------------------------------------------------------------ revno: 356 tags: ArimaaCC_2011 committer: janzert branch nick: opfor timestamp: Mon 2011-02-21 16:51:05 -0500 message: Quite a few eval modifications, primarily to hostage/blockades and rabbit strength ------------------------------------------------------------ revno: 355 committer: janzert branch nick: opfor timestamp: Fri 2011-02-18 11:04:18 -0500 message: Make search extension time a minimum of 5 seconds or a quarter of the time left ------------------------------------------------------------ revno: 354 committer: janzert branch nick: opfor timestamp: Thu 2011-02-17 10:07:01 -0500 message: Keep searching if current move is a loss and there are more moves and time to search. Add goal defense search back into quiescence search, still not done right but the current method doesn't seem to hurt either. Reduce mobility eval weight. Increase hostage value. ------------------------------------------------------------ revno: 353 committer: janzert branch nick: opfor timestamp: Sun 2011-01-23 05:48:30 -0500 message: Reduce effect of distance pieces on rabbit strength ------------------------------------------------------------ revno: 352 committer: janzert branch nick: opfor timestamp: Fri 2011-01-21 21:00:10 -0500 message: Reduce second victim in one trap weight ------------------------------------------------------------ revno: 351 committer: janzert branch nick: opfor timestamp: Wed 2011-01-19 22:29:37 -0500 message: Fix a few hostage evaluation bugs ------------------------------------------------------------ revno: 350 committer: janzert branch nick: opfor timestamp: Tue 2011-01-18 19:35:30 -0500 message: Failed experiment with trying to encourage weak rabbits to the sides Add static eval to parseboard Add eval_eval to find positions the static eval may be badly misevaluating ------------------------------------------------------------ revno: 349 committer: janzert branch nick: opfor timestamp: Wed 2011-01-12 17:09:36 -0500 message: Don't search all steps for a goal defense in qsearch (in testing could find no benefit) ------------------------------------------------------------ revno: 348 committer: janzert branch nick: opfor timestamp: Wed 2011-01-05 23:41:39 -0500 message: Increase static mobility eval weight Get rid of really bad, scoring of an opponent's goal as a forced goal A few debug message tweaks ------------------------------------------------------------ revno: 347 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2010-07-26 16:10:21 -0400 message: Fix depth limited search when a timecontrol is also given ------------------------------------------------------------ revno: 346 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2010-04-24 17:50:49 -0400 message: log search time with tenths of a second ------------------------------------------------------------ revno: 345 tags: ArimaaCC_2010 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2010-02-27 10:58:11 -0500 message: Update list of libraries needed for a statically linked build ------------------------------------------------------------ revno: 344 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2010-02-27 10:49:41 -0500 message: Add readme and license ------------------------------------------------------------ revno: 343 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2010-02-27 10:33:50 -0500 message: Use full width search in the q-search when goal threat is present ------------------------------------------------------------ revno: 342 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2010-02-01 13:34:15 -0500 message: Clear weakframers before use ------------------------------------------------------------ revno: 341 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2010-01-31 06:25:37 -0500 message: Reduce weight of partial blockades Check root score during parallel search and restart if it changes ------------------------------------------------------------ revno: 340 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2010-01-28 09:08:33 -0500 message: Make on trap penalty relative to current FAME value Adjust blockade weights Add false blockade checking code in static mobility testing ------------------------------------------------------------ revno: 339 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2010-01-27 14:29:38 -0500 message: Check for rabbit frames in new mobility code. Check cat mobility. Count squares that leave the piece frozen towards mobility. Make frames and hostage values relative to FAME value. Include number of nearby supporting pieces around trap in hostage value. Reduce hostage value if the holder is frozen. Remove old blockade and hostage code. ------------------------------------------------------------ revno: 338 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2010-01-26 20:24:41 -0500 message: Add pushes after one free step to static mobility ------------------------------------------------------------ revno: 337 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2010-01-26 05:33:30 -0500 message: Extend static mobility to recognize more legal moves ------------------------------------------------------------ revno: 336 committer: Brian Haskin Jr branch nick: opfor timestamp: Sun 2010-01-24 13:39:11 -0500 message: More mobility unification work ------------------------------------------------------------ revno: 335 committer: Brian Haskin Jr branch nick: opfor timestamp: Fri 2010-01-22 15:42:13 -0500 message: preliminary unification of mobility eval ------------------------------------------------------------ revno: 334 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2010-01-21 10:29:16 -0500 message: First minor steps toward unification of various mobility related evaluation methods ------------------------------------------------------------ revno: 333 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2010-01-15 02:20:45 -0500 message: tweak mobility weights ------------------------------------------------------------ revno: 332 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2010-01-14 00:56:17 -0500 message: remove almost frozen from frozen piece eval, should be covered by threat eval now ------------------------------------------------------------ revno: 331 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2010-01-13 15:58:40 -0500 message: Mobility eval work, split threat and mobility weight ------------------------------------------------------------ revno: 330 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2010-01-12 03:45:43 -0500 message: Refactor blockades, hostages and frames Tweak general mobility ------------------------------------------------------------ revno: 329 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2010-01-07 12:04:55 -0500 message: Add random position tester for static mobility search Small fix for static mobility when supporting piece moves away from a trap ------------------------------------------------------------ revno: 328 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2010-01-07 11:59:18 -0500 message: Assert that search result is not provisional when search depth is finished ------------------------------------------------------------ revno: 327 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2010-01-03 09:06:57 -0500 message: Only use the result of a search at the next depth after we're sure it's better than the previous best move ------------------------------------------------------------ revno: 326 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-12-31 17:21:09 -0500 message: Check legality of steps when getting pv for display ------------------------------------------------------------ revno: 325 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-12-31 00:59:14 -0500 message: Catch and log sync exceptions while waiting in search threads ------------------------------------------------------------ revno: 324 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2009-12-30 01:12:23 -0500 message: Add better logging for search thread exception Crash the program if a search thread crashes ------------------------------------------------------------ revno: 323 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-12-24 00:14:12 -0500 message: Fix crash when trying to log a report if no positions have been generated ------------------------------------------------------------ revno: 322 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2009-12-23 23:09:38 -0500 message: Fix bug causing random moves in forced wins, from confusion between best score found in search and best score by positions updated in main thread ------------------------------------------------------------ revno: 321 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-12-22 15:05:21 -0500 message: Fix crash bug in threaded engine from not clearing msg queue after search ------------------------------------------------------------ revno: 320 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-12-21 16:45:02 -0500 message: Fix error in queue clear function ------------------------------------------------------------ revno: 319 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-12-19 10:28:52 -0500 message: Forgot to initialize board setup generator in threaded engine ------------------------------------------------------------ revno: 318 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-12-18 20:15:24 -0500 message: Make sequential or threaded engine selectable from the commandline ------------------------------------------------------------ revno: 317 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-12-18 20:13:57 -0500 message: Remove some debug messages Rearrange code a bit to avoid a compiler bug ------------------------------------------------------------ revno: 316 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-12-15 10:34:22 -0500 message: Initial implementation of root parallelization Split some of the goal search and trap move functions to improve compile time ------------------------------------------------------------ revno: 315 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-11-29 22:47:19 -0500 message: Work around for compiler bug ------------------------------------------------------------ revno: 314 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-11-29 08:10:09 -0500 message: Various fixes for use with tango > 0.99.8 Expanded material population packing debug check ------------------------------------------------------------ revno: 313 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-11-27 13:28:57 -0500 message: Separate transposition table construction from the searcher ------------------------------------------------------------ revno: 312 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-11-24 14:13:19 -0500 message: Move thread safe queue to its own module out Better detection for hash step pushes that should really be pulls ------------------------------------------------------------ revno: 311 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-11-23 22:20:54 -0500 message: Fix a goal search bug ------------------------------------------------------------ revno: 310 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-11-23 14:59:49 -0500 message: Only search pull steps in qsearch when push and pull are available. Make position and steplist allocation thread safe. ------------------------------------------------------------ revno: 309 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-11-12 06:12:33 -0500 message: Remove old aborted search log message. Stop max search from going negative. ------------------------------------------------------------ revno: 308 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-11-12 00:51:07 -0500 message: Make logging transposition table stats configurable. Fix time management when using a time control with uncapped reserve time. ------------------------------------------------------------ revno: 307 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-11-06 02:14:09 -0500 message: More changes to work with tango library ------------------------------------------------------------ revno: 306 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-10-27 04:05:36 -0400 message: Remove old bot engines ------------------------------------------------------------ revno: 305 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-10-27 04:04:08 -0400 message: Remove commented out imports ------------------------------------------------------------ revno: 304 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-10-27 04:01:38 -0400 message: Don't try and link the winsock library unless on windows ------------------------------------------------------------ revno: 303 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-10-26 17:51:16 -0400 message: Initial pass at complete port to Tango library ------------------------------------------------------------ revno: 302 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-10-22 00:54:22 -0400 message: Change node count to count every call to search method. Add more transposition table logging. Undo transposition table node copying in search, current usage should be fine without it. ------------------------------------------------------------ revno: 301 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-09-05 01:15:45 -0400 message: Remove old now incorrect comment ------------------------------------------------------------ revno: 300 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-09-05 01:04:56 -0400 message: Rework logged eval to not use an aei command ------------------------------------------------------------ revno: 299 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-08-30 01:08:16 -0400 message: Fix a couple really bad bugs in static mobility evaluation Update to conform with aei protocol version 1 ------------------------------------------------------------ revno: 298 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-08-24 16:33:35 -0400 message: Correctly parse long positions using g and w for side to move ------------------------------------------------------------ revno: 297 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2009-03-04 23:10:43 -0500 message: Fix a bug in each of goalsearch and capture generation ------------------------------------------------------------ revno: 296 tags: ArimaaCC_2009 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-28 08:53:46 -0500 message: Add some more 6 step captures to the capture generator ------------------------------------------------------------ revno: 295 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-28 06:31:32 -0500 message: reduce the weight of an elephant blockade ------------------------------------------------------------ revno: 294 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-28 05:19:26 -0500 message: Add threat area term to eval ------------------------------------------------------------ revno: 293 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-28 02:16:35 -0500 message: More evaluation tweaks ------------------------------------------------------------ revno: 292 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-27 23:28:24 -0500 message: Some more minor evaluation adjustments ------------------------------------------------------------ revno: 291 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-27 16:56:27 -0500 message: Only prune unrelated steps when depth less than half of max depth or less than 5 ------------------------------------------------------------ revno: 290 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-27 07:47:11 -0500 message: Greatly reduce the chance of a four rabbit forward setup Further loosen the definition of related steps for unrelated step pruning Adjust some more static evaluation weights ------------------------------------------------------------ revno: 289 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-27 01:12:28 -0500 message: Adjust some static eval parameters for captures, threats and hostages ------------------------------------------------------------ revno: 288 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-02-26 15:50:21 -0500 message: Add capture evasion steps to queiscense search Another capture generation fix ------------------------------------------------------------ revno: 287 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2009-02-25 02:56:35 -0500 message: Get rid of unused option to capture generation to not generate all captures ------------------------------------------------------------ revno: 286 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2009-02-25 01:42:45 -0500 message: Add method to clear capture generation state ------------------------------------------------------------ revno: 285 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-02-24 13:14:14 -0500 message: Add early beta test pruning ------------------------------------------------------------ revno: 284 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-02-24 10:35:43 -0500 message: Another goal search fix and a capture generation fix ------------------------------------------------------------ revno: 283 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-02-23 14:59:56 -0500 message: More mobility optimization work Add penalty for strong blockaders ------------------------------------------------------------ revno: 282 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-02-23 09:41:19 -0500 message: A few more minor optimizations and small bug fixes to mobility evaluation ------------------------------------------------------------ revno: 281 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-02-22 12:18:27 -0500 message: Don't count moves that freeze the threat ------------------------------------------------------------ revno: 280 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-02-22 04:01:42 -0500 message: Fix some uninitialized variables in the evaluation A bit more optimization of mobility and piece threat ------------------------------------------------------------ revno: 279 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-21 17:11:03 -0500 message: Rework some of the threat eval ------------------------------------------------------------ revno: 278 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-21 01:27:12 -0500 message: Fix error in piece strength comparison in mobility evaluation Another capture generation fix ------------------------------------------------------------ revno: 277 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-20 19:56:47 -0500 message: Goal detection fix ------------------------------------------------------------ revno: 276 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-20 18:50:10 -0500 message: Add some threat analysis to mobility, almost certainly way out of balance right now More capture generation fixes ------------------------------------------------------------ revno: 275 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-02-19 13:09:43 -0500 message: Change mobility to apply evenly, not dependent on hostages Another capture generation fix ------------------------------------------------------------ revno: 274 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2009-02-18 02:54:49 -0500 message: Add consideration for hostages held to mobility ------------------------------------------------------------ revno: 273 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-02-17 22:25:19 -0500 message: Initial static eval of mobility of strongest pieces. Goal search fix ------------------------------------------------------------ revno: 272 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-02-17 13:04:20 -0500 message: Add a function to do a quick conservative mobility check on a piece Some more trap generation fixes. ------------------------------------------------------------ revno: 271 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-02-16 01:01:16 -0500 message: Fix backwards equality check on fixed depth search ------------------------------------------------------------ revno: 270 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-02-16 00:50:02 -0500 message: Cut goal threat weight by about a third Make sure that a fixed depth search does not use the timecontrol More trap generation fixes ------------------------------------------------------------ revno: 269 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-02-15 03:19:12 -0500 message: More goal search and trap generation fixes ------------------------------------------------------------ revno: 268 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-14 21:59:17 -0500 message: Whitespace cleanup of position.d ------------------------------------------------------------ revno: 267 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-14 21:06:29 -0500 message: Make regression_test.py a little more resilient More trap generation fixes Another goal search fix ------------------------------------------------------------ revno: 266 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-13 05:27:23 -0500 message: A few corner case fixes to the transposition table Another trap generation fix ------------------------------------------------------------ revno: 265 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-02-12 14:38:00 -0500 message: Add the ability to build a statically linked bot_opfor to build.py Give regression_test.py slightly nicer messages Another fix to capture generation ------------------------------------------------------------ revno: 264 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-02-12 03:58:10 -0500 message: Add a script to regression test goal search and trap generation with parseboard Some more goal search and trap generation fixes ------------------------------------------------------------ revno: 263 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-02-10 18:51:51 -0500 message: Reduce goal threat weight More capture generation fixes ------------------------------------------------------------ revno: 262 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2009-02-10 02:18:15 -0500 message: Static goal search fix Static capture generation fix ------------------------------------------------------------ revno: 261 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-02-09 14:55:15 -0500 message: Add goal extensions to search More capture generation fixes ------------------------------------------------------------ revno: 260 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-02-07 13:16:36 -0500 message: More capture gen fixes ------------------------------------------------------------ revno: 259 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-02-06 15:10:56 -0500 message: More static capture fixes Reduce weight of friendly captures in static eval to reduce "stand pat" value in qsearch ------------------------------------------------------------ revno: 258 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-02-05 21:26:26 -0500 message: Some fixes to capture generation testing Several fixes in capture generation ------------------------------------------------------------ revno: 257 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-02-05 02:07:11 -0500 message: Fix some major bugs in refactored capture generation code ------------------------------------------------------------ revno: 256 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2009-02-04 20:26:10 -0500 message: Refactor 2 opposing neighbors to trap capture generation Expand the trap checking in trap fuzzer to follow a capture through its steps Move trap checking to its own module and add it to parseboard program ------------------------------------------------------------ revno: 255 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-02-01 09:17:49 -0500 message: Add random static trap tester Various static trap fixes ------------------------------------------------------------ revno: 254 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-01-31 16:57:04 -0500 message: Properly wrap some debug code into a debug statement ------------------------------------------------------------ revno: 253 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-01-29 03:10:25 -0500 message: Yet another goal search fix ------------------------------------------------------------ revno: 252 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2009-01-26 17:46:27 -0500 message: Fix goal detection bug in search Fix static goal search bug ------------------------------------------------------------ revno: 251 committer: user branch nick: opfor timestamp: Sun 2009-01-25 22:12:26 -0500 message: Modify rabbit strength to better take into account trap safety ------------------------------------------------------------ revno: 250 committer: user branch nick: opfor timestamp: Fri 2009-01-23 20:57:53 -0500 message: Modify on trap penalty with the calculated trap safety Use seperate values for framed pieces ------------------------------------------------------------ revno: 249 committer: user branch nick: opfor timestamp: Thu 2009-01-22 22:15:46 -0500 message: Fix some bugs in calculating trap safety ------------------------------------------------------------ revno: 248 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-01-16 15:25:00 -0500 message: Add in some comments explaining what steps are allowed by the related step check Always allow all capture steps since the related step check still misses some 3 step combos that can be important for some captures ------------------------------------------------------------ revno: 247 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2009-01-11 03:46:33 -0500 message: Fix disallowing pulling an opponents rabbit to goal and back off again Fix an array overflow in killer move heuristic on searches over 24 steps deep ------------------------------------------------------------ revno: 246 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-01-03 16:51:16 -0500 message: Have goal threat eval take into account trap safety ------------------------------------------------------------ revno: 245 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-01-03 14:11:17 -0500 message: Small fix to goal threat eval ------------------------------------------------------------ revno: 244 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2009-01-03 00:39:49 -0500 message: Clean out some old code for an arimaa score eval ------------------------------------------------------------ revno: 243 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2009-01-02 18:10:50 -0500 message: Goal search fix ------------------------------------------------------------ revno: 242 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-01-01 22:35:58 -0500 message: Have the history heuristic correctly handle pass steps ------------------------------------------------------------ revno: 241 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2009-01-01 17:42:28 -0500 message: Cleanup and a few fixes in step sorting routine ------------------------------------------------------------ revno: 240 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2008-12-30 04:03:06 -0500 message: Null steps are never related ------------------------------------------------------------ revno: 239 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2008-12-30 