Actually, it's even worse than that. Continuation headers ALWAYS begin
with white space. Is there a rule for what to do with that when the
line break is between two encoded-words?
It is very interesting that your B-encoded example splits a word from
the real text ("you") in between two encoded-words on two different
lines of the header. So far, I have been unable to come up with any
correct algorithm that won't leave any white space between them. I'd be
grateful for an explanation of what you think a displayer SHOULD be
doing with this white space. -- Nathaniel