The example input file has been slightly modified from the analogous 
Java example: the non-ascii characters have been removed and replaced
by some ascii data.

The lexer would fail, if it tries to compare a byte >0x7f from the input
to a unicode constant.
It would work fine, if you feed unicode into the lexer, but then 
sys.stdout.write() would probably fail, when it tries to print non-ascii
data.

It will also work with non-ascii input (the lexer operates on unicode()
strings), but then you'll have to do some more work (which I omitted for
simplicity):
 - decode the input using the appropriate encoding into a unicode string.
 - encode the data as you send it to the console or store it in a file.