Belarusian-HOWTO

Belarusian-HOWTO

Alexander Mikhailian,

mikhailian@altern.org

v.0.1.5, 25 February 2001

Short guide in setting up Belarusian language support in Linux

console, X Window System, web-browsers, text editors, etc. Charsets used in the setup are either windows-1251 or iso-8859-5. Belarusian characters can also be found in koi8-ub, koi8-c, koi8-ru and, of course, Unicode. Although this HOWTO is Linux-specific, many advices are applicable to other UNIX-like systems.

----; Table of Contents; 1. Introduction: ; 1.1. History; 1.2. Comments; 1.3. New Versions; 1.4. Copyright; 1.5. Acknowledgements and Thanks; 2. System-wide setup: ; 2.1. Choosing Charset; 2.2. windows-1251 in the kernel; 2.3. ISO-8859-5 in the kernel; 2.4. Setting locale; 2.5. Belarusian in console; 2.6. Belarusian in X Window System; 3. Editing texts: ; 3.1. Emacs; 3.2. Spell-checking; 3.3. TeX; 4. Browsers: ; 4.1. Netscape; 4.2. Lynx; 5. Mailers: ; 5.1. Netscape Messenger; 5.2. Mutt; 6. Further support

1. Introduction

1.1. History

This document was started in September 15, 1999 by Alexander Mikhailian

1.2. Comments

Comments on this HOWTO may be directed to the author

mikhailian@altern.org.

1.3. New Versions

The newest version can always be found at

the bellinux page.

1.4. Copyright

This manual may be reproduced in whole or in part without restrictions.

1.5. Acknowledgements and Thanks

Thanks to everyone who gave comments as I was writing this.

2. System-wide setup

2.1. Choosing Charset

As for now, the choice is mainly between windows-1251 and

iso-8859-5. Windows-1251 provides compatibility with M$ Windows and it is is by far the most popular charset for nearly all slavic languages. On the other side, iso-8859-5 is better supported and easier to set up.

The present HOWTO will explain the use of both charsets in parallel. Unless otherwise stated in the text, Belarusian support package from the bellinux page contains all the files mentioned in the present HOWTO.

2.2. windows-1251 in the kernel

Windows-1251 support in the kernel is needed to visualize MS

Windows filenames in cyrillic while the console and X Window system are localized in windows-1251. As latest FAT file systems store filenames in Unicode, we have to define the output charset of the Virtual File System layer and install the support for windows-1251 in the kernel.

There is a patch for 2.2.14 kernels that is included in the

Belarusian support package. To apply the patch,

*

cd /path_to_your_kernel_source_tree

* *

patch -p0 ` patch_for_the_kernel.patch

* *

Recompile and reinstall the kernel modules.

Refer to Kernel-HOWTO for the details.

* *

Add

codepage=866,iocharset=microsoft-cp1251

to your mount options in order to get the filenames in windows-1251.

To test the setup, execute

mount /dev/hda -t vfat -o codepage=866,iocharset=iso8859-5 /mnt/hda

as root. Of course, you should replace /dev/hda by the name of your vfat partitions containing names in cyrillic.

*

2.3. ISO-8859-5 in the kernel

There is built-in support for iso-8859-5 in Linux kernel.

To enable iso-8859-5, make sure you compile this module in the nls section of the kernel setup. Normally, iso-8859-5 is compiled by default and can be loaded either automatically or by issuing

insmod nls_iso8859-5

as root.

Add

codepage=866,iocharset=iso8859-5

to your mount options in order to get the filenames in iso-8859-5.

2.4. Setting locale

*

Belarusian locale is available in glibc 2.2 and later.

If you do not have the belarusian locale installed in your system, you can compile it yourself. The source is available at the bellinux page. Issue

localedef -f CP1251 -i be_BY be_BY.CP1251

or

localedef -f ISO-8899-5 -i be_BY be_BY.ISO-8899-5

to compile the locale in windows-1251 encoding or iso8859-5 encoding.

* *

Check how it works by setting the locale

set LANG=be_BY.CP1251; export LANG

or

set LANG=be_BY.ISO-8859-5; export LANG

and running a locale-aware program like date or cal.

