[CentOS] Re: ls : not UTF-8 compliant?

Thu Feb 21 00:01:49 UTC 2008
Mufit Eribol <hme at onart.com.tr>

Scott Silva wrote:
> on 2/20/2008 3:14 PM Mufit Eribol spake the following:
>> Michael A. Peters wrote:
>>> Mufit Eribol wrote:
>>>> Sorry bugging you for this simple command.
>>>>
>>>> ls command displays question marks for the local characters (ones 
>>>> not included in 8859-1 space) in filenames.
>>>>
>>>> ie.
>>>> [root at server aa]# touch �arp
>>>> [root at server aa]# ls
>>>> ??arp
>>>> [root at server aa]# ls -b                    #for octal escapes
>>>> \303\247arp
>>>> [root at server aa]#
>>>>
>>>> However, ls|less, ls|more or vi <directory name> all display 
>>>> filename correctly. Also, the <tab> completes such filenames in the 
>>>> correct way. Even, logsave command for the ls output prints the 
>>>> right characters.
>>>>
>>>> So, I assume the filesystem keeps the filenames in UTF-8 encoding, 
>>>> but somehow ls can not show them properly.
>>>>
>>>> Any workaround or a replacement for ls? BTW The system is Centos 
>>>> 5.1 and locale shows the encoding as UTF-8.
>>>>
>>>> Thank you.
>>>
>>> Works for me.
>>>
>>> [mpeters at jerusalem tmp]$ touch �arp
>>> [mpeters at jerusalem tmp]$ ls
>>> çarp
>>> [mpeters at jerusalem tmp]$ echo $LANG
>>> en_US.UTF-8
>>> [mpeters at jerusalem tmp]$
>>> _______________________________________________
>>> CentOS mailing list
>>> CentOS at centos.org
>>> http://lists.centos.org/mailman/listinfo/centos
>> Interesting! Perhaps it is a quirk of ssh using PuTTY. I haven't 
>> tried it on the monitor connected. Did you try in on the monitor and 
>> CLI (no X, no Gnome etc)?
> Remember that putty defaults to an iso character set unless you change 
> the defaults.
>

No way! I use UTF-8 for "Character set translation on received data" of 
PuTTY. Centos is a fresh install with the default LANG setting. What 
else should I try?
Thank you for your support!