1 files changed, 36 insertions, 5 deletions
diff --git a/src/nvim/mbyte.c b/src/nvim/mbyte.c
index bd680330ca..cc488d486f 100644
--- a/src/nvim/mbyte.c
+++ b/src/nvim/mbyte.c
@@ -346,7 +346,6 @@ static int enc_canon_search(const char_u *name)
 }
 
 
-
 /*
  * Find canonical encoding "name" in the list and return its properties.
  * Returns 0 if not found.
@@ -1587,7 +1586,7 @@ void show_utf8(void)
   line = get_cursor_pos_ptr();
   len = utfc_ptr2len(line);
   if (len == 0) {
-    MSG("NUL");
+    msg("NUL");
     return;
   }
 
@@ -1610,7 +1609,7 @@ void show_utf8(void)
     }
   }
 
-  msg(IObuff);
+  msg((char *)IObuff);
 }
 
 /// Return offset from "p" to the first byte of the character it points into.
@@ -1883,6 +1882,40 @@ int mb_tail_off(char_u *base, char_u *p)
   return i;
 }
 
+
+/// Return the offset from "p" to the first byte of the character it points
+/// into. Can start anywhere in a stream of bytes.
+///
+/// @param[in] base  Pointer to start of string
+/// @param[in] p     Pointer to byte for which to return the offset to the previous codepoint
+//
+/// @return 0 if invalid sequence, else offset to previous codepoint
+int mb_head_off(char_u *base, char_u *p)
+{
+  int i;
+  int j;
+
+  if (*p == NUL) {
+    return 0;
+  }
+
+  // Find the first character that is not 10xx.xxxx
+  for (i = 0; p - i > base; i--) {
+    if ((p[i] & 0xc0) != 0x80) {
+      break;
+    }
+  }
+
+  // Find the last character that is 10xx.xxxx
+  for (j = 0; (p[j + 1] & 0xc0) == 0x80; j++) {}
+
+  // Check for illegal sequence.
+  if (utf8len_tab[p[i]] == 1) {
+    return 0;
+  }
+  return i;
+}
+
 /*
  * Find the next illegal byte sequence.
  */
@@ -2398,8 +2431,6 @@ static char_u *iconv_string(const vimconv_T *const vcp, char_u *str, size_t slen
 #endif  // HAVE_ICONV
 
 
-
-
 /*
  * Setup "vcp" for conversion from "from" to "to".
  * The names must have been made canonical with enc_canonize().