00:28:47 -0500 message: Add holding piece moving away after stepping off a trap as a related step ------------------------------------------------------------ revno: 238 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2008-12-29 22:27:59 -0500 message: Fix a step sorter leak ------------------------------------------------------------ revno: 237 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2008-12-29 21:48:11 -0500 message: Add pruning of moves with only unrelated steps ------------------------------------------------------------ revno: 236 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-12-28 18:02:12 -0500 message: Goal search work ------------------------------------------------------------ revno: 235 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-12-21 02:22:32 -0500 message: Only do null move search if not 4 steps left ------------------------------------------------------------ revno: 234 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-12-21 02:21:05 -0500 message: Goal search ------------------------------------------------------------ revno: 233 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-12-19 11:51:41 -0500 message: More goal search work ------------------------------------------------------------ revno: 232 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2008-12-18 16:11:21 -0500 message: Add negascout (I really thought this was in a long time ago) Only do lmr when not in a pv node ------------------------------------------------------------ revno: 231 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2008-12-18 16:09:23 -0500 message: goal search bugs ------------------------------------------------------------ revno: 230 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2008-12-16 01:18:41 -0500 message: Add height parameter to alphabeta call to measure distance from root ------------------------------------------------------------ revno: 229 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-12-14 17:14:30 -0500 message: Rename some distance from root variable to a 'height' instead of doubling the meaning of 'depth' A few more goal search fixes ------------------------------------------------------------ revno: 228 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-12-12 00:46:36 -0500 message: Goal search fixing ------------------------------------------------------------ revno: 227 committer: janzert@janzert.com branch nick: opfor timestamp: Tue 2008-12-09 18:45:13 -0500 message: Change trap safety to look at strongest nearby piece A few more goal search fixes ------------------------------------------------------------ revno: 226 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-12-05 11:40:59 -0500 message: Fix trap safety test ------------------------------------------------------------ revno: 225 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-12-05 08:14:56 -0500 message: A few more edge cases, a few more bugs, in goal search ------------------------------------------------------------ revno: 224 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2008-12-03 13:54:47 -0500 message: Correctly do a re-search if a root move produce a value above beta ------------------------------------------------------------ revno: 223 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2008-12-03 08:20:52 -0500 message: Never ending new goal search corner cases and bugs ------------------------------------------------------------ revno: 222 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-11-30 23:18:38 -0500 message: More goal search fixing ------------------------------------------------------------ revno: 221 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2008-11-29 19:35:22 -0500 message: Add static evaluation cache ------------------------------------------------------------ revno: 220 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2008-11-29 19:34:55 -0500 message: Ever ongoing fixes to the new goal search ------------------------------------------------------------ revno: 219 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-11-28 02:10:53 -0500 message: Fix some more, add some more to the goal search ------------------------------------------------------------ revno: 218 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2008-11-26 13:35:08 -0500 message: Fix opfor to expect correct value for goal not found ------------------------------------------------------------ revno: 217 committer: janzert@janzert.com branch nick: opfor timestamp: Wed 2008-11-26 01:24:22 -0500 message: More bugs fixed in decision tree goal search Initial integration of new goal search into opfor ------------------------------------------------------------ revno: 216 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2008-11-24 16:57:13 -0500 message: More goal search work including refactoring the way results are recorded ------------------------------------------------------------ revno: 215 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2008-11-24 02:14:44 -0500 message: More goal search work ------------------------------------------------------------ revno: 214 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-11-23 01:58:07 -0500 message: More work on the new goalsearch Extended goal fuzzer to check each step of a goal move ------------------------------------------------------------ revno: 213 committer: janzert@janzert.com branch nick: opfor timestamp: Thu 2008-11-20 11:39:48 -0500 message: Many fixes to new goal search. Updates to goal_fuzzer to test the reversed position for black goals ------------------------------------------------------------ revno: 212 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-11-14 03:24:40 -0500 message: Differentiate result messages between old and new goal search in parseboard. ------------------------------------------------------------ revno: 211 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-11-14 03:13:15 -0500 message: Have parseboard run the new goal search as well. ------------------------------------------------------------ revno: 210 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-11-14 03:09:05 -0500 message: Set position to null after clearing to prevent double free. ------------------------------------------------------------ revno: 209 committer: janzert@janzert.com branch nick: opfor timestamp: Fri 2008-11-14 02:11:05 -0500 message: New decision tree static goal search, still quite buggy though. Small fuzzer for testing the new goal search, not yet working either as it is causing assertion errors in position code at the moment. ------------------------------------------------------------ revno: 208 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-11-09 01:05:35 -0500 message: Fix build script and set executable bit ------------------------------------------------------------ revno: 207 committer: janzert@janzert.com branch nick: opfor timestamp: Sat 2008-11-08 17:02:08 -0500 message: Simple build script ------------------------------------------------------------ revno: 206 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2008-11-03 16:36:54 -0500 message: Change setup weights to be more favorable to 99of9 and less toward Fritzlein's ------------------------------------------------------------ revno: 205 committer: janzert@janzert.com branch nick: opfor timestamp: Mon 2008-11-03 12:21:01 -0500 message: Reduce weight of some rabbit structure eval features ------------------------------------------------------------ revno: 204 committer: janzert@janzert.com branch nick: opfor timestamp: Sun 2008-11-02 22:26:16 -0500 message: Add penalty for high piece holding hostage or blockade ------------------------------------------------------------ revno: 203 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-10-03 18:35:15 -0400 message: Disable GC for live timecontrol games ------------------------------------------------------------ revno: 202 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-10-03 17:37:58 -0400 message: Add an option that can be used to force opfor to do an immediate garbage collection and report how long it takes. ------------------------------------------------------------ revno: 201 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-10-02 21:58:42 -0400 message: Fix GC NO_SCAN calls to use the correct memory address Add NO_SCAN flag for StepLists ------------------------------------------------------------ revno: 200 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-09-30 01:00:10 -0400 message: Get rid of kludge to stop pondering after the regular move time now that an incoming move is detected consitently fast. Reduce work of calling code for PositionNode freeing. Add GC NO_SCAN flags to the hash table and position class. ------------------------------------------------------------ revno: 199 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-09-28 00:13:28 -0400 message: Add some debug code and message to help catch memory leaks. Fix a leak of StepLists in the search ------------------------------------------------------------ revno: 