Chinese Debian Mini Howto

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.2 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license is included in the section entitled "GNU Free Documentation License".

0. Introduction

• Generate relevant locales
• Install Chinese fonts
• Install an input method (IM) engine
• Set locale
• Adjust application settings (if necessary)

In the rest of this text, I will explain these steps one by one. Most of the commands below need to be run in a terminal with root privilege.

1. Generating Locales

• en_US ISO-8859-1
• zh_CN GB2312
• zh_CN.GBK GBK
• zh_CN.UTF-8 UTF-8
• zh_TW BIG5
• zh_TW.UTF-8 UTF-8

NOTE: Sometimes you need to reboot to get the new locales working. To avoid potential problems, it is strongly recommended that you get this correct from the beginning when you install your system.

2. Installing Chinese Fonts

• ttf-arphic-gbsn00lp (AR PL SungtiL GB)
• ttf-arphic-gkai00mp (AR PL KaitiM GB)
• ttf-arphic-bsmi00lp (AR PL Mingti2L Big5)
• ttf-arphic-bkai00mp (AR PL KaitiM Big5)

These packages can be installed by running

apt-get install ttf-arphic-bkai00mp ttf-arphic-bsmi00lp ttf-arphic-gbsn00lp ttf-arphic-gbsn00lp

There are other fonts available. You can find them by searching for "xfonts" using dselect

3. Installing Input Method (IM) Engine(s)

There are several packages related to scim. The easiest way to install it is by running

apt-get install scim scim-chinese scim-tables-zh

The package names may change in the future (as they did in the past). In that case, you can search for packages beginning with "scim" using dselect, examine their descriptions, and choose the input methods you need. After that dselect will do the rest of the job by selecting all dependent packages.

After that, create a new file /etc/X11/Xsession.d/95xinput with the following content.

/usr/bin/scim -d
XMODIFIERS="@im=SCIM"
export XMODIFIERS

This script will be run every time X windows starts. In case you want to be more flexible, you can put something more complicated in the file. For example,

case "$LANG" in
zh_TW*)
    /usr/bin/scim -d
    XMODIFIERS="@im=SCIM"
    ;;
zh_CN*)
    /usr/bin/scim -d
    XMODIFIERS="@im=SCIM"
    ;;
esac

export XMODIFIERS

To use scim, simply press "Ctrl-Space", and a small window will appear at the lower right corner of the desktop.

It is advisable that you configure scim (right click on its icon on the panel, then configure) and remove all unwanted input methods. You will need to restart scim (probably X, too) to make this to take effect.

4. Setting Locale

If you are using a X display manager that does not support this, you will have to put an additional line such as

export LANG=zh_CN.gb2312

in /etc/X11/Xsession.d/95xinput. NOTE: It would not work if you run this after you login. You will need to restart X for this to work.

REMARK: One "side effect" of this is that once you set the language to be Chinese, then all the menu become Chinese. For those who want to keep the English menu but still want to view/input Chinese, you can set locale to be zh_CN.gb2312, but change the settings for some environment variables. For example, I have the following lines in the above file.

ENCODING="en_US"
#export LC_ALL=$ENCODING
export LC_MESSAGES=$ENCODING
#export LC_COLLATE=$ENCODING
#export LC_CTYPE=$ENCODING
export LC_TIME=$ENCODING
export LC_NUMERIC=$ENCODING
#export LC_MONETARY=$ENCODING
#export LC_PAPER=$ENCODING
#export LC_NAME=$ENCODING
export LC_ADDRESS=$ENCODING
export LC_TELEPHONE=$ENCODING
export LC_MEASUREMENT=$ENCODING
export LC_IDENTIFICATION=$ENCODING

Note that "scim" works fine no matter what locale you choose.

5. Application Settings

5.1 Web Browsers

For Mozilla version 1.6, there are some locale packages such as mozilla-locale-zh-cn or mozilla-locale-zh-tw.

