UTF-8 locale support

[Dima Veselov]

Hello!

My servers keep many files in cyrillic naming. Serving big user loads
it is hard to keep files in old encodings with outside world is already
living in UTF-8. Storing files not in UTF-8 cause some problems with Samba
and fatal problems with Linux & NFS, which don't have conversions at all.

So I feel there is time to move on UTF-8 on NetBSD too, and it seems
NetBSD 6 has ru_RU.UTF-8 support, however it is still is not complete.

Fresh installed 6.1.4 can store files in UTF-8. It also can share these
via SMB or NFS, but I can't make it work in shell.

As I see it has support only for LC_CTYPE and LC_MESSAGES via locale.alias
having no native ru_RU.UTF-8 support.

My linux rxvt-unicode terminal (working locally as expected) with ssh to
NetBSD box show:

[***@gloria ~]$ locale
LANG="ru_RU.UTF-8"
LC_CTYPE="ru_RU.UTF-8"
LC_COLLATE="C"
LC_TIME="C"
LC_NUMERIC="C"
LC_MONETARY="C"
LC_MESSAGES="ru_RU.UTF-8

This cause cyrillic filenamse being shown good, but I cannot access it,
because shell print hex code (f.e. \:\262\321\320) instead of letters.
Bash is 4.3.0(1) out of the box. (By the way https://wiki.netbsd.org/unicode/
says it will work out of the box)

Two questions on that:

Am I right and aliasing ru_RU.UTF-8 to en_US.UTF-8 make this that bad?

If I am right - what I shall do to complete ru_RU.UTF-8 locale and have
no problems in writing cyrillic filenames?

[Sergey Hromov]

В Fri, 15 Aug 2014 03:44:17 +0400

Post by Dima Veselov
My linux rxvt-unicode terminal (working locally as expected) with ssh
LANG="ru_RU.UTF-8"
LC_CTYPE="ru_RU.UTF-8"
LC_COLLATE="C"
LC_TIME="C"
LC_NUMERIC="C"
LC_MONETARY="C"
LC_MESSAGES="ru_RU.UTF-8
This cause cyrillic filenamse being shown good, but I cannot access
it, because shell print hex code (f.e. \:\262\321\320) instead of
letters. Bash is 4.3.0(1) out of the box. (By the way
https://wiki.netbsd.org/unicode/ says it will work out of the box)
Am I right and aliasing ru_RU.UTF-8 to en_US.UTF-8 make this that bad?
If I am right - what I shall do to complete ru_RU.UTF-8 locale and
have no problems in writing cyrillic filenames?

Dont' tried rxvt, but using xfce4-terminal I can operate with files on
NetBSD file server (with LC_ALL=ru_RU.utf-8 and ksh as shell
for user on it) using ssh. But with this combination there is some
issues, for example if I

$ touch тест

$ mv тест тес2 (here used tab for autocompletion and
backspace for deleting last letter "т")

$ ls
тес?2

Looks like backspace can't handle two-byte coded symbols.