Introduction

The e2compr package consists of a set of patches and utilities written primarily by Antoine Dumesnil de Maricourt (dumesnil@etca.fr) to provide transparent compression and decompression in the second extended (ext2) filesystem.

Reasons you might wish to use e2compr:

• It reduces hard disk usage. (Robert Budzynski: ``Everyone knows that you never really have enough disk space, no matter how much you do have.'') Whether you're trying to squeeze everything you can onto that 386 with its 50MB hard drive, or you like to keep lots of source code / documentation / books on your multi-gigabyte hard drive, e2compr is just the thing.
• You can keep things in plain form that you otherwise would have gzipped. For example, Debian keeps all its documentation in gzipped form, which makes it harder to deal with. Postscript files and source trees especially come to mind.
• Unlike Zlibc, new files can be compressed automatically without your explicit action each time.
• Much safer than DouBle (if the filesystem becomes full or the system crashes). Also, with e2compr you can compress different parts of your filesystem in different ways; in particular, you can choose to leave some areas uncompressed.
• Unlike tcx, your data files can be compressed as well as your binaries.
• Although still something of a ``work in progress'' and not widely released, it is usable now (for home systems) and is relied on by many people.

Reasons you may wish not to use e2compr yet:

• Bugs are still being discovered, and there are numerous `fixme' comments in the code.
• Reduced speed and responsiveness.
• Backward compatibility with older e2compr releases is not yet considered very important. (Backward compatibility is a yoke that results in kluges and reduced end product quality.) In future, command-line options may change, behaviour may change, and even the file format.

Antoine passed maintainership to Peter Moulder (reiter@netspace.net.au) when Peter requested permission to distribute an e2compr patch updated for the then-new Linux-2.0.0.

Since then, Peter, Jan Rêkorajski, Gábor J.Tóth (jtoth@iki.kfki.hu) and Tom Lees (tom@lpsg.demon.co.uk) and others have developed and maintained e2compr.

In February 2000, Riley Williams rhw@MemAlpha.CX became joint e2compr maintainer with Peter. The pjm@e2compr.memalpha.cx address at the bottom of each e2compr web page forwards to both Riley and Peter.

Antoine no longer has the time to maintain e2compr; communication should probably be sent to e2compr maintainers or to the e2compr mailing list. (See below for how to subscribe to the lists).

Current state of project

The system is already usable now, and has been since at least 1997. Many people already rely on it.

Nevertheless, there are many avenues for improvement.

Error detection and handling could be improved, which is an issue for those with unreliable hard drive I/O.

See the TODO file in the distribution for a longer list of possibilities for further development (i.e. `wishlist').

I wouldn't use e2compr on a ``production'' machine (where neither faults nor delays are tolerated).