For Mozilla-Firefox, you might need to install one of the mozilla-firefox-locale-zh-cn or mozilla-firefox-locale-zh-tw packages.

5.2 Editors

set enc=euc-cn
set tenc=euc-cn
set fileencoding=euc-cn
set guifont=AR\ PL\ KaitiM\ GB\ 12

The last line specifies the font and the font size to use. You can change it to any of the four fonts as in Section 2 above, and adjust the font size until you feel comfortable.

Note that even when you set LC_MESSAGE to en_US, gvim might still display Chinese menu if your locale is set to Chinese. In this case, you need a bit trick here.

Firstly, you need to create a file with the content something like the following.

#!/bin/sh
# Start application $1 with English environment

if [ -z "$1" ]; then
  echo "Usage: $0 app arg1 arg2 ... "
  exit 1
fi

export LANG=en_US

PROG=$1
shift
exec $PROG $*

This small shell script will set $LANG so that the application it runs would think that it is running in a full English system.

To run gvim, we run enstart.sh gvim instead. You can add this to your desktop/panel shortcut, or make it an alias.

5.3 Terminal Emulators

The gnome-terminal, the default terminal emulator in the GNOME environment, supports Chinese by selecting the character encoding from the terminal menu.

The default terminal emulator in KDE, Konsole, works fine with Chinese automatically if the locale is set correctly. You can also select/change character encoding from the menu.

5.4 Display of Chinese File Names in FAT Partitions

fat
vfat

nls_cp936 (for simplified Chinese)
nls_cp950 (for traditional Chinese)
nls_utf8 (for Unicode characters)

Then in the file /etc/fstab, add another line like the following

/dev/hda5 /mnt/dos vfat noauto,user,codepage=936,iocharset=cp936 0 0

for simplified Chinese, and replace the number 936 to 950 for traditional Chinese, and to utf8 for Unicode characters. NOTE: you should change the partition (/dev/hda5) and mount point (/mnt/dos) to the partition you want to mount, and the directory you want it to be mounted to, respectively.

5.5 XMMS

Fortunately, similar trick as in the case of gvim can be applied here, but this time we want to run with Chinese support. Let us create the following file and call it chstart.sh

#!/bin/sh
# Start application $1 with Chinese environment

if [ -z "$1" ]; then
  echo "Usage: $0 app arg1 arg2 ... "
  exit 1
fi

export LANG=zh_CN.gb2312

PROG=$1
shift
exec $PROG $*

To run XMMS, just type chstart.sh xmms in a terminal, or create a shortcut on desktop/panel to do it, or use an alias.

5.6 Instant Messengers

A work-around for AMSN is to use the enstart.sh to start it instead (but of course you cannot expect to input Chinese with it).

Fortunately, there are instant messengers that do support Chinese. Skype is an example, but only works when gdm/kdm has the locale set to zh_CN.gb2312 otherwise it will not work even if you start it with the chstart.h script above.

GAIM is another cute IM that was reported to work with Chinese. It is compatible with AIM, Yahoo! Messenger, MSN Messenger, Google Talk (Jabber), etc.

5.7 Java

One of things trouble me a lot is the Java on Debian do not support Chinese natively. I got the solution to Java1.5 recently (tested for GB2312).

Here is the way (two steps):
1. make a directory fallback under JavaHome/lib/fonts/
2. add a Chinese TrueType or Type 1 font under fallback

An example:

mkdir /usr/lib/j2sdk1.5-sun/jre/lib/fonts/fallback
ln -s /usr/share/fonts/truetype/arphic/gbsn00lp.ttf .

Note the trailing dot at the end of the above line.

Some explanations for the fallback:
"... If the runtime environment has a directory lib/fonts/fallback and this directory contains valid TrueType or Type 1 fonts, the runtime automatically adds these fonts as fallback fonts for 2D rendering..." --- from http://java.sun.com/j2se/1.5.0/docs/guide/intl/fontconfig.html

Note: This solution should be considered as quick and dirty, the best solution would be change the fontconfig file for Java. But, I haven't figured out that yet.