*

2.5. Belarusian in console

2.5.1. Setting Belarusian with windows-1251

Some cyrillic console fonts are in fact russian fonts and

lack many cyrillic characters. Fortunately, !UniCyr? fonts by Vadinm Zhitnikov have all the glyphs from cp866, cp1251, iso8859-5 and all printable symbols from koi8-r. Besides the belarusian keyboard maps which you can find at the the bellinux page , other files are fairly standard and are available in most linux distributions. To set up window-1251 in console,

*

Load a !UniCyr? font

consolechars -f !UniCyr?_8x16

* *

Load by2_win.kmap keymap.

loadkeys by2_win

* *

Load Application-Charset Map cp1251.acm

consolechars -m cp1251

*

2.5.2. Setting Belarusian with iso-8859-5

There are two ways to set up Belarusian with iso-8859-5

*

Load iso-8859-5 font

consolechars -f iso05.f16

* *

Load belarusian keymap

by2.kmap or by.kmap

loadkeys by.kmap

loadkeys by2.kmap

* This method has one serious drawback - you will loose all pseudographic characters and, say, you Midnight Commander will look somewhat naked. The second method described below preserves all pseudographic characters:

*

Load !UniCyr? font

consolechars -f !UniCyr?_8x16

* *

Load by2.kmap

keymap or by.kmap

loadkeys by.kmap

loadkeys by2.kmap

* *

Load Application-Charset Map iso05.acm

consolechars -m iso05

* Beware that different Linux distributions have different console-related packages - console-tools or kbd. Abovementioned scripts are meant to work with console-tools which is by far more popular.

2.6. Belarusian in X Window System

2.6.1. ISO-8859-5 in X Window System through XKB

This is is quite easy to set up.

*

Install ISO-8859-5 fonts for X Window and make sure they are before other fonts in the fontpath. Refer to Font-HOWTO for details on how to install fonts.

* *

Get the belarusian keyboard layout by from the

the bellinux page and put it on the place of any other keyborad layout in /usr/X11R6/lib/X11/xkb/symbols/, e.g. instead of the belgian keyboard layout /usr/X11R6/lib/X11/xkb/symbols/be

Put the following stings in your

/etc/X11/XF86COnfig

XkbModel? "microsoft"

XkbLayout? "be"

XkbOptions? "grp:caps_toggle"

* *

Don't forget to set up the system locale to

be_BY.ISO8859-5 as described in Section 2.4. You will also want to make sure that iso-8859-5 fonts are before any other fonts in your font path.

2.6.2. Windows-1251 in X Window System through XKB

This works only with XFree 4.0.2 and higher.

*

Install ISO-8859-5 fonts for X Window and make sure they are before other fonts in the fontpath.

* *

Get the belarusian keyboard layout by from the

the bellinux page and put it on the place of any other keyborad layout in /usr/X11R6/lib/X11/xkb/symbols/, e.g. instead of the belgian keyboard layout /usr/X11R6/lib/X11/xkb/symbols/be

* *

Put the following strings in your

/etc/X11/XF86Config

XkbModel? "microsoft"

XkbLayout? "be"

XkbOptions? "grp:caps_toggle"

* *

Don't forget to set up the system locale to be_BY.CP1251

as described in Section 2.4.

*

To use windows-1251 with XFree 3.3.5 and 4.0, you have to

apply a patch from Aleksey Novodvorsky that allows the use of windows-1251 with XKB. The original location is at ftp.logic.ru/pub/logic/linux/be-locale and it is also available from the bellinux page. Lucky users of Linux-Mandrake RE get a patched XFree86 out of box.

Belarusian keyboard layout is palnned for all XFree releases

after 4.0.2. To make it work, you will have to add the following lines

into /etc/X11/XF86Config

XkbModel? "microsoft"

XkbLayout? "by"

XkbOptions? "grp:caps_toggle"

2.6.3. Windows-1251 in X Window System through Xmodmap

*

Uncomment

XkbDisable?

line in your /etc/X11/XF86COnfig.

*

Install windows-1251 fonts for X Windows and

make sure thay are before other fonts in the path.

* *

Make and install xruskb package which can be

downloaded from http://bellinux.sourceforge.net

* *

Replace .xmm files in your xruskb directory by those

found at the bellinux page.

* *

Add the following lines in your .Xdefaults file

xrus*modeButton1.labelString: BEL xrus*modeButton1.label: BEL

* *

Run

xrus jcuken-cp1251

or

xrus jcuken-iso5

to start keyboard switcher.

* *

3. Editing texts

3.1. Emacs

Emacs is able to display Belarusian characters out of box in many

Linux distributions. However, the following conditions should satisfy:

*

Cyrillic iso-8859-5 fonts should be installed.

* *

Emacs-mule package should be installed.

*

The user may want to change the default keyboard layout

("input method" in emacs slang) to jcuken. This feature is provided by belarusian.el along with some other goodies. Read the comments in the beginning of belarusian.el for details on the setup procedure.

