This chapter describes how to perform the following tasks to establish a basic Japanese environment on a Tru64 UNIX system.
Install the Japanese environment
Set up the internationalization environment
Set up the Common Desktop Environment (CDE) language
Set up the Japanese locale
Set up the Japanese terminal
Perform file code conversions with
iconv()
Set up the message catalogs
Set up the Japanese reference pages
2.1 Installing the Japanese Environment
To use the Japanese features that Tru64 UNIX offers, you must
install the Worldwide Language Support subsets and Japanese subsets as well
as the Tru64 UNIX base software subsets.
Figure 2-1: Software Subsets Required to Use the Japanese Environment
The subsets in Figure 2-1 are needed to use the Japanese features and are automatically installed when you select Japanese in the installation procedure. For the detailed steps on the installation procedure, refer to the Installation Guide and Installation Guide Advanced Topics.
To use any of the following Japanese features, you must install the associated software kit after installing the Tru64 UNIX Japanese environment:
The installation kit for the VJE Japanese input system is contained on a separately orderable CD-ROM. For ordering information, refer to the Worldwide Language Support web page at http://www.tru64unix.compaq.com/unix/i18n.htm. For installation information, refer to the VJE User's Guide.
WX3
The Japanese input system WX3 kit is contained on a separately orderable CD-ROM. For ordering information, refer to the Worldwide Language Support web page at http://www.tru64unix.compaq.com/unix/i18n.htm. For installation information, refer to the WX3 User's Guide.
midoc
and
miman
The Japanese document
full-text search tool
(midoc)
and reference page full-text
search tool
(miman)
are contained on the
Tru64 UNIX Japanese Extensions
CD-ROM.
For
installation and usage information, refer to
Chapter 6.
2.2 Setting Up the Internationalization Environment
To use the Tru64 UNIX Japanese features, first set the following environment variables for the internationalization environment:
Set the directory that will be used as the virtual root directory in the internationalization environment.
PATH
Set
$I18NPATH/bin
at the head of PATH to execute the commands included in the Japanese subsets
(for example,
jvi
) and to give precedence to the execution
of Japanese subsets.
Tru64 UNIX provides the following script files to help you properly set the environment variables:
/etc/i18n_profile
Contains
descriptions of the settings you should include in the user's
.profile
file
/etc/i18n_login
Contains
descriptions of the settings you should include in the user's
.cshrc
file
/etc/i18n_path
Contains descriptions of the path settings that other commands require to use any of the internationalization commands and utilities
The scripts are template files.
The system manager can modify either
the
/etc/i18n_profile
or the
/etc/i18n_login
file, as appropriate.
Note
Do not modify the
/etc/i18n_path
file because it depends on the system directory structure.
If you use the
/usr/sbin/i18nsetup
command, the contents of the skeleton file
are added to the end of each of the user's startup scripts ($HOME/.profile
and
$HOME/.cshrc
).
Use an ASCII editor to
reposition these additions within the file.
If you are logged in as superuser (root), you can use the
/usr/sbin/i18nsetup
command to modify the startup script of any
user.
Consider the following example that modifies the startup files of users
kanai
and
shimada
.
# /usr/sbin/i18nsetup kanai shimada
If
the startup files are properly modified, the commands and utilities for the Tru64 UNIX
internationalization environment will be available when you log in to the
system, or the settings will be reflected in the current shell environment
using
eval
(for
sh
) or
source
(for
csh
).
Note
The descriptions of Japanese environment settings and operations in the remainder of this chapter assume that the environment variable settings for I18NPATH and PATH described in this section have been made. If the Japanese environment is not set up properly, examine the values of these environment variables.
2.3 Setting up the CDE Language
To set the language for each user in the
CDE, select the appropriate locale from the Language submenu of the login
dialog Options when logging in to the system.
To set the system language to
Japanese, select one of the Japanese locales (listed in
Table 2-1)
from the Language submenu.
Table 2-1: Japanese Locales That Can Be Selected for the CDE Language Setting Option
Language | Locale |
Japanese | ja_JP |
Japanese (DEC Kanji) | ja_JP.deckanji |
Japanese (EUC) | ja_JP.eucJP |
Japanese (Super DEC Kanji) | ja_JP.sdeckanji |
Japanese (Unicode) | ja_JP.UTF-8 |
Japanese (Shift JIS) | ja_JP.SJIS |
The value of the LANG environment
variable for CDE sessions is automatically set using the language you select
during login.