198 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-09-22 22:07:05 -0400 message: Check for search aborts in the qsearch ------------------------------------------------------------ revno: 197 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-09-22 15:35:38 -0400 message: Check for aborted searches after null move search and qsearch ------------------------------------------------------------ revno: 196 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-09-21 23:54:02 -0400 message: Fix forced move reporting as a losing move Remove unneeded import ------------------------------------------------------------ revno: 195 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-09-16 02:00:43 -0400 message: Add option to use socket communication, make stdio communication the default ------------------------------------------------------------ revno: 194 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-09-16 00:05:35 -0400 message: Fix bug in stdio interface ------------------------------------------------------------ revno: 193 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-09-08 13:42:31 -0400 message: Make opfor more likely to actually exit after receiving the quit command ------------------------------------------------------------ revno: 192 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-09-05 20:40:33 -0400 message: Add a callback to check for any commands that indicate the current search should abort ------------------------------------------------------------ revno: 191 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-09-05 20:38:07 -0400 message: Add a missed file in last commit. ------------------------------------------------------------ revno: 190 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-09-05 14:45:20 -0400 message: Add ability for aei interface to communicate over stdio ------------------------------------------------------------ revno: 189 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-08-27 00:09:58 -0400 message: Remove the tournament rules option (these rules are used in all games now) ------------------------------------------------------------ revno: 188 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-08-27 00:07:13 -0400 message: Have random setups prefer 99of9 over Standard over Fritz ------------------------------------------------------------ revno: 187 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-08-25 00:51:52 -0400 message: Send moves that are predicted to lose directly to the losing moves list. Add a second pass when all moves are predicted to lose to double check and also find the longest loss. ------------------------------------------------------------ revno: 186 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-07-26 01:52:12 -0400 message: Add ability to specify the number of playouts from the command line to handicaptest ------------------------------------------------------------ revno: 185 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-07-25 15:30:12 -0400 message: Clean up interface (command line and output) for handicaptest ------------------------------------------------------------ revno: 184 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-07-24 10:34:52 -0400 message: Change another node count variable to be 64bit, would overflow and cause problems after about 20 hours of search otherwise. ------------------------------------------------------------ revno: 183 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-04-25 18:55:35 -0400 message: Some refactoring of board setup New small program that does random playouts from a specified handicap ------------------------------------------------------------ revno: 182 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-04-21 13:01:05 -0400 message: Add max target time as a possible abort search time ------------------------------------------------------------ revno: 181 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-04-21 12:59:52 -0400 message: Display number of wins from random playouts ------------------------------------------------------------ revno: 180 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-04-13 20:18:36 -0400 message: Fix expanded qsearch ------------------------------------------------------------ revno: 179 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-04-13 08:33:32 -0400 message: Default to random minor piece placement ------------------------------------------------------------ revno: 178 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-04-13 08:31:04 -0400 message: Check capture moves in qsearch that can't be completed in current move ------------------------------------------------------------ revno: 177 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-04-12 11:28:20 -0400 message: Revert inner loop of alpha-beta replacement had wrong negation and even after correcting that still gave different results ------------------------------------------------------------ revno: 176 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-04-12 09:25:40 -0400 message: Add null move pruning ------------------------------------------------------------ revno: 175 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-04-11 16:09:49 -0400 message: Rearrange inner alpha-beta loop to be arguably cleaner and slightly faster ------------------------------------------------------------ revno: 174 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-04-11 12:56:52 -0400 message: Add support for target search time different from time control time. Mainly useful for postal games. ------------------------------------------------------------ revno: 173 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-04-08 07:43:37 -0400 message: Add option to randomize cats and dogs in the setup ------------------------------------------------------------ revno: 172 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-04-04 04:14:29 -0400 message: Add multiple piece setups ------------------------------------------------------------ revno: 171 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-04-04 04:12:41 -0400 message: Add logging when setting transposition table size ------------------------------------------------------------ revno: 170 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-04-01 22:22:19 -0400 message: change a few of the new log info report names ------------------------------------------------------------ revno: 169 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-03-30 22:55:42 -0400 message: Several changes and clean ups to make parseboard nicer. ------------------------------------------------------------ revno: 168 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-03-30 22:19:00 -0400 message: Add more search information logging Add a few more levels of search reductions at the root for really bad moves during deep searches. ------------------------------------------------------------ revno: 167 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-03-29 14:34:53 -0400 message: Call debug level that checks position invariants 'verify_position' instead of '2' ------------------------------------------------------------ revno: 166 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-03-29 14:28:05 -0400 message: Remove the undo move stuff since it turned out to be slower ------------------------------------------------------------ revno: 165 [merge] committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-03-29 14:19:15 -0400 message: Merge work on undoing steps ------------------------------------------------------------ revno: 164 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-03-24 04:06:50 -0400 message: Fix indentation in Step class ------------------------------------------------------------ revno: 163 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-03-24 04:06:00 -0400 message: Get rid of extraneous set and copy methods from Step class ------------------------------------------------------------ revno: 162 tags: ArimaaCC_2008 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-03-01 00:12:45 -0500 message: Revert almost frozen penalty to previous level ------------------------------------------------------------ revno: 161 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-03-01 00:02:29 -0500 message: Reduce penalty for trapping two pieces in the same trap ------------------------------------------------------------ revno: 160 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 23:28:05 -0500 message: Reduce long range trap threat value ------------------------------------------------------------ revno: 159 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 23:03:08 -0500 message: Take trap safety into account for rabbit strength ------------------------------------------------------------ revno: 158 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 19:31:30 -0500 message: Small speed up for capture generator ------------------------------------------------------------ revno: 157 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 19:10:10 -0500 message: Only ponder for the permove time