hyphenation and ideographic characters

Hi all.

After hyphenation feature was merged into main branch, line break
doesn't seem to work well for ideographic characters.

This is screenshot from 2387:81f484e70da6 (I think this is the last
changeset before hyphenation).

http://furaisanjin.blog.so-net.ne.jp/_images/blog/_9f7/furaisanjin/1-c9360.png

But the current one (2493:f99afbd38ae7) shows something like this.

http://furaisanjin.blog.so-net.ne.jp/_images/blog/_9f7/furaisanjin/2-fe1a8.png

I added red box to highlight the difference. We had similar problem
before adding a_Utf8_ideographic on Html_process_word in html.cc.
Hyphenation also needs similar checking.

Regards,
furaisanjin