        release                 : 1.5.9 2012-01-22
                                : Solves a buf with snappy compression, 
                                : which could lead to crashes.
                                : Solves a bug in the fsync code which could lead to
                                : incomplete writes at unmount.
                                : Fixes BACKGROUND_DELETE not being parsed right as
                                : well as a deadlock when BACKGROUND_DELETE is disabled. 
        release                 : 1.5.8 2011-09-29
                                : Code clean-ups.
        release                 : 1.5.7 2011-09-17
                                : Fix a possible deadlock (timing) introduced in 1.5.6
                                : Adds support for google SNAPPY compression.
        release                 : 1.5.6 2011-09-16
                                : Added support to force replog rotation 
                                : echo 1 >/.lessfs/replication/rotate_replog
                                : Solved a bug that caused the replog not to rotate after
                                : REPLOG_DELAY had expired.
                                : lessfs_stats now shows the compression/deduplication ratio.
                                : Bug fix that increases the IOPS performance of lessfs from
                                : approx 1000 IOPS to > 6000 IOPS. (Measured with BDB on an
                                : Intel X25M SSD).
        release                 : 1.5.4 2011-09-12
                                : Important file_io reclaim bug fixes.
        release                 : 1.5.3 2011-09-06
                                : More replication fixes that insure data integrity 
                                : after an unclean shutdown of the master or the slave.
        release                 : 1.5.2 2011-09-05
                                : Fixes with replication after an unclean shutdown of the master.
                                : In this case the master would revert the running transactions but
                                : failed to remove them from the replication logfile.
                                : Removes flock when used with chunk_io
        release                 : 1.5.0 2011-08-04
                                : Lessfs now supports database partitioning (BerkeleyDB only)
                                : and a new data backend called chunk_io. Use chunk_io only with
                                : btrfs. An advantage of chunk_io is that deleting data will
                                : actually free up space. However the overall performance is lower
                                : and very depended on the type of filesystem that is used.
                                : btrfs works well, others are significantly slower.
        release                 : 1.4.10 2011-06-26
                                : Solves a problem with mklessfs when used with tokyocabinet
                                : and tc as datastore. Using TC as datastore is now DEPRECATED
                                : BerkeleyDB now supports using DB_PRIVATE. 
                                : Instead of using db_stat use telnet localhost bdb_status
        release                 : 1.4.9 2011-06-22
                                : Lessfs can now log to a file as well as syslog.
                                : File logging can be enabled with configure --enable-filelog
                                : after which lessfs will log to : /var/log/lessfs.log
        release                 : 1.4.8 2011-06-06
                                : Solves a problem that would cause lessfschk to segfault.
                                : Fixes a problem with rename and inode caching.
                                : Adds backward compatability with <= lessfs-1.4.2
        release                 : 1.4.7 2011-05-31
                                : Solves a problem with the new block sort routines that
                                : can cause a segfault. Also fixes a problem where hardlinking
                                : a hardlink would sometimes return a not null terminated string.
        release                 : 1.4.6 2011-05-29
                                : Solves a problem with sorting the blocks.
        release                 : 1.4.5 2011-05-27
                                : The worker threads now sort the work that needs to be done
                                : one inode-blocknr so that data will be written sequential
                                : on disk. -ENOSPC handling. Lessfs now no longer freezes
                                : when the disk fills up. Instead it returns -ENOSPC.
                                : When the disk is filled to MIN_SPACE-FREE - 3% lessfs 
                                : will now reclaim space much more agressive.
                                : Lessfs now uses HASH databases instead of BTREE databases
                                : when compiled with BerkeleyDB.
        release                 : 1.4.2 2011-05-09
                                : since BerkeleyDB 4.3
                                : Berkeley DB supports degree 2 isolation, which discards 
                                : read locks held by a transaction in order to minimize 
                                : the number of locks required by, for example, a cursor 
                                : iterating through a database. Lessfs now uses this isolation
                                : level and as a result use less locks and is 10% faster then 
                                : lessfs-1.4.0 with BDB. A nasty bug with hamsterdb was solved.
        release                 : 1.4.0 2011-05-05
                                : This is a new major release of lessfs.
                                : It's focus has been on reliability and crash recovery.
                                : Lessfs now comes with support for Berkeleydb which is the