of opponents turn ------------------------------------------------------------ revno: 156 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 16:18:23 -0500 message: Increase almost frozen weight and slightly lower minimum thinking time ------------------------------------------------------------ revno: 155 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 15:54:16 -0500 message: Tweaks to rabbit evaluation weights ------------------------------------------------------------ revno: 154 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 13:05:46 -0500 message: Add detection of 'all push' 6 step traps. ------------------------------------------------------------ revno: 153 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-29 10:08:20 -0500 message: Rework reductions done at the root based on previous incremental search score ------------------------------------------------------------ revno: 152 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-28 15:12:44 -0500 message: Some work on static trap and goal threat evals. ------------------------------------------------------------ revno: 151 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-27 04:36:32 -0500 message: Fix capture generator case with enemy on trap, one adjoining enemy and powerful piece 2 safe steps away ------------------------------------------------------------ revno: 150 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-26 18:47:23 -0500 message: Only report number of nodes in quiescent search if actually doing a quiescent search ------------------------------------------------------------ revno: 149 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-26 14:11:27 -0500 message: Correctly reset the active traps at the beginning of evaluation ------------------------------------------------------------ revno: 148 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-26 13:12:26 -0500 message: Remove expanding quiescent depth ------------------------------------------------------------ revno: 147 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-26 13:10:15 -0500 message: Refactor trap safety eval, include check that two neighboring pieces aren't trappable Add ability for logged eval to check current game position Add report method to engine and search ------------------------------------------------------------ revno: 146 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-25 11:10:13 -0500 message: Split out static evaluation components into a seperate file and class from opfor ------------------------------------------------------------ revno: 145 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 23:27:32 -0500 message: Fix last checkin to actually work ------------------------------------------------------------ revno: 144 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 22:50:43 -0500 message: Don't ponder if I'm low on time for the next move ------------------------------------------------------------ revno: 143 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 21:44:57 -0500 message: Correctly set the alphabeta node number to check for abort time ------------------------------------------------------------ revno: 142 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 21:23:12 -0500 message: Fix stronger piece calculation for unfrozen pieces ------------------------------------------------------------ revno: 141 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 18:50:22 -0500 message: Add penalty for being next to a stronger enemy even if not frozen ------------------------------------------------------------ revno: 140 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 18:18:05 -0500 message: Modify frozen piece score by total number of pieces left on the board ------------------------------------------------------------ revno: 139 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 11:58:23 -0500 message: Change abort think time to 20 seconds ------------------------------------------------------------ revno: 138 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 10:26:05 -0500 message: Correct a couple of code comments ------------------------------------------------------------ revno: 137 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 07:49:23 -0500 message: Add distance from trap factor to hostage score ------------------------------------------------------------ revno: 136 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 07:11:36 -0500 message: Fix to static trap victim sorting ------------------------------------------------------------ revno: 135 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 05:37:45 -0500 message: Add frozen penalty to blockade that is actually a hostage ------------------------------------------------------------ revno: 134 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 04:20:49 -0500 message: More adjustments to rabbit strength ------------------------------------------------------------ revno: 133 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 03:19:23 -0500 message: Adjustments to rabbit strength ------------------------------------------------------------ revno: 132 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 01:26:47 -0500 message: Fix the goal length sorting then disable it and only record the shortest goal since that's all opfor uses at the moment anyway ------------------------------------------------------------ revno: 131 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-23 01:10:39 -0500 message: Increase ageing of the history heuristic to only carry over 10 percent ------------------------------------------------------------ revno: 130 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-22 15:52:29 -0500 message: Slightly increase the weight of forward strong rabbits ------------------------------------------------------------ revno: 129 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-21 16:57:07 -0500 message: Change abort long search message to standard log instead of warning ------------------------------------------------------------ revno: 128 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-21 15:47:39 -0500 message: Add log message when playing forced move ------------------------------------------------------------ revno: 127 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-21 15:04:36 -0500 message: Fix abort time during pondering ------------------------------------------------------------ revno: 126 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-21 06:49:45 -0500 message: Add clearer depth reporting ------------------------------------------------------------ revno: 125 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-21 06:08:05 -0500 message: Always reset socket send buffer size, seems to work better than just leaving it if it's larger. Stop reporting aborted searches after a certain number in one move Set maximum abort time to 30 seconds when pondering ------------------------------------------------------------ revno: 124 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-21 04:56:40 -0500 message: Add amount of data already sent to socket send exception Increase small socket send buffers Decrease weight of advanced strong rabbits ------------------------------------------------------------ revno: 123 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-20 23:50:01 -0500 message: Speed optimizations to goal search ------------------------------------------------------------ revno: 122 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-20 20:01:07 -0500 message: Small speed optimizations to mobility and rabbit strength evals ------------------------------------------------------------ revno: 121 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-20 16:05:55 -0500 message: Use alpha reached to update killer heuristics ------------------------------------------------------------ revno: 120 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-20 14:59:39 -0500 message: Add killer move heuristic ------------------------------------------------------------ revno: 119 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-18 04:35:08 -0500 message: Allow server ip and port to be specified on the command line Add more socket error and closing detection ------------------------------------------------------------ revno: 118 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-02-17 18:09:44 -0500 message: Change most direct console printing to go through the log system Add a log of the minimum and maximum search times set ------------------------------------------------------------ revno: 117 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-15 15:38:56 -0500 message: Fix infinite loop in pv reporting when finding repeated positions in the transposition table ------------------------------------------------------------ revno: 116 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-15 11:59:13 -0500 message: More work on piece strength, mobility and a few other pieces of eval ------------------------------------------------------------ revno: 115 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-14 