Thus, you do not need to set the LANG environment variable in
a startup script (such as
.login
).
Note
If you use this method to select a language for a CDE session, you must not define the LANG environment variable in the startup script. If there is a mismatch between the language you selected during login and the value of LANG in the startup script, the session might be unable to start properly.
2.3.1 Setting the CDE Default Language
If you select only Japanese when you install the Worldwide Language Support software, the default language for the window environment is automatically set to Japanese EUC (ja_JP.eucJP).
You can use the following procedure to change the default language for the Japanese window environment. This procedure is useful if you want to set a Japanese locale other than EUC for the default language for the window environment or if you selected more than one language when installing the Worldwide Language Support software.
If an
/etc/dt/config/Xconfig
file exists,
delete it.
Create a new
/etc/dt/config/Xconfig
file
using the following commands:
# cd /etc/dt/config # ln -s Xconfig.locale-name Xconfig
For example, to set the default language to Japanese EUC, use the following command.
# ln -s Xconfig.ja_JP.euc.JP Xconfig
This setting is enabled for both the login screen and default session
languages.
2.4 Setting Up the Japanese Locale
To associate the Tru64 UNIX internationalized commands and libraries with a particular Japanese environment, specify the locale that represents the desired combination of language, territory, and codeset.
To specify the locale name representing the locale, use the following format:
Language_Territory.Codeset
For Language, Territory, and Codeset, specify the respective language, territory, and character codeset names defined in the system.
Tru64 UNIX provides the Japanese locales listed in
Table 2-2.
All of these locales specify the Japanese language currently used in Japan;
they differ with respect to the character codeset name used.
Table 2-2: Japanese Locale Names Available for Tru64 UNIX
Locale Name | Character Codeset Name Used |
ja_JP.deckanji | DEC Kanji |
ja_JP.sdeckanji | Super DEC Kanji |
ja_JP.eucJP | Japanese EUC |
ja_JP.SJIS | Shift JIS |
ja_JP | Japanese EUC |
ja_JP.UTF-8 | UTF-8 |
At installation, Tru64 UNIX
stores the Japanese locale databases in the paths listed in
Table 2-3.
Table 2-3: Locale Databases
Locale Name | Locale Databases |
ja_JP.deckanji | /usr/lib/nls/loc/ja_JP.deckanji |
ja_JP.sdeckanji | /usr/lib/nls/loc/ja_JP.sdeckanji |
ja_JP.eucJP | /usr/lib/nls/loc/ja_JP.eucJP |
ja_JP.SJIS | /usr/lib/nls/loc/ja_JP.SJIS |
ja_JP.UTF-8 | /usr/lib/nls/loc/ja_JP.UTF-8 |
To specify a Japanese locale, enter the appropriate locale name listed in Table 2-2 as the value for the LANG environment variable. The commands and utilities of the Tru64 UNIX internationalization environment will be associated with the locale name value set for the LANG environment variable.
In Tru64 UNIX, there are six categories of internationalized
commands and libraries that differ from one locale to another.
Table 2-4
describes the categories and the settings defined for each.
The category names
correspond with environment value names.
You can independently specify a locale
name suitable for each of these environment variables.
You can also set up
the locales with a variety of environment variable definitions, depending
on the processing you want to accomplish.
Table 2-4: Available Category Names
Category Name (Environment Variable Name) | Settings |
LC_COLLATE | Collating sequence |
LC_CTYPE | Character class, character attribute (uppercase, lowercase, etc.) |
LC_MESSAGES | Message display format |
LC_MONETARY | Monetary format |
LC_NUMERIC | Numeric data format |
LC_TIME | Date/time format |
To assign the same locale name to all these categories, use the LC_ALL environment variable.
The system applies the following priority in determining which environment variables control the locale settings:
LC_ALL environment variable (when a valid locale name has been specified)
LC_COLLATE, LC_CTYPE, LC_MESSAGES, LC_MONETARY, LC_NUMERIC, LC_TIME environment variables (when a valid locale name has been specified)
LANG environment variable (when a valid locale name has been specified)
If you specified no locale name or an invalid locale name as an environment variable value, Tru64 UNIX assumes that the C locale has been specified. The C locale is the most basic environment and can process only ASCII characters. With the C locale, Japanese characters cannot be processed.