After installing belarusian.el, you will

be able to switch the input method, by typing C-\-cyrillic-belarusian-RET

For a more user-friendly setup, put the following lines in the

.emacs file

;; enables PC-like key mappings

(pc-bindings-mode) ;; enables PC-like selection (pc-selection-mode) ;; highlight selection (transient-mark-mode t) ;; use as many multilingual fonts as possible (set-default-font "fontset-standard") ;; set up windows-1251 as default encoding everywhere (codepage-setup 1251) (set-terminal-coding-system 'cp1251) (set-keyboard-coding-system nil) (setq default-buffer-file-coding-system 'cp1251) (set-selection-coding-system 'cp1251) (prefer-coding-system 'cp1251) (define-coding-system-alias 'windows-1251 'cp1251) (define-coding-system-alias 'microsoft-1251 'cp1251) (define-coding-system-alias 'microsoft-cp1251 'cp1251) (define-coding-system-alias 'windows-cp1251 'cp1251) (define-coding-system-alias 'koi8-u 'cyrillic-koi8) (define-coding-system-alias 'KOI8-R 'cyrillic-koi8) (modify-coding-system-alist 'file "\.*" 'cp1251) (modify-coding-system-alist 'file "\\.el$" 'iso-2022-7bit)

This configuration is for windows-1251 setup.

3.2. Spell-checking



*

If you do not have ispell 3.1.20, install it from any Linux

distribution CD.

* *

Download the package

belspell.tar.gz. Unpack it and run

buildhash belarusian.sml belarusian.aff belarusian.hash

* *

Copy

belarusian.hash to `/usr/lib/ispell or wherever your ispell hash tables are.

* *

in order to spell-check a file, type

ispell -d belarusian yourfile.txt

*

Getting belarusian ispell dictionary working with emacs is a bit

tricky.

*

Copy

ispell.el and ispell.elc into your $EMACSBIN/../lisp directory.

* *

Copy temporarily

$EMACSBIN/../lisp/loaddefs.el to /usr/src/emacs-XX.X/lisp/loaddefs.el where XX.X stand for emacs version number.

* *

M-x-update-file-autoloads-RET

and enter the path to the ispell.el.

* *

Copy back loaddefs.el

* *

Restart emacs

*

3.3. TeX

from Aleksey Novodvorsky:

You need TeX + babel + T2, e.g. teTeX b= 1.0. in order to get partial

support for Belarusian,

*

put

babel.sty , belarusianb.ldf and belarusianb.sty into /usr/share/texmf/tex/generic/babel

* *

run

texhash

* *

To use Belarusian in LaTeX: add the

following lines in the preamble:

\documentclass[belarusian?{article} \usepackage[cp1251?{inputenc} \usepackage{babel}

or

\documentclass[belarusian?{article} \usepackage[iso88595?{inputenc} \usepackage{babel}

*

4. Browsers

4.1. Netscape

It is often a problem to correctly visualize

Belarusian-specific characters with Netscape. This is due to that Netscape 4.x looks for any koi8-r font and, if it finds one, they refuse to use other fonts for displaying Cyrillic pages.

The problem is solved by removing koi8-r fonts from everywhere - X

Window System font path and font server path.Then install koi8-ru, iso-8859-5 or windows-1251 and run

xset fp rehash

or restart X server. You will probably have to delete bpreferences and preferences.js from the Netscape home directory .netscape and even then you are not guaranteed to have Netscape work right. Netscape's handling of fonts has always been an obscure issue.

Links to koi8-ru, iso-8859-5 and windows-1251 cyrillic fonts for X

Window System can be found at the bellinux page

4.2. Lynx

In order to view Belarusian sites, you have to set up you console to

handle the encoding you need. See section Section 2.5 for more details.

Add the following lines to your .linxrc file.

character_set=Cyrillic (windows-1251) preferred_language=be preferred_charset=windows-1251

or

character_set=Cyrillic (ISO-8859-5) preferred_language=be preferred_charset=iso-8859-5

if you have set up iso-8859-5 in console.

If the page you are browsing does not have an explicit charset

declaration, press o and set the document charset manually.

5. Mailers

5.1. Netscape Messenger

Netscape Messenger can not handle belarusian texts properly.

The same applies to Mozilla.

5.2. Mutt

Mutt handles a multitude of charsets and encodings with ease. If your

console has windows-1251 support, add the following lines to your

.muttrc

charset-hook "windows-1251" "cp1251" set charset="windows-1251"

Mutt automatically converts all incoming messages for the screen

output. It can also convert outgoing messages basing itself on a versatile pattern matching mechanism, e.g. the following line in .muttrc will force mutt to convert all messages sent to be-locale@iatp.unibel.by to koi8-r.

send-hook 't ^be-locale@iatp\.unibel\.by$' 'set send_charset="koi8-r"'

6. Further support

A mailing list devoted entirely to Belarusian language support

is available at be-locale@iatp.unibel.by. To subscribe, send a message to majordomo@iatp.unibel.by with the string "subscribe be-locale youremailaddress" in the body of the message

Another mailing list that treats mostly linguistic issues is

available at movaznaustva@egroups.com. To subscribe to it, send a message with empty body to movaznaustva-subscribe@egroups.com.