case reflect.String:
+		s := val.String()
+		if !utf8.ValidString(s) {
+			utf8Str, err := latin1ToUTF8(s)
+			if err != nil {
+				return fmt.Errorf("failed to convert %s to UTF-8: %v", path, err)
+			}
+			val.SetString(utf8Str)
+		}
+
+	case reflect.Slice:
+		if val.Type().Elem().Kind() == reflect.Uint8 {
+			return nil // skip []byte
+		}
+		for i := 0; i < val.Len(); i++ {
+			if err := sanitizeValue(val.Index(i), fmt.Sprintf("%s[%d]", path, i)); err != nil {
+				return err
+			}
+		}
+
+	case reflect.Map:
+		for _, key := range val.MapKeys() {
+			valItem := val.MapIndex(key)
+			newItem := reflect.New(valItem.Type()).Elem()
+			newItem.Set(valItem)
+			if err := sanitizeValue(newItem, fmt.Sprintf("%s[%v]", path, key)); err != nil {
+				return err
+			}
+			val.SetMapIndex(key, newItem)
+		}
+	}
+
+	return nil
+}
+
+func latin1ToUTF8(input string) (string, error) {
+	reader := charmap.ISO8859_1.NewDecoder().Reader(bytes.NewReader([]byte(input)))
+	result, err := io.ReadAll(reader)
+	if err != nil {
+		return "", err
+	}
+	return string(result), nil
+}