You could use

HTML::Entities::decode_entities_old($string)
[download]

thats the old perl fallback. Or make your string utf8 before you pass it to decode_entities.

my $string = "’";
chop ( $string .= chr(0x1234)); 
print "string: $string\n"; # &rsquo 
decode_entities($string);
[download]

and since we like to print a utf8 char put STDOUT into utf8 mode too.

# at the top of the program
binmode STDOUT, ":utf8";
[download]

UPDATE: or even better update HTML::Parser and it works out of the box. Tested with 3.43.