Sorry for my late reply.
To support UTF-8, we need to change many code. mknmz script depends on
old perl behavior. Recent perl has native UTF-8 string support, but
mknmz was written in old time.
At Thu, 12 Aug 2004 00:18:29 +0200,
Julien Gourdon wrote:
Would using the fedora patch solve my problem ? Is there a way to
"clean" the index ? A mknmz option perhaps ? (I've not found it btw)
The fedora patch is a dirty workround and it considered about *only*
Japanese. I think it shouldn't help you...
--
NOKUBI Takatsugu
E-mail: knok(_at_)daionet(_dot_)gr(_dot_)jp
knok(_at_)namazu(_dot_)org / knok(_at_)debian(_dot_)org