1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
|
Thu Jan 14 11:53:06 CET 2010
Releasing doodle 0.7.0.
Fri Jan 1 21:07:59 CET 2010
Updating implementation for GNU libextractor 0.6.x API.
Mon Apr 28 08:41:16 MDT 2008
Updated Dutch translation.
Sun Apr 20 20:52:16 MDT 2008
Fixed a problem in the code handling iconv (character
set translation) errors; if iconv was not working, the
input string should just be returned without conversion
(thanks to Alan Post for pointing this out).
Wed Dec 19 11:06:08 MST 2007
Updated Vietnamese translation.
Mon Dec 10 17:10:08 MST 2007
Updated Swedish translation.
Wed Dec 5 18:26:24 MST 2007
Added "missing" include file for pthread_kill.
Releasing doodle 0.6.7.
Tue Nov 6 12:01:10 MST 2007
Updated Dutch translation.
Mon Feb 12 13:44:03 MST 2007
gettext update to 0.16.1.
Tue Jan 2 21:25:13 MST 2007
Patching bug on big-endian platforms (thanks to Volker Weiss).
This changes the database format (for everybody), so you will
need to regenerate your indices.
Patching bug in handling of empty metadata (again, thanks to
Volker Weiss).
Releasing doodle 0.6.6.
Mon Jan 1 19:10:40 MST 2007
Added pkgconfig support.
Tue May 2 14:56:42 PDT 2006
doodled shutdown sequence was broken, fixed.
doodled used stat instead of lstat, resulting in
symbolic links being added and then removed from
the database; fixed. doodled no longer verifies
all entries in the database on startup (too costly).
We assume that "doodle -b" has been used to create
an up-to-date database and that doodled will always
run in the background keeping it (sufficiently)
up-to-date.
Added out-of-process execution of GNU libextractor.
This ensures that doodle(d) does not crash even if
GNU libextractor does (i.e. due to bugs or corrupted
file systems).
Releasing doodle 0.6.5.
Sat Jan 7 02:37:53 PST 2006
Added Swedish translation.
Sun Sep 18 22:14:17 PDT 2005
Actually implemented keyword logging. Releasing doodle 0.6.4.
Sun Sep 18 16:30:56 PDT 2005
Updated Irish translation.
Fri Sep 16 19:06:42 PDT 2005
Updated German translation.
Thu Sep 15 21:55:19 PDT 2005
Fixing compiler warnings given by gcc 4.0.
Sun Sep 4 17:56:10 PDT 2005
Fixed -P option (short form did not work, Mantis #927).
Made prune-paths much faster by not iterating over the pruned
directories (Mantis #926).
Wed Aug 3 11:01:18 PDT 2005
Updated Vietnamese translation.
Thu Jul 28 16:35:50 EST 2005
Small changes to support API of libextractor >= 0.5.0.
Releasing doodle 0.6.3.
Tue Apr 5 17:11:33 EST 2005
Added Vietnamese translation.
Mon Feb 7 11:06:05 EST 2005
Added /sys to default prunepath.
Wed Jan 26 03:20:11 EST 2005
Added Italian translation.
Wed Jan 5 17:34:06 EST 2005
Fixed bug with approximate searches trying to write to the database.
Releasing doodle 0.6.2. (SVN revision is 52).
Fri Dec 31 20:33:44 EST 2004
Updated Irish and German translation.
Mon Dec 27 21:21:35 CET 2004
Do not follow symbolic links (in doodled, doodle was ok already).
Sun Dec 26 01:44:00 CET 2004
doodled debugging. Added multithreading, code cleanup. Works much
better now. Added conversion of search strings to UTF-8 to support
libextractor 0.4.0 and higher. Releasing doodle 0.6.1.
Sat Nov 13 15:00:03 EST 2004
Added locking (flock) to avoid concurrent modification of the doodle
DB. More doodled bugfixes and optimizations. In particular fine-grained
modification tracking (per node) and batched truncation. This
dramatically speeds up updating an existing database after files were
removed (O(n) instead of O(n*k) where n is the size of the DB and k
is the number of removed files).
Thu Nov 11 21:50:01 EST 2004
Various major bugfixes to doodled and also in doodle
(with respect to logging).
Fri Oct 22 22:04:32 EST 2004
Added doodled, a daemon that watches changes in directories
that were indexed by doodle and updates the database whenever
files in those directories change. Highly experimental code.
Uses libfam. Releasing doodle 0.6.0.
Thu Oct 21 21:55:31 EST 2004
Major bugfixes in the new code. The format was again
changed slightly. Where in 0.5.0 all offsets were relative,
this is no longer the case for the special absolute offset "0"
which is used for termination. Since "0" is not a valid relative
offset (a node may not point to itself) we do not have to encode
0 as a relative offset and can just put an absolute 0. This
safes a couple of bytes. Also we did not check if the same
node was listed (by chance) multiple times for the same filename.
Again, doing that check costs hardly any time but improves
the database size. For a small testcase we're now down
20% in DB size from 0.5.0 (270458 bytes vs. 223311 bytes).
Wed Oct 20 21:33:34 EST 2004
Added multi-link support. Multi-link support allows
much faster scanning of the doodle tree by allocating
n-nodes from the same level of the tree in the same
block in memory. This changes the database format
slightly; the cost and benefits in terms of size should
roughly cancel each other out (on some test I saw an
increase of 1.9%). For searching this should not make
much of a difference; single search invocations
should be minimally slower, on the other hand multiple
searches on the same tree in memory should be faster.
In general, building the database should be significantly
faster (at least as far as doodle is concerned) since in
essence this entails doing lots of searches in-memory.
Using the proftree benchmark the time for 100.000
in-memory 2-searches goes down from 150s to 110s on my
machine.
Did I mention that this feature is highly experimental and
the current CVS code is thus far from stable (for example
proftree2 still crashes!) If you want something stable,
stick to doodle 0.5.0 for now!
Sat Oct 16 15:32:48 EST 2004
Added translation to Albanian.
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.
|
