[Bug-tar] Feature Request: specifying encoding of file names inside the .tar file

[范德成]

I am currently using Fedora Core 2 and I find that
exchanging tar balls
between Windows and Fedora is difficult because my
Windows uses Chinese
(CP936) encoding while Fedora uses mostly UTF-8
encoding. Tar balls made
under Windows store file names in CP936 so it would be
fussy to make
them expanded under Fedora. A workaround is to remount
the file system
and set use the -o switch of `mount' to specify
iocharset=cp936, and
then export LC_ALL=zh_CN.GBK, (in my case CP936 equals
GBK), and then
use tar to extract the files.

Anyway, I think such a problem could be solved without
remounting a
partition. The easiest way I think is to add an option
for tar that
specifies the encoding to use when packing and
unpacking. So I request
this feature: add an option to tar that specifies the
encoding to use
when making tar balls and extracting from tar balls
while the file
system IO character set still remains the same as the
locale value. e.g.
if the option is like -o charset, then the command
line would be like this:

# locale
LANG=zh_CN.UTF-8
LC_ALL=zh_CN.UTF-8
...
# tar -c -o charset=cp936 .>packed.tar
...
# tar -x -o charset=cp936<packed.tar
...

Thanks for reading this message.





        

        
                
___________________________________________________________ 
雅虎免费G邮箱－中国第一绝无垃圾邮件骚扰超大邮箱 
http://cn.mail.yahoo.com