utf8 "\xE1" does not map to Unicode at split2.pl line 11, <> chunk 2.
Close with partial character at (eval 21) line 67, <> chunk 2.