I've found that filenames with non-ASCII characters can be unpacked successfully in the /tmp directory (rather than the FAT filesystems where the ebooks are stored). So, I added a configuration option to enable/disable this behaviour. You have to open a non-problematic archive file to access the configuration options (otherwise the program kicks you out without any chance of changing options), then set the option to use /tmp for unpacking to "On". You should be able to open the problematic archives after that.
The disadvantage of using /tmp is that it is limited to 128 MB, so really big archives will probably fail when space runs out on that device.
Hopefully, this problem will go away when we change the program to unpack the images one at a time via internal library code rather than all at once via external commands. But that could take a while before it gets done.
I've updated the attached zip files in the first message in this thread with this new version. You only have to copy the new pbimageviewer.app program over; unzip and unrar haven't changed.
Last edited by rkomar; 01-24-2012 at 05:28 AM.