Add a comment to the Turkish casefiddle test

diff --git a/test/src/casefiddle-tests.el b/test/src/casefiddle-tests.el
index 164adbc19ef14dfa606990fa6dc4662c7ac46d6e..dbbe9f30925f83c2c63cc918593fe4a1d2067e7e 100644 (file)
--- a/test/src/casefiddle-tests.el
+++ b/test/src/casefiddle-tests.el
@@ -280,6 +280,9 @@
 
 (ert-deftest casefiddle-turkish ()
   (skip-unless (member "tr_TR.utf8" (get-locale-names)))
+  ;; See bug#50752.  The point is that unibyte and multibyte strings
+  ;; are upcased differently in the "dotless i" case in Turkish,
+  ;; turning ASCII into non-ASCII, which is very unusual.
   (with-locale-environment "tr_TR.utf8"
     (should (string-equal (downcase "I ı") "ı ı"))
     (should (string-equal (downcase "İ i") "i̇ i"))