Fix UTF-8 corruption when pasting at multibyte character

col('.') is the first byte of the character under the cursor, so splitting
the line there left one byte in head and the rest in tail. Advance past the
whole character instead.

Close #1627

autoload/fzf/vim.vim

@@ -498,7 +498,10 @@ function! fzf#vim#paste(items) abort
     return s:warn('Cannot paste into a nomodifiable buffer')
   endif
   let line = getline('.')
-  let idx = empty(line) ? 0 : col('.')
+  " col('.') is the first byte of the character under the cursor. Skip the
+  " whole character so that a multibyte character is not split in half.
+  let cursor = col('.') - 1
+  let idx = empty(line) ? 0 : cursor + strlen(matchstr(line, '.', cursor))
   let head = strpart(line, 0, idx)
   let tail = strpart(line, idx)
   let pad = (!empty(head) && head !~ '\s$') ? ' ' : ''