Sat Oct  9 15:31:18 EST 2004
	Added translation to Dutch.  Releasing 0.5.0.

Tue Oct  5 02:05:02 EST 2004
	Implemented logreplay.  Fixed some bugs in swaping code,
	added some more checks.

Mon Oct  4 02:22:22 EST 2004
	Added support for searching multiple databases 
	(like with locate).  Added support for pruning
	paths (option -P).  Updated man-pages and other docs.

Sun Sep 26 17:14:13 EST 2004
	Added support for 64-bit files (database larger than
	2^31 bytes).  Made offsets to other nodes relative
	(reduces DB size since relative offsets are almost
	always significantly smaller).  This is an incompatible
	database format change!  Added option to allow user
	to change the memory limit (8 MB is a bit small for
	very large databases if plenty of memory is available).

Sun Sep 26 16:04:10 EST 2004
	Fixing potential segfault when reading corrupted 
	database.

Tue Sep 14 22:29:53 EST 2004
	Bounding search-string size to avoid quadratic
	complexity for very long meta-data strings.
	
Thu Sep  9 00:29:37 EST 2004
	Fixed bugs, fixed I18N to make it work properly if
	libdoodle is used in applications other than doodle.
	Releasing 0.4.0.

Sat Sep  4 18:38:06 IST 2004
	Changed API (needed to make memory consumption bound
	truely work).  New API should also be more intuitive
	and results in less code.  The Java API did not change.  
	Improved IO by making reads aligned to blocks.  
	Improved copying on expansion to avoid quadratic 
	complexity (in the number of search-strings).  Added 
	benchmark for profiling.

Fri Sep  3 00:50:00 IST 2004
	Added code to reduce memory consumption (by using
	a temporary file and serializing part of the index
	to that file if we cross a certain memory threshold).
	Also reduced memory consumption on 64-bit machines
	by replacing pointers with 32-bit integers.  This also
	improves CPU usage by avoiding sorting and binary
	searches.  Reduced IO (mostly for search) by keeping 
	certain additional values in memory (feasible now 
	thanks to the memory threshold code).

Sat Aug 21 22:52:22 IST 2004
	Added Basque translation (incomplete).

Fri Aug 13 21:00:06 IST 2004
	JNI/Java interface to libdoodle works.  Improved
	error messages and provided complete translation to
	German.  Additional testcases.  Updated and improved
	man-pages.  Releasing 0.3.0.

Thu Aug 12 20:01:59 IST 2004
	gettextization of the package (with gettext 0.14)

Thu Aug 12 19:47:57 IST 2004
	Fixing problem with approximate searches (-i and -a options).
	
Fri Jul 23 20:05:20 IST 2004
	Added -B option.  Added man-page for libdoodle.
	Releasing 0.2.1.

Fri Jul 23 19:11:40 IST 2004
	Split doodle into library (libdoodle) and main binary.
	Logging now done via logger method.  Lots of code
	refactoring and renaming to make library split nicer.

Fri Jul  2 19:51:15 IST 2004
	Removed SHA-1 (now using timestamps instead) resulting
	in slighly smaller databases and significantly increased
	indexing performance.  Database format changed.
	Releasing 0.2.0.

Fri Jul  2 18:48:20 IST 2004
	Added code to avoid printing filenames multiple times
	if multiple keywords match.  Made option -e work when
	building database.  Added option -f for (full) filenames
	as keywords.  Added testcases for io and tree module.

Sun Jun 20 14:34:28 IST 2004
	Fixed segfault.  Added sorting of filenames array
	and cix to speed up database serialization (improves
	speed by a factor of 60).  Checking for special devices
	(char, blk, fifo, sock) added.  Releasing 0.1.1.
	
Fri Jun 11 19:27:11 IST 2004
	Reduced IO overhead (using buffers).  Reduced space
	consumption even further (integer-pair packing, do not
	canonicalize tree more than required, limit split to
	minimum split required).  Fixed various bugs.  Releasing 0.1.0.

Sun Jun  6 19:29:37 IST 2004
	Reduced space consumption (to 57%), added incrementality.
	
Mon May 31 15:36:33 EST 2004
	Re-worked suffix-tree code.  Multiple results and serialization
	now work fine.
	
Mon May 24 19:56:01 EST 2004
	Code prints first result (but only first hit at the moment).

Mon May 24 01:00:51 EST 2004
	Project started.


