This version fixes a bug reported by Brian, who gets my thanks.

This bug blocked the recursive save of the root folder to one level (just the root folder), but did not happen when saving any other folder.

Since I was once more in the sources, I brought some changes to the automatic save of the playlist mode, which is triggered after each drag-n-drop operation when that option is checked. From now on, this automatic save will no longer care if the subfolders save option is active. In effect, it makes the playlist mode much faster, since it only saves what has just changed, not the whole directory tree from where we made the change.
This was what made the playlist mode so slow when I just moved one album before another on the 2GB SD Card of my Sony-Ericsson K750i.

Well, it's not really a *new* program, since I've started coding this at least two years ago, and I already spread it somewhat at my current work and at one of its client.

This program enables you to read IBM iSeries (formerly AS/400) save files on Windows, linux, or whatever you're working on provided it has a fairly recent Java Runtime Environment (v1.5 and above). It might be useful if you need to check what's inside a SAVF just before mailing it to a client, or if you want to extract a file from a SAVF without having to upload the whole SAVF on an iSeries.

Since this version is fairly stable, I'm coding a new version which will improve the performance (multithread, lazy analysis, ...) and reduce the memory usage (unpacking on disk, ...). With some luck I will at last have an occasion to port my prototypes for *MEDIUM and *HIGH unpacking from REXX to Java to manage these compression methods.

I just finished rewriting the site in a cleaner PHP, with some valid HTML, some layout changes, and a real host (OVH). I still have some things to change to ease the navigation, and hope you'll apreciate the changes.

After a long pause, I'm back with a patch for DriveSort, following a bug report from Jan-Eric (thanks :]). This bug triggered a rampant memory allocation in some cases on disks bigger than 4 GB. The bug is now fixed, and I used this opportunity to add a bit more progress report when opening the directory table, since on big volumes, it stood stuck for a while at 80%.