                                : only backend that survived all our tests. Also changed is
                                : the way that deleted chunks are recycled.
                                : Other changes : The lessfs replication sequence can be read
                                : from .lessfs/replication/sequence. This file is only used
                                : with batch replication.
        release                 : 1.3.3.12 2011-04-28
                                : Added support for hamsterdb in listdb.
                                : When used with batch replication the master will now 
                                : close the replication log after reaching the configured size
                                : or after 15 minutes, whatever comes first.
                                : The interval can be adjusted at compile time in lib_repl.h
                                : #define REPLOG_DELAY 15*60
        release                 : 1.3.3.11 2011-04-17
                                : Lessfs now supports 'batch' replication. The master will
                                : create replication logfiles with a sequence number. The slave
                                : will scan a 'watchdir' and process the replication logfiles.
        release                 : 1.3.3.9  2011-04-06
                                : A new tool that allows commiting replication logfiles to
                                : a slave. This enables the use of for example sftp to transfer
                                : replication logfiles. Replication has been improved.
        release                 : 1.3.3.8  2011-03-23
                                : lessfs_read has been optimized.
        release                 : 1.3.3.7  2011-03-19
                                : Solves a problem where entering 0.0.0.0 as replication
                                : listen IP failed to work. Solves a problem that would
                                : cause fsx-linux to fail on lessfs running on i386.
        release                 : 1.3.3.6  2011-03-18
                                : Minor code cleanups. Lessfs now compiles with -Werror.
        release                 : 1.3.3.5  2011-03-15
                                : Minor code cleanups.
        release                 : 1.3.3.4  2011-03-08
                                : Replication connect fix.
        release                 : 1.3.3.3  2011-03-07
                                : Major rewrite of the replication code.
        release                 : 1.3.3.1  2011-02-02
                                : Fixes cache corruption bug in lessfs with hamsterdb.
        release                 : 1.3.3    2011-02-02
                                : Disable mmap use with hamsterdb. 
                                : Improved replication. Many code cleanups.
        release                 : 1.3.2    2011-01-19
                                : Make hamsterdb work without transactions.
        release                 : 1.3.1    2011-01-15
                                : Fixes a memory leak in add2cache.
                                : Fixes a problem that would cause lessfs to segfault when 
                                : used with encryption. Adds support for hamsterdb.
        release                 : 1.2.2.6  2010-12-22
                                : Fixes a silly mistake that may cause a segfault upon mounting.
        release                 : 1.2.2.5  2010-12-22
                                : Fixes a problem with lessfs_stats.
        release                 : 1.2.2.4  2010-12-19
                                : Improved cache eviction.
                                : When a lot of small files are copied, exceeding cache size,
                                : this change improves the performance with 40%.
                                : time tar xjf /home/maru/Downloads/linux-2.6.34.tar.bz2
                                : now takes : 1m11.810s where it takes 2m9.838s with lessfs 1.2.2.3.
        release                 : 1.2.2.3  2010-12-19
                                : This release improves the speed of metadata operations 
                                : by 30~50%.
        release                 : 1.2.2.2  2010-12-12
                                : This release enables some sane cache settings by default.
                                : It also solves a problem with sending the replication backlog.
                                : The lib_tc code has been split in two files.
        release                 : 1.2.2    2010-12-3
                                : Solved some minor configuration problems with replication.
        release                 : 1.2.1    2010-11-31
                                : Added support for a commercial only backend.
        release                 : 1.2.0    2010-10-31
                                : Improves the speed of metadata operations with 8%.
                                : lessfsck now defaults to a faster mode of operation.
                                : It used to check the inodes as well as the metadata.
                                : Now it only checks the metadata by default.
                                : It is still possible to check the file integrity
                                : by selecting the -t flag with lessfschk.
        release                 : 1.1.9.10 2010-10-27
                                : Introduces a cache layer for metadata making this
                                : release up to 12 times faster when small files are copied.
        release                 : 1.1.9.8  2010-10-18
                                : Fixes a memory leak when used with encryption enabled.
                                : Improved write performance.
        release                 : 1.1.9.7  2010-10-17
                                : Fixes a bug where truncation would fail when encryption had been
                                : selected with the tc datastore.
