Tokenizes and labels a header value just like MIME.tokenize_labled() ,
but also converts encoded words using MIME.decode_word() . The result
is an array where each element is an array of two or more elements, the
first being the label. The rest of the array depends on the label:
"special"
One additional element, containing the character code for the special
character as an int.
"word"
Two additional elements, the first being the word, and the second
being the character set of this word (or 0 if it did not originate
from an encoded word).
"domain-literal"
One additional element, containing the domain literal as a string.