fulltext parser strange behave

Hello

I am writing tsearch2 wrapper and I testing functionality. I found
some little bit strange on default parser. It can't parse tags with
numbers:

test=# select * from parse('<h1>zluty kun se napil <b>zlute</b> vody</h2>');tokid | token
-------+-------   12 | <    3 | h1   12 | >    1 | zluty   12 |    1 | kun   12 |    1 | se   12 |    1 | napil   12 |
13 | <b>    1 | zlute   13 | </b>   12 |    1 | vody   12 | <              <=====   19 | /h2   12 | >
<=====
(19 rows)

It is correct?

Regards
Pavel Stehule