        release                 : 1.1.9.6  2010-10-13
                                : This enables lessfs background truncation threads
                                : to resume after an umount or shutdown. Background truncation
                                : is now enabled by default. Because lessfs uses the freelist
                                : database to keep truncation state this database now needs
                                : to be configured in lessfs.cfg even when using the tc datastore.
        release                 : 1.1.9.1  2010-10-01
                                : Fixes a problem with replication.
                                : Prioritizes reads and writes above truncation.
        release                 : 1.1.9.1  2010-09-29
                                : Fixes some nasty bugs with replication. Improves the du, ls -als
                                : statistics as well as better .lessfs_stats when lessfs
                                : truncates a file. .lessfs_stats is now fairly accurate when
                                : used with file_io. A problem that is not solved is that when
                                : a file is stored twice on lessfs, the second time it will
                                : show that the file uses 0 bytes. This is true until we remove
                                : the first copy of the file. A that time the second copy of the
                                : file is the reason that the blocks are still on the filesystem.
                                : It is however very difficult if not impossible with the current
                                : implementation to link these blocks back to file2.
        release                 : 1.1.9    2010-09-25
                                : Lessfs now supports asynchronous replication as well as
                                : synchronous replication. Please read the README.replication
                                : Background delete now appears to be very stable. This may
                                : become the default at the next release.
        release                 : 1.1.8.1  2010-09-21
                                : Fixes a small bug that disabled lessfs_stats 
        release                 : 1.1.8    2010-09-20
                                : This release contains a number of small bug fixes as well
                                : as improved truncation code. Lessfs can now truncate
                                : files as a background process. Files that are being truncated
                                : are no longer read/write locked but are now only write locked.
                                : During truncation a file with the inodenumber of the truncated file
                                : as filename will be visible in /.lessfs/locks/ to indicate that
                                : the inode is still being processed.
        release                 : 1.1.6    2010-08-22
                                : This release adds a number of features to lessfs.
                                : A. Lessfs is now able to spawn a background thread for 
                                : delete operations. Truncate and delete operations will
                                : appear to be very fast when this feature is enabled.
                                : B. Lessfs now supports native synchronous replication.
                                : This should be considered alpha code, although it
                                : works for me ;-). For now the lessfs master will wait
                                : for ever on the slave. So if the slave dies or disconnects
                                : the master will try to reconnect and freeze until this 
                                : has succeeded. In the future it will be possible to
                                : have the master write a backlog after a certain amount of time
                                : or after manual intervention.
        release                 : 1.1.5    2010-07-21
                                : Improves crash resilience by inserting test transactions
                                : when lessfs is mounted. If a test transaction fails the 
                                : database is automatically optimized (repaired).
                                : This release also fixes a bug where lessfs could deadlock
                                : when used with samba.
        release                 : 1.1.4    2010-07-16
                                : Improves lessfschk.
        release                 : 1.1.3    2010-07-15
                                : This release fixes a problem where lessfs would leave
                                : orphaned data chunks in the system under high load.
                                : A number of problems with lessfschk has been solved.
        release                 : 1.1.2    2010-07-07
                                : This release fixes a bug discovered by Dave.
                                : Lessfs now survives this test script:
                                : #! /bin/bash
                                : for n in $(seq -w 1 10 ); do
                                :  mkdir /data1/data/fsx$n
                                :  cd /data1/data/fsx$n
                                :  fsx-linux all >/tmp/fsx$n &
                                :
                                :  mkdir /data1/data/bonnie$n
                                :  bonnie++ -u root -d /data1/data/bonnie$n >/tmp/bonnie$n 2>&1 &
                                :
                                :  mkdir /data1/data/iozone$n
                                :  cd /data1/data/iozone$n
                                :  iozone -a >/tmp/iozone$n &
                                : done
                                : Performance has also improved... 
        release                 : 1.1.1    2010-07-02
                                : This release fixes a race condition that will manifest