To view the settings of the current locale, use the
locale
command:
% locale
For more information on the
locale
command, see
locale
(1)2.5 Setting Up the Japanese Terminal
To display Japanese characters on the terminal or input them through the keyboard, use the following command to set the terminal driver to the line discipline for Japanese characters:
% stty jdec
Immediately after execution of this command, the terminal driver is set as follows:
Item | Appropriate stty Option | Settings |
Terminal I/O Code | tcode dec |
DEC Kanji code |
Application code | acode
codeset-name |
Japanese codeset being set for the locale (eucJP if no locale has been set) |
Japanese Input Mode | imode hiragana |
Input an 8-bit code as a 2-byte Hiragana code |
Kanji Conversion | -ikk |
Configuration Off |
If your environment does not match the above settings,
use the
stty
command to establish the appropriate setting.
The following sections detail how to use the
stty
command
to set handling of terminal and application codes and to set 8-bit data for
your environment.
2.5.1 Setting the Terminal and Application Codes
You can independently set Kanji codes for each of the terminal and application codes. Various combinations are available.
The application program uses the terminal driver atty
[Footnote 1]
to communicate with the Tru64 UNIX terminal.
The terminal driver
atty is incorporated in the kernel (see
Figure 2-2).
Figure 2-2: Terminal Driver (Asian tty)
Tru64 UNIX allows you to select
the respective character codes interchanged between the terminal, terminal
driver, and application.
Throughout this document, the character codes that
the terminal uses for communication with the host are called terminal codes
and the character codes that the application processes are called application
codes.
2.5.1.1 Setting the Terminal Codeset
To set a character codeset other than DEC Kanji for the terminal codes,
use the
stty
command with the
-tcode
option:
% stty tcode codeset-name
Table 2-5
lists the codeset names that you can set.
Table 2-5: Possible Codeset Names for the Terminal Codesets
Codeset Name | Associated Mnemonic |
DEC Kanji | dec |
Shift JIS | SJIS |
Japanese EUC | eucJP |
7-bit JIS | jis7 |
8-bit JIS | jis8 |
UTF-8 | UTF-8 |
2.5.1.2 Setting the Application Codes
To set the application codeset, use the
stty
command
with the
-acode
option:
% stty acode codeset-name
Table 2-6
lists the codeset names that you can set.
Table 2-6: Possible Codeset Names for the Application Codesets
Codeset Name | Associated Mnemonic |
DEC Kanji | deckanji |
Super DEC Kanji | sdeckanji |
Shift JIS | SJIS |
Japanese EUC | eucJP |
UTF-8 | UTF-8 |
The Tru64 UNIX system interprets 8-bit data (that is, data with the eighth bit on) that was input through the terminal in one of two ways:
The data is interpreted as a 1-byte Kana character, and:
Passed to the application as a 1-byte, or half-width Kana character without any conversion
Converted into the 2-byte, or full-width Hiragana code associated with the Kana character of interest
Converted into the 2-byte, or full-width Katakana code associated with the Kana character of interest
The data is interpreted as part of a Kanji code of two or more bytes.
You can use the
stty
command with the
-imode
option to specify one of the following:
1-byte, or half-width Kana code
To pass the 8-bit data input as a 1-byte Kana character to the application,
use the following
stty
command:
% stty imode hankaku
Unless a DEC Kanji code that does not support a 1-byte Kana character has been set for the application code, the 1-byte Kana character represented by this code is converted into the associated 2-byte Katakana code.
2-byte, full-width Hiragana code
To convert a Kana character represented by the 1-byte (8-bit) input
data into the 2-byte Hiragana code, use the following
stty
command:
% stty imode hiragana
2-byte, or full-width Kana character
To convert a Kana character represented by the 1-byte (8-bit) input
data into the 2-byte Katakana character, use the following
stty
command:
% stty imode katakana
part of character of two or more bytes
To handle the 1-byte (8-bit) input data as part of a Kanji code when
DEC Kanji has been set for the terminal codeset, use the following
stty
command:
% stty imode kanji
2.5.3 Checking the Current Terminal Settings
To view the current terminal settings, execute the
stty
command without flags or options:
% stty atty disc;speed 9600 baud; -parity hupcl erase = ^? brkint -inpck -istrip icrnl -ixany onlcr echo echoe echok imode=hiragana tcode=dec acode=deckanji %
The settings displayed correspond
to the
stty
options.
The options prefixed by a hyphen (-)
are currently disabled; the others are currently enabled.
See
stty
(1)2.6 File Code Conversion
The Tru64 UNIX
iconv
command converts file codes.
The command uses the following format:
iconv -f input-code -t output-code file-name ...
