[XeTeX] Polyglossia: Support for romanization of CJK

[Arthur Reutenauer]

>                                                     tsu-kattari,
> tsuka-ttari or tsukatta-ri ; second hyphenation preferable to
> tsukat-tari because of how the word is segmented when spoken.

  Is there any established tradition on this?  Hyphenation often doesn't
respect syllable boundaries, and the original poster seemed to assume
that we wanted to break in the middle of a pair of geminated consonants.
That's what the patterns I sent yesterday are doing; but of course it
would be easy to change (and would make the file even shorter,
actually).

  And, for that matter, does what you say mean that in 新聞, for
example, the first syllable ends after the /i/?

	Arthur