fixed auto-completion of Unicode file names

* teco_string_diff() could return a number of bytes in the middle of
  an Unicode sequence. It now also requires Unicode strings.
* Added a missing Unicode-validity check when replacing command lines (`{` and `}`).
  teco_cmdline_insert() should really be refactored, though (FIXME).
* Added test case

1 files changed, 6 insertions, 4 deletions

diff --git a/src/string-utils.c b/src/string-utils.c
index e9dd148..d98b6b0 100644
--- a/src/string-utils.c
+++ b/src/string-utils.c
@@ -87,8 +87,10 @@ teco_string_get_coord(const gchar *str, gsize off, guint *pos, guint *line, guin
 }
 
 /**
- * Get the length of the prefix common to two strings.
- * Works with UTF-8 and single-byte encodings.
+ * Get the length of the prefix common to two UTF-8 strings.
+ *
+ * The UTF-8 strings must be validated, which should be the case
+ * for help labels and short Q-Register names.
  *
  * @param a Left string.
  * @param b Right string.
@@ -103,8 +105,8 @@ teco_string_diff(teco_string_t a, const gchar *b, gsize b_len)
 	gsize len = 0;
 
 	while (len < a.len && len < b_len &&
-	       a.data[len] == b[len])
-		len++;
+	       g_utf8_get_char(a.data+len) == g_utf8_get_char(b+len))
+		len = g_utf8_next_char(b+len) - b;
 
 	return len;
 }