                                : itself when lessfs is under heavy IO.
        release                 : 1.1.0    2010-06-30
                                : Marc Christiansen has provided a patch that fixed a crash
                                : and enhances lessfs_stats output. Many problems have
                                : been fixed with the new cache code.
        release                 : 1.1.0-beta5    2010-06-20
                                : Fixes a bug where reads would mostly mis the cache.
                                : Read performance has now dramatically (300%) increased
                                : for chunks of data that are found in the cache.
        release                 : 1.1.0-beta4    2010-06-19
                                : Fixes a (rare) race condition with the file_io backend.
                                : Improved write performance when writing smaller then 
                                : BLKSIZE data chunks. General code cleaning.
        release                 : 1.1.0-beta3    2010-06-15
                                : Under some circumstances a newly written block of data
                                : with hash (A) could be overwritten before the previous
                                : write had finished. A new 'per hash' locking mechanisme 
                                : now makes sure that this never happens.
                                : See create_hash_note for details.
        release                 : 1.1.0-beta2    2010-06-13
                                : Fixed a deadlock. Lessfs now supports deadlock reporting 
                                : and debugging. ;-)
        release                 : 1.1.0-alpha1   2010-06-02
                                : This release changes lessfs internals in a major way.
                                : Lessfs-1.1.0 is _not_ compatible with previous lessfs versions.
                                : This new version provides a much improved cache layer and
                                : way better performance. Threading has been improved and lessfs
                                : is now capable of using many threads /CPU's.
        release                 : 1.0.8   2010-11-30
                                : This release enables lessfs to be mounted without the
                                : need to specify other options then the configuration file
                                : and the mountpoint. Please consult the manual for more
                                : details. Eric D. Garver contributed a patch that
                                : makes the build process less picky about missing
                                : GNU files like INSTALL and NEWS. A bug in lessfs_read
                                : has been found by extensive testing with fsx-linux. In
                                : cases where a sparse block of data would be followed by
                                : a normal block lessfs_read would return wrong data in 
                                : some cases. Added automatic migration support for older 
                                : lessfs versions.
        release                 : 1.0.7   2010-11-27
                                : This release fixes a problem where data copied
                                : from windows to lessfs (samba) would show the wrong
                                : nr of blocks. This would result in du reporting wrong
                                : numbers.
        release                 : 1.0.6   2010-11-07
                                : Fixes a segfault that may occur when lessfs is used 
                                : without transactions enabled. The segfault occurs when
                                : lessfs is unmounted after closing the databases. The 
                                : impact of the bug is therefore low.
        release                 : 1.0.5   2010-03-07
                                : Fixes a small problem with logging.
        release                 : 1.0.4   2010-03-04
                                : This release enables support for transactions/checkpointing.
                                : Lessfs now no longer needs fsck after a crash. Also new is the ability to
                                : run a program when disk space drops below a certain amount of space. This program
                                : can be used to free up space when the tokyocabinet datastore is used.
        release                 : 1.0.1   2010-01-24
                                : Fixes a rare race condition that can cause lessfs to segfault and crash.
        release                 : 1.0.0   2009-12-30
                                : Removed all the bugs. ;-)
        release                 : 0.9.6   2009-12-21
                                : Fix an erroneous free() that can crash lessfs upon startup
                                : when the tiger hash is selected. Changes mklessfs so that
                                : it supports automatic directory creation and database overwrites.
                                : mklessfs now has improved error reporting.
        release                 : 0.9.5   2009-12-17
                                : This version of lessfs has been scrutinized for memory leaks.
                                : One small issue with lessfs_utimes has also been solved.
        release                 : 0.9.4   2009-12-16
                                : Fixes two memory leaks that are related to hardlink operations.
                                : Solves a problem caused by not initializing st_[a/c/m]time.tv_nsec.
                                : Thanks to Wolfgang Zuleger for doing a great job on analyzing these bugs.
                                : Fixed a memory leak in file_io.c
        release                 : 0.9.3   2009-12-13
                                : Partial file truncation encryption caused data corruption.