The
iconv
command converts the input codeset to the output codeset and writes the resulting
data to standard output or to the specified filename(s).
You can specify any
codeset name in
Table 2-7.
Table 2-7: Possible Codeset Names for use with iconv
Codeset Name | Associated Mnemonic |
DEC Kanji | deckanji |
Super DEC Kanji | sdeckanji |
Japanese EUC | eucJP |
Shift JIS | SJIS |
7-bit JIS | JIS7 |
7-bit JIS [Footnote 2] | jiskanji7 |
ISO-2022-JP Japanese Mail Code | ISO-2022-JP |
ISO-2022-JP Extended Japanese Mail Code | ISO-2022-JPext |
UCS-4 | UCS-4 |
UCS-2 | UCS-2 |
UTF-8 | UTF-8 |
For example, the following command converts the
foobar.txt
Shift JIS coded text file into the
foobar.sdc
Super DEC Kanji text file:
% iconv -f SJIS -t sdeckanji foobar.txt >foobar.sdec
The
user-defined character areas vary from codeset to codeset.
When you use
iconv
for code conversion, make sure that the code values assigned
to the character codes of the user-defined characters are also converted.
For information on the user-defined character areas of the respective codesets
and the character code correspondence between the codesets, see
udc
(5)
In addition to the codesets listed in
Table 2-7, Tru64 UNIX
supports file code conversion into codesets used with machines other than
those manufactured by HP.
Table 2-8
lists these codesets.
Table 2-8: Possible Codeset Names for use with iconv (for Machines Not Manufactured by HP)
Codeset Name | Associated Mnemonic |
IBM Kanji System Character | ibmkanji |
Fujitsu JEF Code | JEF |
Hitachi KEIS | KEIS |
For details on the various codesets, see
iconv_ibmkanji
(5)iconv_JEF
(5)iconv_KEIS
(5)2.7 Message Catalogs
Programs output various messages during execution. When software contains message character strings embedded in the program, the contents of the output cannot vary according to the operating environment.
Instead of embedding message character strings in the program during execution, internationalized Tru64 UNIX commands and libraries reference a file containing logged message characters and output its contents in message form. A message catalog is the mechanism that references a file during execution and outputs the appropriate message.
Note
The contents of files containing messages have been compiled by the
gencat
command in binary form. Thus, they cannot be displayed using thecat
command. To view the file contents, use thedspcat
command.
During installation, Tru64 UNIX stores the
Japanese locale message catalogs in directories.
Table 2-9
describes the directory path associated with each of the Kanji codes specified
with a locale name
[Footnote 3]
.
Table 2-9: Directories Where Japanese Message Catalogs Exist
Kanji Code | Directory of Japanese Message Catalog |
Japanese EUC | /usr/lib/nls/msg/ja_JP |
DEC Kanji | /usr/lib/nls/msg/ja_JP.deckanji |
Super DEC Kanji | /usr/lib/nls/msg/ja_JP.sdeckanji |
Japanese EUC | /usr/lib/nls/msg/ja_JP.eucJP |
Shift JIS | /usr/lib/nls/msg/ja_JP.SJIS |
UTF-8 | /usr/lib/nls/msg/ja_JP.UTF-8 |
2.8 Setting Up the Japanese Reference Pages
Tru64 UNIX provides Japanese reference pages that describe the commands
and libraries available in the Japanese environment.
Reference pages are documents
you can view online using the
man
command.
For example:
% man n title
The document
associated with
title
is output to the terminal
screen.
The value
n
is the section number of the
reference page.
The reference pages are classified into the sections listed
in
Table 2-10.
Table 2-10: Reference Page Section Numbers
Classification of Contents | Associated Section Number |
Commands | 1 |
System Calls | 2 |
Library Routines | 3 |
File Formats | 4 |
nroff macros, etc. | 5 |
Special Files | 7 |
System Management Commands | 8 |
If the LANG environment variable is set for
use in the Japanese environment, the
man
command searches
for the directory that contains the simplified online operation file of the
Japanese reference page (/usr/share/ja_JP.deckanji/man
if LANG is ja_JP.deckanji) and displays the Japanese reference page.
If the LANG environment variable is set to ja_JP.SJIS, the
man
command searches for the
/usr/share/ja_JP.eucJP/man
directory and the codeset converter converts the format into SJIS.
When you need to view an English reference page, use
the
man
command with the
-M
option as follows:
% man -M /usr/share/man title
For details about the
man
command, see
man
(1)