2 files changed, 10 insertions, 0 deletions
diff --git a/lib/unicode.ml b/lib/unicode.ml
index 1765e93dcd..cfaa73cc11 100644
--- a/lib/unicode.ml
+++ b/lib/unicode.ml
@@ -168,6 +168,13 @@ let next_utf8 s i =
        (c land 0x3F) lsl 6 + (d land 0x3F)
   else err ()
 
+let is_utf8 s =
+  let rec check i =
+    let (off, _) = next_utf8 s i in
+    check (i + off)
+  in
+  try check 0 with End_of_input -> true | Invalid_argument _ -> false
+
 (* Check the well-formedness of an identifier *)
 
 let initial_refutation j n s =
diff --git a/lib/unicode.mli b/lib/unicode.mli
index 520203d435..65e75a20d6 100644
--- a/lib/unicode.mli
+++ b/lib/unicode.mli
@@ -26,3 +26,6 @@ val lowercase_first_char : string -> string
 (** For extraction, turn a unicode string into an ascii-only one *)
 val is_basic_ascii : string -> bool
 val ascii_of_ident : string -> string
+
+(** Validate an UTF-8 string *)
+val is_utf8 : string -> bool