        release                 : 0.9.2   2009-12-12
                                : This release fixed some problems where permissions where not properly
                                : set on open files. It also fixes a problem with the link count 
                                : of directories. Performance for some meta data operations has improved.
        release                 : 0.9.0   2009-12-09
                                : Problems with truncation have finally been resolved.
                                : Less now passes fsx-linux. rsync now works without problems.
                                : Major changes of the truncation code.
        release                 : 0.8.3   2009-11-15
                                : Fixes a major bug in the truncation code.
                                : This bug will crash lessfs when used with ccrypt or rsync –inplace.
        release                 : 0.8.2   2009-11-09
                                : Fixes a bug that causes lessfsck and mklessfs to segfault when compiled
                                : with encryption support and encryption disabled in the config.
                                : Fixes a bug that causes lessfs to segfault on umount when compiled
                                : with encryption support and encryption disabled in the config.
                                : lessfsck,listdb and mklessfs are now installed in /usr/sbin 
                                : instead of /usr/bin.
        release                 : 0.8.1   2009-11-07
				: Fixes a bug that causes mklessfs to segfault when DEBUG is not set.
				: Adds lessfsck. lessfsck can be used  to check, optimize and repair
				: a lessfs filesystem. lessfsck only supports the tc data store for now.
                                : Support for the file_io backend will come with release 0.8.2
        release                 : 0.8.0   2009-10-26
                                : Fixes a possible segfault when lessfs is used with lzo compression.
                                : Fixes a problem when compiling lessfs without encryption on
                                : a system without openssl-devel.
                                : Enhances the logging facility.
                                : Performance has improved on higher latency storage like drbd or iscsi.
                                : Reduces the number of fsync operations when sync_relax>0.
                                :
                                : Thanks to : Roland Kletzing for finding and assisting 
                                :             with solving some of the problems mentioned.
                                :
        release                 : 0.7.5   2009-10-16
                                : Fix a segfault on free after unmounting lessfs without
                                : encryption support. Fix a problem that could lead to a
                                : deadlock when using file_io with NFS.
                                : A performance improvement, changed a mutex lock for a
                                : spinlock.  
        release                 : 0.7.4   2009-10-12
                                : This version of lessfs introduces a new hash named
                                : Blue Midnight Whish.
                                :
                                : http://www.q2s.ntnu.no/sha3_nist_competition/start
                                :
                                : This is a very fast hash that increases lessfs performance
                                : significantly. The implementation makes it easy to use any
                                : of the hashes from the NIST hash competition. MBW was
                                : choosen for lessfs because of the speed.
                                : To use BMW : configure --with-sha3
        release                 : 0.7.3   2009-10-06
                                : Error on free.
        release                 : 0.7.2   2009-10-06
                                : Fixes a typo in lib_tc.c that could lead to data corruption.
                                : @@ -990,7 +991,7 @@
                                : if (NULL != cachedata) {
                                : blk = (BLKCACHE *) cachedata->data;
                                : binhash(blk->blockdata, BLKSIZE, res);
                                : -                stiger=(unsigned char *) &res;
                                : +                dtiger=(unsigned char *) &res;
                                : One character does make difference...
                                : Includes a small performance enhancement.
        release                 : 0.7.1   2009-10-05
                                : A complete new way to store data has been introduced.
                                : The file_io backend has some advantages over tokyocabinet 
                                : for storing the actual data blocks.
                                : The performance has been improved.
        release                 : 0.6.1   2009-09-06
                                : Never improve code minutes before releasing it.
                                : Fix a silly bug with mklessfs.
        release                 : 0.6.0   2009-09-06
                                : Added encryption support to lessfs. 
                                : Fixed one small bug that would leave orphaned meta data in the
                                : metadatabase when hardlinks where removed.
                                : configure now has the option to select lzo compression as
                                : well as an option for debugging. See ./configure --help
        release                 : 0.5.0   2009-08-26
                                : Improved thread locking that leads to much better performance.
                                : Many NFS related problems have been solved and debugging
                                : is now easier.

