Garbled characters with ja_JP.eucJP encoding

[bukatlib]

tcsh garbles characters when eucJP encoding is used.
$ LANG=ja_JP.eucJP LC_ALL=ja_JP.eucJP tcsh
tcsh $ unknown_command
unknown_command: <garbled_characters>.

Note that it works when UTF-8 encoding is used instead.
$ LANG=ja_JP.UTF-8 LC_ALL=ja_JP.UTF-8 tcsh
tcsh $ unknown_command
unknown_command: コマンドが見つかりません.

I got the same results on Solaris and Linux with tcsh 6.22.02 installed.
It looks like that only UTF-8 is considered in NLSClassify function (commit 4b12ecb).

Thank you for your help and feedback.

[zoulasc]

How do I reproduce this? What do I need to install and run on what OS? christos

…

On Nov 9, 2020, at 8:44 AM, Libor Bukata ***@***.***> wrote: LANG=ja_JP.UTF-8 LC_ALL=ja_JP.UTF-8 tcsh

[bukatlib]

On Solaris:
$ nlsadm set-system-locale ja_JP.eucJP
$ init 6

On Fedora Linux:
$ dnf install vlgothic-fonts
$ localectl set-locale LANG=ja_JP.eucJP

Afterwards, you can run the following command to see garbled characters:
LANG=ja_JP.eucJP LC_ALL=ja_JP.eucJP tcsh
tcsh $ unknown_command
unknown_command: <garbled_characters>.