03:16:15 -0500 message: Fixes and changes to eval after looking at the split out component values ------------------------------------------------------------ revno: 114 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-14 01:25:14 -0500 message: New piece strength eval method Add a command to show each of the eval components for a given position ------------------------------------------------------------ revno: 113 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-13 20:24:08 -0500 message: Rip the option to no "stand pat" back out of quiescent search as it was definitely worse Take into account piece strength for hostages ------------------------------------------------------------ revno: 112 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-13 05:28:44 -0500 message: Fix shutting off stand pat quiesce ------------------------------------------------------------ revno: 111 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-13 04:57:25 -0500 message: Stop evaluating in the middle of a push in quiescent search ------------------------------------------------------------ revno: 110 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-13 04:52:33 -0500 message: More eval work Add option to test not allowing pre-queisce eval as alpha ------------------------------------------------------------ revno: 109 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-12 22:48:29 -0500 message: More work on static trap eval ------------------------------------------------------------ revno: 108 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-12 20:51:06 -0500 message: Go back to nodes instead of time for search run lengths, much faster for shallow searches Small tweaks to eval ------------------------------------------------------------ revno: 107 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-12 19:49:36 -0500 message: Tweak static trap weight ------------------------------------------------------------ revno: 106 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-12 18:33:41 -0500 message: Tweak on trap/framed evaluation Cap the value of piece strength and reduce effect of more distant pieces Reduce goal threat weight as it just misses to many threats and leads to an unbalanced evaluation when both sides really have threats but it only sees one Do some more work on static trap evaluation ------------------------------------------------------------ revno: 105 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-12 05:03:11 -0500 message: Age history heuristic entries by a larger amount Adjust stronger piece hostages to be worth slight less and weaker pieces a little higher Fix a couple of bad bugs in static trap evaluation Adjust goal threat evaluation ------------------------------------------------------------ revno: 104 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-11 23:26:27 -0500 message: Add search reductions at the root for moves with few steps and moves the previously had low scores Change the maximum default quiescent search depth to 16 steps Tweak mobility eval ------------------------------------------------------------ revno: 103 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-11 18:53:47 -0500 message: More eval tweaking ------------------------------------------------------------ revno: 102 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-11 16:11:57 -0500 message: Change long search warnings ------------------------------------------------------------ revno: 101 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-11 16:00:57 -0500 message: Get rid of estimating how long the next search run will last as it varies too much Fix divide by zero error when calculating number of nodes to search before checking time ------------------------------------------------------------ revno: 100 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-11 15:37:52 -0500 message: Add time checks within alphabeta search to abort the search if going to long Add logging of illegal moves from the capture generator in regular builds ------------------------------------------------------------ revno: 99 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-02-10 03:40:27 -0500 message: Fix nasty bug in LMR ------------------------------------------------------------ revno: 98 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-02-10 01:46:44 -0500 message: Make mobility evaluation more accurate ------------------------------------------------------------ revno: 97 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-09 21:55:39 -0500 message: Make tournament rules setting configurable ------------------------------------------------------------ revno: 96 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-09 21:33:40 -0500 message: Remove mtdf use from bot_opfor Make lmr more conservative (no recursive reductions) Add a form of lmr reduction to the root of the search ------------------------------------------------------------ revno: 95 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-09 15:50:29 -0500 message: Give late move reductions a try ------------------------------------------------------------ revno: 94 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-09 01:54:15 -0500 message: Implement slightly flexible logging system ------------------------------------------------------------ revno: 93 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-08 16:46:18 -0500 message: Tweaking after reviewing some recent games ------------------------------------------------------------ revno: 92 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-08 14:54:32 -0500 message: Convert to unix line endings ------------------------------------------------------------ revno: 91 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-08 14:33:32 -0500 message: Fixup mobility eval ------------------------------------------------------------ revno: 90 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-08 05:25:43 -0500 message: Adjust rabbit strength balance ------------------------------------------------------------ revno: 89 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-08 04:51:55 -0500 message: Change hostage criteria ------------------------------------------------------------ revno: 88 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-08 04:43:29 -0500 message: Add rabbit_steps function to expand a bitboard with appropriate steps Change maximum quiescence depth to be proportional to regular search depth Finish blockade/hostage eval ------------------------------------------------------------ revno: 87 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 21:51:52 -0500 message: Don't re-search moves that lead to a loss Move immediately when only one move left under consideration. ------------------------------------------------------------ revno: 86 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 21:25:55 -0500 message: Small adjustment to time management to better handle deep searches ------------------------------------------------------------ revno: 85 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 20:36:59 -0500 message: More eval adjusting Handle bad step from tranposition table better ------------------------------------------------------------ revno: 84 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 18:45:16 -0500 message: Disable home rabbit eval ------------------------------------------------------------ revno: 83 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 18:43:48 -0500 message: Adjusted default eval weights ------------------------------------------------------------ revno: 82 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 18:29:33 -0500 message: More work on mobility and rabbit strength eval ------------------------------------------------------------ revno: 81 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 17:31:27 -0500 message: Smooth out weak rabbit eval ------------------------------------------------------------ revno: 80 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-02-07 11:16:56 -0500 message: Tweak trap safety/control eval ------------------------------------------------------------ revno: 79 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-06 14:29:37 -0500 message: Tweak framed piece eval ------------------------------------------------------------ revno: 78 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-02-06 02:09:21 -0500 message: Disable central elephant eval ------------------------------------------------------------ revno: 77 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-02-05 08:10:17 -0500 message: Decrease effect of goal threat on weak rabbit eval ------------------------------------------------------------ revno: 76 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-04 23:39:43 -0500 message: Lessen the effect of goal threat on strong rabbits ------------------------------------------------------------ revno: 75 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-04 21:21:22 -0500 message: Tweak strong rabbit eval ------------------------------------------------------------ revno: 