You can now already use any Unicode characters in file names. No kernel
or file utilities need modifications. This is because file names in the
kernel can be anything not containing a null byte, and '/' is used to
delimit subdirectories. When encoded using UTF-8, non-ASCII characters
will never be encoded using null bytes or slashes. All that happens is
that file and directory names occupy more bytes than they contain characters.
For example, a filename consisting of five greek characters will appear
to the kernel as a 10-byte filename. The kernel does not know (and does
not need to know) that these bytes are displayed as greek.
This is the general theory, as long as your files stay inside Linux. On
filesystems which are used from other operating systems, you have mount
options to control conversion of filenames to/from UTF-8:
The "vfat" filesystems has a mount option "utf8".
See
file:/usr/src/linux/Documentation/filesystems/vfat.txt.
When you give an "iocharset" mount option different from the default
(which is "iso8859-1"), the results with and without "utf8" are not
consistent. Therefore I don't recommend the "iocharset" mount option.
The "msdos", "umsdos" filesystems have the same mount option, but it
appears to have no effect.
The other filesystems (nfs, smbfs, ncpfs, hpfs, etc.) don't convert
filenames; therefore they support Unicode file names in UTF-8 encoding only
if the other operating system supports them.
Recall that to enable a mount option for all future remount, you add to
the fourth column of the corresponding /etc/fstab line.
Ttys are some kind of bidirectional pipes between two program, allowing
fancy features like echoing or command-line editing. When in an xterm,
you execute the "cat" command without arguments, you can enter and edit
any number of lines, and they will be echoed back line by line. The
kernel's editing actions are not correct, especially the Backspace (erase)
key and the tab bey are not treated correctly.
if you are using glibc2, apply the patch
glibc211-tty.diff
and recompile your libc (or if you are not so adventurous, it is sufficient
to patch an already installed include file:
glibc-tty.diff),
apply the patch
stty.diff
to GNU sh-utils-1.16b, and rebuild the "stty" program, then test it using
"stty -a" and "stty iutf8".
add the command "stty iutf8" to the "unicode_start" script, and
add the command "stty -iutf8" to the "unicode_stop script.
apply the patch
xterm.diff
to xterm-109, and rebuild "xterm", then test it by starting
"xterm -u8"/"xterm +u8" and running "stty -a" and interactive "cat" inside it.
To make this fix persistent across rlogin and telnet, you also need to:
Define new values for the TERM environment variable, "linux-utf8" as an
alias to "linux", and "xterm-utf8" as an alias to "xterm".
If your system has the ncurses library and the /usr/lib/terminfo (or
/usr/share/terminfo) database, do this by running
as non-root (this will create the terminfo entries in your $HOME/.terminfo
directory). Here are
linux-utf8.terminfo and
xterm-utf8.terminfo.
I don't recommend running this as root, because it will create
the terminfo entries in /usr/lib/terminfo where they might be erased next
time you upgrade your system.
If your system has an /etc/termcap file, you should also edit that file:
copy the linux and xterm entries and give them the new names "linux-utf8"
and "xterm-utf8". For an example, see
termcap.diff.
Each time you call "unicode_start" or "unicode_stop" from the console, also
execute "export TERM=linux-utf8" or "export TERM=linux", respectively.
Apply the patch
xterm2.diff
to xterm-109, rebuild "xterm", and remove any "XTerm*termName" line from
/usr/X11R6/lib/X11/app-defaults/XTerm and $HOME/.Xdefaults. Now xterm sets
the TERM variable to "xterm-utf8" instead of "xterm" when running in UTF-8
mode.
Apply the patches
netkit.diff,
netkitb.diff and
telnet.diff and
rebuild "rlogind" and "telnetd". Now rlogin and telnet put the tty into
UTF-8 editing mode whenever the TERM environment variable is "linux-utf8"
or "xterm-utf8".
You will need a program to convert your locally (probably ISO-8859-1) encoded
texts to UTF-8. (The alternative would be to keep using texts in different
encodings on the same machine; this is not fun in the long run.)
One such program is `iconv', which comes with glibc-2.1. Simply use
Here are two handy shell scripts, called
"i2u"
i2u.sh (for ISO to UTF conversion) and
"u2i"
u2i.sh (for UTF to ISO conversion).
Adapt according to your current 8-bit character set.
individual variables for:
character types and encoding,
natural language messages,
sorting rules,
number formatting,
money amount formatting,
date and time display
LANG
default value for all LC_* variables
(See `man 7 locale' for a detailed description.)
Each of the LC_* and LANG variables can contain a locale name of the
following form:
language[_territory[.codeset]][@modifier]
where language is an
ISO 639 language code
(lower case), territory is an
ISO 3166
country code (upper case), codeset denotes a character set, and
modifier stands for other particular attributes (for example indicating
a particular language dialect, or a nonstandard orthography).
LANGUAGE can contain several locale names, separated by colons.
In order to tell your system and all applications that you are using UTF-8,
you need to add a codeset suffix of UTF-8 to your locale names. For example,
if you were using
If you have glibc-2.1 or glibc-2.1.1 or glibc-2.1.2 installed, first check
using "localedef --help" that the system directory for character maps is
/usr/share/i18n/charmaps. Then apply to the file /usr/share/i18n/charmaps/UTF8
the patch
glibc21.diff
or
glibc211.diff
or
glibc212.diff, respectively.
Then create the support files for each UTF-8 locale you intend to use, for
example:
You typically don't need to create locales named "de" or "fr" without
country suffix, because these locales are normally only used by the
LANGUAGE variable and not by the LC_* variables, and LANGUAGE is only
used as an override for LC_MESSAGES.
The glibc-2.2 will support multibyte locales, in particular the UTF-8 locales
created above. But glibc-2.1 or glibc-2.1.1 do not really support it.
Therefore the only real effect of the above creation of the
/usr/share/locale/de_DE.UTF-8/* files is that `setlocale(LC_ALL,"")'
will return "de_DE.UTF-8", according to your environment variables, instead
of stripping off the ".UTF-8" suffix.
To add support for the UTF-8 locale, you should build and install the
`libutf8_plug.so' library, from
libutf8-0.5.2.tar.gz.
Then you can set the LD_PRELOAD environment variable to point to the
installed library:
Then, in every program launched with this environment variable set, the
functions in libutf8_plug.so will override the original ones in
/lib/libc.so.6. For more info about LD_PRELOAD, see "man 8 ld.so".
This entire thing will not be necessary any more once glibc-2.2 comes out.
Now let's fill these new locales with content. The following /bin/sh
commands can convert your message catalogs to UTF-8 format. They must
be run as root, and require the programs `msgfmt' and `msgunfmt' from
GNU gettext-0.10.35 to be installed.
convert-msgcat.sh
This too will not be necessary any more once glibc-2.2 comes out, because
by then, the gettext function will convert the strings appropriately from
the translator's character set to the user's character set, using either
iconv or librecode.
