xenocara/app/xterm/README.i18n

-- $XFree86: xc/programs/xterm/README.i18n,v 1.1 2003/11/13 01:16:37 dickey Exp $

Using xterm in your language
============================

Since XFree86 version 4.0, the internationalization (i18n) feature of
xterm is gradually improved.  Xterm is being improved even now.  You
need only set the standard locale environment variables such as
LC_CTYPE, LC_ALL, LC_CTYPE, or LANG.  Once the locale is set up you can
use xterm in your favorite character encoding.

This document explains how the i18n feature is realized and how to
configure xterm for your character encoding.

Refer to locale(7) for details of the locale mechanism.


Basic i18n-related settings and resources
=========================================

These settings apply to XFree86 xterm patch #181, and the program luit
which is distributed with XFree86 4.4

1.  Usage of "locale mode"

    On startup, xterm must be in "locale mode" to make it follow the
    current locale.  You can invoke xterm in locale mode in these ways:

    a.  Set "vt100.locale" resource "true".  This resource was
        introduced since XFree86 4.3.  The default value of the "locale"
        resource is "medium", which means xterm follows the locale only
        in Chinese, Japanese, Korean, or Thai locales.  For example,

          XTerm*locale: true

        in your ~/.Xresources file.

    or

    b.  Invoke xterm with the "-lc" option.

2.  Converter program "luit"

    The "luit" must be available in the standard XFree86 binary
    directory.  It is usually available because it is part of the
    XFree86 distribution.  The standard binary directory may differ from
    system to system.  /usr/X11R6/bin/luit is an example.

    "luit" is used to convert between Unicode and the character encoding
    for your locale.  When built for XFree86, xterm includes logic for
    invoking luit.

3.  Locale setting

    Finally, you will need to configure your locale.  We expect that you
    have already configured your locale for other software.  For example,

      LANG=de_DE@euro
      export LANG

    in your ~/.xsession file.  There are many ways to configure locale. 
    For example, your display manager may have a mechanism to invoke a
    window manager in your favorite locale, or you may have system-wide
    locale setting in /etc/environment.  You may also have set the
    LC_ALL variable instead of the LANG variable.


How to use xterm in different locale temporarily
================================================

You may sometimes need to invoke xterm in a different character encoding
than your current locale.  For example, use xterm to login remote systems
in different locale.

Do this by invoking xterm in the target locale.  For example,

  $ LANG=ru_RU.KOI8-R xterm &

Previously, font setting has been used in such cases.

  $ xterm -fn -misc-fixed-medium-r-normal--10-*-*-*-*-*-koi8-r &

This does not work well in conjunction with the "locale" resource,
because luit and xterm combined rely upon Unicode fonts.


How to set fonts for UTF-8/locale modes
=======================================

Since xterm patch #181, xterm can automatically use Unicode fonts in
UTF-8 mode and locale mode.  Few of you will need to modify the default
setting to display your language.  In particular, Unicode fonts in
combination with locale mode will satisfy the needs of not only
ISO-8859-1 users but also East Asian and other non-ISO-8859-1 users.

If you want to set your favorite Unicode font for UTF-8 and locale
modes, you should add a line such as the following in your ~/.Xresources
file:

    XTerm*VT100.utf8Fonts.font: \
         -misc-fixed-medium-r-semicondensed--13-120-75-75-c-60-iso10646-1

The leading "XTerm*" pattern is more specific than the system's
app-defaults file, therefore it overrides the corresponding line
beginning with

    *VT100.utf8Fonts.font:

Here is an additional note.  If you want to display East Asian
doublewidth characters (CJK Ideogram, Hiragana, Katakana, Hangul,
and so on), we recommend using

    -misc-fixed-medium-r-semicondensed--13-*-*-*-*-*-iso10646-1

or

    -misc-fixed-medium-r-normal--18-*-*-*-*-*-iso10646-1

because these two fonts have corresponding doublewidth fonts.  These
fonts are used as default font and default "Large" font, respectively.


The internals of xterm i18n
===========================

You do not need to read this section if you only want to configure your
xterm.  Here we describe how xterm is implemented to support i18n.

The original version of xterm does not support locale or character
encoding.  Its I/O stream is interpreted as a mere 8-bit index for a
font.

Beginning with XFree86 4.0, xterm supported UTF-8.  It was implemented
as a separate UTF-8 mode from the conventional 8-bit mode.  Character
encodings had no effect on the 8-bit mode.  The UTF-8 mode has been
extended to support doublewidth characters (for East Asian characters)
and combining characters (such as accents for Latin alphabets and Thai
vowels/tone marks).

Doublewidth characters are characters that occupy two continuing
columns on the terminal.  Xterm uses separate fonts for normal
(singlewidth) characters and doublewidth characters.  Though xterm has
configuration items for specifying doublewidth fonts, it will
automatically search for a font with exactly twice as wide and the same
name as the specified normal font.

The default behavior of xterm was modified to use this UTF-8 mode in
UTF-8 locales.  A command line option of "-u8" and a resource of "utf8"
were introduced to choose UTF-8 mode.

"luit" was introduced to XFree86 at version 4.2.  It converts between
UTF-8 and other encodings.  When luit is invoked in a UTF-8 terminal,
the terminal acts as if it is really running in the other encoding.

Since XFree86 version 4.3, xterm provides a new mode to invoke luit
automatically to support various encodings.  The mode where xterm
invokes luit is called "locale mode".  It is the third mode following
conventional 8-bit mode and UTF-8 mode.  In the locale mode, xterm is
aware of the current locale and character encoding.  Since locale mode
uses luit, it is based on the UTF-8 mode.  That is, xterm works in UTF-8
mode and luit works as a converter between UTF-8 and the character
encoding for your locale.  This is why the locale mode always needs
Unicode fonts.  The default behavior of xterm is modified so that the
"locale mode" will be adopted in Chinese (Big5 and GB2312), Japanese
(EUC-JP), Korean (EUC-KR), and Thai (ISO-8859-11, as known as TIS-620)
locales.  Locale mode is chosen for these character encodings because
these encodings are not supported by conventional 8-bit mode even by
changing fonts (ISO-8859-11 needs combining characters and others need
doublewidth characters).

To control the locale mode, command line options of "-lc" and "-en" and
a resource of "locale" were introduced.  The command line option of
"-u8" and a resource of "utf8" were made obsolete by them, though
retained for compatibility.

Since XFree86 version 4.4, xterm can have two sets of default fonts,
one for conventional 8-bit mode and another for UTF-8 and locale modes,
by introducing the "utf8Fonts" subresource.


Future TODO Items
=================

We anticipate that xterm's locale mode will be used increasingly in the
future.  Since the UTF-8 and locale modes use more resources than
conventional 8-bit mode (because it needs larger fonts and another
process "luit"), faster hardware may be needed to gain complete
acceptance by users.  However, the locale mechanism allows users
to manipulate data in a standard form.  Its usefulness compensates
in part for reduced performance.

Xterm supports antialiased fonts ("-fa" and "-fs" command line options). 
Currently UTF-8 nor locale modes do not work with antialiased fonts.

Xterm does not support bi-directional or RTL languages such as Hebrew
and Arab.  A simple standard how terminal should behave for these
languages is needed.

Xterm does not support Unicode characters above U+10000.