Correction on the XXX should we check this: I was thinking about the wrong section of the code.  But it is still 'no': by postel's law we should accept dirty data.  Currently the consumer of the library can then decide whether or not to reject the dirty data by building a subclass.  It might be nice to provide an option to control rejection of invalid characters in commands, but that should be a separate issue.  (And, we are ending up with so many options that we might want to think about whether or not there is a better API for controlling them).