[debian-devel:05887] [jshin@xxxxxxxxxxxxxxxxx: glibc-wcsmbs and locale data for Korean EUC-KR]

[yosshy@debian.or.jp (Akira YOSHIYAMA)]

  吉山＠目黒です。皆様お久しぶりです。

  謎の在米韓国人 shin さんから、これまた久しぶりなメールが来ました。
  でも、なんて書いてあるか分かんない…

------- Forwarded Message
Date: Wed, 4 Nov 1998 04:46:16 JST
From: Jungshik Shin <jshin@xxxxxxxxxxxxxxxxx>
Reply-To: jshin@xxxxxxxxxxxxxxxxx
To: yosshy@debian.or.jp
Subject: glibc-wcsmbs and locale data for Korean EUC-KR
Message-ID: <Pine.LNX.3.96K.981103142707.21545A-100000@xxxxxxxxxxxxxxxxxxx>
MIME-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7BIT
X-UIDL: 30303030303030303330373032393232


  Hello,

   I'm not sure if you still remember me. Earlier this year, I wrote to
you several times about your wcsmbs patch to glibc 2.0.x and provided
you with some information on Korean encodings, EUC-KR, UHC, and JOHAB.

  As you may know, most Korean linux users have used liblocale.so (see
http://pantheon.yale.edu/~jshin/faq/qa36.html. This work around was
first found in Japan) to enable Korean Input Method to work with
Netscape. However, there are a couple of problems with entering Korean
in 'textarea'(form) and mail/news composition window. I guess the
same is true of Japanese input. 

 A few days ago, I heard that some trouble with inputting Japanese and
Korean in Netscape in Linux/FreeBSD with liblocale.so preloaded to
replace setlocale() with Xsetlocale()  has to do with abnormal behavior
of wcs*mb*()/mb*wc*(). Japanese FreeBSD users came up with a workaround,
preloading libansi.so which redefines wc*mb*()/mb*wc*()) but that isn't
applicable to Linux. Then, it occurred to me that your patch to glibc
2.0.x with wc*mb*()/mb*wc*() support would solve the problem easily.  Is
this correct? Can you enter Japanese in any place of Netcape
communicator without any trouble when using glibc-2.0.7-wcsmbs?

 So, I downloaded the newest version of glibc2.0.7-wcsmbs and
wcsmbs-0.4.x from ftp.debian.or.jp. For Japanese locales, everything
seems to be in order, but locale data(LC_CTYPE, LC_TIME, etc) is missing
for Korean locales. Without that, fontList/fontset spec. for X11 (e.g.
Netscape) wouldn't work, would it? Hence, my question is how I can
generate those locale files(LC_CTYPE, LC_TIME,etc) for Korean
EUC-KR(what version of localedef and what format of locale and charmap
source did you use?)  I saw you wrote that those for Japanese are sort
of 'dummy' only to make fontList/fontset work with X11 programs. I guess
that would be sufficient for Korean users, too. Once I succeed in
building 'dummy' Korean locales, I'll send them to you for inclusion
in your package. 

   Thank you much in advance, 

        Jungshik Shin







------- End of Forwarded Message

---
   Name: 吉山あきら  Akira Yoshiyama
   E-mail: yosshy@debian.or.jp  (runlevel1@xxxxxxxxxxxxxxxxx)
   URL: http://jedi.seg.kobe-u.ac.jp/~yosshy/linux.html