how to convert filename mojibake back into original encoding??

wordsmithy · 05-12-2008, 11:45 PM

I used the Japanese version of slackware to back up all of my japanese-encoded files onto an external device but when I try to access those files they are not in the encoding windows or mac are expecting so the filenames show up as garbage(the contents are intact and okay, just the filenames are messed up). If I access the device from slax, it can read them so I think the issue is that slax/linux in general is by default in EUC or UTF-8 and windows and mac are looking for shift-JIS? I'm not an expert on this and it usually "just works" so I am looking for the easiest way to convert or have these files recognized outside of a linux environment. The original files can't be recopied, either, so I am stuck. Any suggestions?

hro · 05-13-2008, 02:13 AM

You can use 'recode' command within script to change encoding of the filenames. You should make some testing and read 'info recode' or at least 'man recode'. You will need more complicated script in order to traverse directories. Here is example scripts to rename files using ISO-8859-1 encoding and to rename them back using default encoding.

Code:

~/xDocs/test> ll *.txt
-rw-r--r-- 1 hro users 0 2008-05-13 09:22 ÄÖÜ.txt
-rw-r--r-- 1 hro users 0 2008-05-13 09:13 öäå.txt
~/xDocs/test> cat utf2iso.sh
#!/bin/bash
# hro, 2008-05-13
#
for fin in *.txt
        do (
        fout=$(ls $fin | recode ISO-8859-1)
        mv $fin $fout
        ) done

~/xDocs/test> utf2iso.sh
~/xDocs/test> ll *.txt
-rw-r--r-- 1 hro users 0 2008-05-13 09:13 Ã¶Ã¤Ã¥.txt
-rw-r--r-- 1 hro users 0 2008-05-13 09:22 Ã?Ã?Ã?.txt

Code:

~/xDocs/test> cat iso2utf.sh
#!/bin/bash
# hro, 2008-05-13
#
for fin in *.txt
        do (
        fout=$(ls $fin | recode ..ISO-8859-1)
        mv $fin $fout
        ) done

~/xDocs/test> iso2utf.sh
~/xDocs/test> ll *.txt
-rw-r--r-- 1 hro users 0 2008-05-13 09:22 ÄÖÜ.txt
-rw-r--r-- 1 hro users 0 2008-05-13 09:13 öäå.txt