teaser of Japanese language often breaks

Automatically generated teaser of Japanese language often breaks, because multibyte character split occurred. I fixed it, and please apply the patch.
When concisely saying;
'。' == "\xE3\x80\x82" == 3 bytes

line 187 at node.module;

before

  $breakpoints = array('</p>' => 4, '<br />' => 0, '<br>' => 0, "\n" => 0, '. ' => 1, '! ' => 1, '? ' => 1, '。' => 1, '?? ' => 1);

after

  $breakpoints = array('</p>' => 4, '<br />' => 0, '<br>' => 0, "\n" => 0, '. ' => 1, '! ' => 1, '? ' => 1, '。' => 3, '?? ' => 1);