cyrillic / unicode filenames

[Colt45]

I can't seem to have files with cyrillic filenames.

My locale is en/gb/UTF8... I've got unicode fonts installed. I'm using "unicode" xterm.



If i have a tarball with files inside with cyrillic names, and i untar it, the unicode letters are replaced with ???, and directories are replaced with _.
So eventually tar gets two files named ????.???, and skips a file cause it tries to overwrite the other, etc.


If I try to transfer a unicode filename file from windows to the linux box with SFTP, it bitches that such a file exists (and it doesn't), normal ascii files transfer fine.


any ideas here?


oh yeah.. If i cat a text file with cyrillic text inside, it shows up fine in the terminal. It's just the filenames that don't work.

 

[Nothinman]

Are you trying to extract the files to a FAT filesystem?

 

[Colt45]

Nope, its ext3.

I know you have to force fat or ntfs as utf8 when you mount it, but there's nothing like that for linux partitions is there?

 

[Nothinman]

I don't believe so, I just created a file called ???????????.test on XFS and ext3 and both worked fine.

 

[Colt45]

Crap.. I just saved a file with cyrillic letters copied out of charmap, and it worked


So maybe the files inside the tarball are using some other encoding for the filename? cp-1251 or something.. It does show up on windows.

 

[Nothinman]

AFAICT there's no encoding information in the tar file format.

 

[Colt45]

well i sort of lied.
it was a .7z

I tried p7zip, it had the same problem. so un unzipped it in windows, filenames showed up fine, and then i made a tarball in windows, and it still has the same problem in linux.



Anyways, I just tried p7zip again and it works now, go figure.