0

I cannot find any documentation on the encoding of file names in zip files.

I have to pack file names with German umlauts into a zip, load it into an Android and unpack there. Then I have to refer to these files with names that are stored in another file. The latter is UTF-8, so there is no problem with umlauts. But I need to make sure, that the unzipped files have the proper names after unpacking on the Android.

Currently, I am using info-zip on Windows 7 to pack the archive, and the API functions on the Android to unpack, and everything worked. However, if I use jar from the Java SDK or 7z for the packaging, the files come out on the Android with wrong names. This makes me wonder, if there is any proper solution for this at all (besides using antique ASCII file names only).

I assume Android uses Unicode file names, so all I had to do is to force the zipper program to encode the file names in Unicode too. It this possible?

Thanks for your time, RG

Rene
  • 3,746
  • 9
  • 29
  • 33

2 Answers2

1

I found a workaround.

public static String iso (String s)
    {
        byte bytes[]=EncodingUtils.getBytes(s,"iso-8859-1");
        return new String(bytes);
    }

This function converts an ISO-8859-1 string to a Java Unicode string. You can apply this to the filename you get from ZipEntry.getName(), and use it to store the file. This lets you generate zips and read the content on the Android.

For transportation from Windows (say) to Android, you can zip and unzip using jar from the JDK. Info-zip does not work!

That's a pretty cool solution, but it will fail for Japanese characters or anything out of ISO-8859-1. Don't know what to do then.

Rene
  • 3,746
  • 9
  • 29
  • 33
0

Last time I checked the SD card has a charset of ISO8859-1 and no UTF-8. This is for compatibility with Windows systems. I guess you will have to convert the Java Unicode strings to ISO8859-1. But I think the best solution would be to use ASCII-Characters only.

mreichelt
  • 12,359
  • 6
  • 56
  • 70
  • That's of course not a solution, but a workaround. Indeed, I found in the Android docs that filenames in Java are converted to utf-8 byte sequences before being sent to the system, and vice versa. This allows consistent use of unicode filenames within Android. The only question remains is, how to transfer these from and to windows using a zip file. – Rene Nov 03 '10 at 12:15
  • I searched for this a minute ago, and I found the bugentry - the problem is known and also pretty old: http://bugs.sun.com/view_bug.do?bug_id=4244499 I guess you will have one of these options: 1) Wait for the bug to be fixed and for Android to catch up 2) Use your own ZIP implementation that does not use UTF-8 3) Stick with ASCII characters for now 4) Recommend a better operating system to your users ;-) – mreichelt Nov 03 '10 at 14:36