Logged In: YES 
user_id=21627

I think there is a design choice here: Should it be an
invariant that all characters in a Unicode string are below
sys.maxunicode? If so, the procedures to create them should
check whether that invariant holds. If not, all code that
deals with them most consider cases where they are larger
than sys.maxunicode or smaller than 0.

Also, should it be a requirement that Py_UNICODE is an
unsigned type? Then tests for <0 could be dropped, of course.