character vs. token based parser in Parsec with haskell

Ok, the first parser did look *ugly*, so I wrote a bit less ugly one, using the token parser. It's not much shorter,
but runs two times faster!

$ time ./packet_template_token >/dev/null

real    0m0.138s
user    0m0.130s
sys     0m0.006s
$ time ./packet_template_char >/dev/null

real    0m0.395s
user    0m0.388s
sys     0m0.005s
$

In case you are interested, both are here