        release                 : 0.2.8   2009-08-17
                                : Many bugfixes, including incorrect filesize on writing
                                : in a file with various offsets using lseek. This also
                                : caused problems with NFS.

        release                 : 0.2.7   2009-08-14
                                : Fixed a problem where dbstat failed to return the proper
                                : filesize. One other bug could leak to a deadlock of lessfs.

        release                 : 0.2.6   2009-07-17
                                : Fixed two bugs, one which could lead to data corruption.
                                : One other that would leave deleted data in the database.

        release                 : 0.2.5   2009-07-05
                                : This release fixes to one minor and one major bug. 
                                : One bug in the code would actually crash lessfs 
                                : upon renaming a file or directory. lessfs-0.2.4
                                : is no longer available for download.
                                
        release                 : 0.2.4   2009-07-05
                                : This release add support for dynamic defragmentation.
                                : Tokyocabinet supports this as of release 1.4.27
                                : This feature is very important because in the long run
                                : it is crucial for a normal filesystem that blocks of data
                                : are really deleted upon request without defragmenting 
                                : the filesystem. The old defragmentation method is still 
                                : usefull and available. See the tokyocabinet documentation
                                : for details about defragmentation.

        release                 : 0.2.3   2009-06-24
                                : This release fixes a small memory leak and improves
                                : write performance in general approx 12%.
        Known issues            : Using direct_io with kernel 2.6.30 causes reads to
                                : continue for ever. I am not sure if this is a kernel
                                : issue or a lessfs bug. With earlier kernels direct_io
                                : works fine.


        release                 : 0.2.2   2009-06-21
                                : This release adds supports for NFS. It also fixes
                                : some potential problems with the internal cache code.
                                : NFS is only supported with kernel >= 2.6.30 earlier
                                : kernels will corrupt large files written with NFS!
                                : A lot of code has changed in this release so please
                                : do not use this version for production.

        release                 : 0.2.1   2009-06-10
        Resolved problems       : Improved the performance of writes smaller then
				  max_write in size. These writes will now remain long
                                  enough in the cache so that subsequent writes to the
                                  same block will update the cache instead of the database.
                                  Mounting lessfs without formatting the filesystem now
                                  logs a warning instead of creating a segfault.
                                  Creating of sparse files now works again after being
                                  broken in release 0.1.19.

        release                 : 0.2.0   2009-05-25
        Improved performance    : Added a cache that increases throughput with almost 30%.

        release                 : 0.1.22  2009-05-14
        Resolved problems       : One last problem with hardlinking symlinks.
                                  Lessfs will now stop all I/O before the disks run out of space
                                  and corrupt the databases.
        Known problems          : Multithreaded I/O in the same file sometimes fails
                                  under exceptional circumstances. (bonnie++).

        release                 : 0.1.21  2009-05-05
        Resolved problems       : Many problems related to hard en softlinks
        Known problems          : Multithreaded I/O in the same file sometimes fails 
				  under exceptional circumstances. (bonnie++).
				  If the disk runs out of space the tokyocabinet database
				  is corrupted.

	release			: 0.1.20  2009-04-22
	Resolved problems	: Fixed two bugs:
				  1. Truncate operations would sometimes fail.
				  2. unlink of hardlinked files would sometimes fail.
      	Known problems          :

	release			: 0.1.19  2009-04-08
	Resolved problems	: Fixed a bug in the truncate code that would leave
				: deleted chunks in the databases. 
        Known problems		: 

	release			: 0.1.18  2009-03-30
	Resolved problems	: Code cleanup and partial rewrite of update_block. 
        Known problems		:

	release			: 0.1.17  2009-03-28
	Resolved problems	: Reenable logging of statistics to syslog.
        Known problems		:

        release                 : 0.1.16  2009-03-27
        Resolved problems       : Added configure script and fixed ARCH check.
        Known problems          :

Initial release                 : 0.1.15  2009-03-23
        Resolved problems       :
        Known problems          :
