#2 Apr 02, 2006 23:36
Thanks for reporting.
IMHO this is a bug, yes. Though the current implementation is that only "a-z" and "0-9" are valid chars and everything else gets replaces by underscores.
I've added it to the todo list and it should get fixed.