74 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-02-04 16:16:36 -0500 message: Move goal threat eval to its own method ------------------------------------------------------------ revno: 73 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-02-03 02:07:31 -0500 message: Make static trap eval add to the score instead of replace it ------------------------------------------------------------ revno: 72 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-02 20:21:42 -0500 message: Make a true static trap weight instead of just a switch ------------------------------------------------------------ revno: 71 committer: janzert@janzert.com branch nick: D timestamp: Sat 2008-02-02 14:40:32 -0500 message: Tweak central elephant weights ------------------------------------------------------------ revno: 70 committer: janzert@janzert.com branch nick: D timestamp: Fri 2008-02-01 13:52:26 -0500 message: Disable home rabbit bonus it showed no benefit ------------------------------------------------------------ revno: 69 committer: Janzert branch nick: D timestamp: Wed 2008-01-30 18:29:15 -0500 message: Fix a few more bugs in capture move generator Add mtd(f) search option ------------------------------------------------------------ revno: 68 committer: Janzert branch nick: D timestamp: Tue 2008-01-29 23:52:59 -0500 message: Finish capture step sorting in main search Speed up piece and rabbit strength eval Fix some more bugs in capture step generator ------------------------------------------------------------ revno: 67 committer: Janzert branch nick: D timestamp: Tue 2008-01-29 01:52:05 -0500 message: Fix another bug in capture step generator ------------------------------------------------------------ revno: 66 committer: janzert@janzert.com branch nick: D timestamp: Mon 2008-01-28 21:28:02 -0500 message: More work on sorting capture moves first in the search. ------------------------------------------------------------ revno: 65 committer: Janzert branch nick: D timestamp: Sun 2008-01-27 15:35:40 -0500 message: Start of ordering capture steps first in the search. Not working ------------------------------------------------------------ revno: 64 committer: Janzert branch nick: D timestamp: Sat 2008-01-26 17:57:41 -0500 message: Split step sorting out into a seperate class ------------------------------------------------------------ revno: 63 committer: Janzert branch nick: D timestamp: Fri 2008-01-25 21:58:45 -0500 message: Add random setup option to bot_score Have the capture generator sort captures by piece value and length ------------------------------------------------------------ revno: 62 committer: Admin branch nick: D timestamp: Fri 2008-01-25 12:17:56 -0500 message: Update random move and random step bots to work again Add random setup creation function Change random step bot to use a random setup ------------------------------------------------------------ revno: 61 committer: Admin branch nick: D timestamp: Thu 2008-01-24 22:31:18 -0500 message: Re-enable history heuristic ------------------------------------------------------------ revno: 60 committer: Admin branch nick: D timestamp: Thu 2008-01-24 15:53:46 -0500 message: Fix some rabbit eval bugs ------------------------------------------------------------ revno: 59 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-01-23 13:37:56 -0500 message: Probe 4 hash table positions instead of 1 ------------------------------------------------------------ revno: 58 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-01-22 15:27:42 -0500 message: Adjust piece on trap and frame evaluation ------------------------------------------------------------ revno: 57 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-01-22 14:01:49 -0500 message: Fix goal depth on the board for frozen rabbits Add some more reporting to parseboard program Fix a few more bugs in capture generator Add piece and trap location reporting to capture generator Add ponder support to bot_opfor Add static capture evaluation term Add transposition table use to quiesense search Remove full quiescense search Adjust evaluation terms and weights ------------------------------------------------------------ revno: 56 committer: janzert@janzert.com branch nick: D timestamp: Sun 2008-01-06 17:58:07 -0500 message: In memory opening book. Better central elephant definition. ------------------------------------------------------------ revno: 55 committer: janzert@janzert.com branch nick: D timestamp: Thu 2008-01-03 09:40:12 -0500 message: Fix various bugs that lead to evaluation bias. ------------------------------------------------------------ revno: 54 committer: janzert@janzert.com branch nick: D timestamp: Wed 2008-01-02 14:17:16 -0500 message: Move position invariants to a special debug level to greatly speed up normal debug builds. Refactor FastFAME into a class like it should have been done Stop capture move generator from returning many duplicate captures Add static trap evaluation Add rabbit and piece strength evaluations Add a random evaluation weight ------------------------------------------------------------ revno: 53 committer: janzert@janzert.com branch nick: D timestamp: Tue 2008-01-01 00:22:23 -0500 message: Add side to check parameter to capture move generator Add option to check for all captures or just find one Add quiescense search to bot_opfor that does a full width search if the opponent has a capture or goal. It causes the search tree to explode though. ------------------------------------------------------------ revno: 52 committer: janzert@janzert.com branch nick: D timestamp: Sun 2007-12-30 14:29:27 -0500 message: Get rid of extra eval in quiescense search ------------------------------------------------------------ revno: 51 committer: janzert@janzert.com branch nick: D timestamp: Sun 2007-12-30 05:15:08 -0500 message: Goal search depth now specified at search time Various bugs in capture move generator fixed Alpha beta eval changed to take alpha-beta parameters Basic quiescense search added to bot_opfor ------------------------------------------------------------ revno: 50 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-12-28 12:24:26 -0500 message: Add ability to set options on alphabeta search through the arimaa engine interface ------------------------------------------------------------ revno: 49 committer: janzert@janzert.com branch nick: D timestamp: Sun 2007-12-23 01:34:19 -0500 message: An, I think, basically working time management system that should do for now. ------------------------------------------------------------ revno: 48 committer: janzert@janzert.com branch nick: D timestamp: Sat 2007-12-22 14:58:03 -0500 message: Change some of the stdout reporting to aei 'info log' commands ------------------------------------------------------------ revno: 47 committer: janzert@janzert.com branch nick: D timestamp: Sat 2007-12-22 03:30:33 -0500 message: Barely functioning time management. The decision to think longer is currently just random ------------------------------------------------------------ revno: 46 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-12-21 14:08:31 -0500 message: bot_opfor: Switch dog and horse placement in opening setup ------------------------------------------------------------ revno: 45 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-12-21 04:29:37 -0500 message: Add timecontrol information to bot_opfor. Now just need to do something with it. ------------------------------------------------------------ revno: 44 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-12-21 00:15:43 -0500 message: Add aei stop command and bot_opfor handling of it ------------------------------------------------------------ revno: 43 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-12-20 00:45:18 -0500 message: Fix bot_score to handle setoption ------------------------------------------------------------ revno: 42 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-12-20 00:29:09 -0500 message: Add aei setoption command Get rid of several aei go command modifiers Add ability to set search depth through 'setoption depth' ------------------------------------------------------------ revno: 41 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-12-19 18:00:05 -0500 message: Handle go ponder and goal by ignoring them ------------------------------------------------------------ revno: 40 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-12-19 15:29:22 -0500 message: aeibot.d: Add UnknownCommand exception to aeibot interface bot_opfor.d: Clean up the aei score reporting Adjust goal search weighting Reduce trap safety weight ------------------------------------------------------------ revno: 39 committer: janzert@janzert.com branch nick: D timestamp: Tue 2007-12-18 17:50:06 -0500 message: Add support for casual play game end rules (no rabbit capture win) Make goal search evaluation better handle multi move threats ------------------------------------------------------------ revno: 38 committer: janzert@janzert.com branch nick: D timestamp: Tue 2007-12-18 01:17:27 -0500 message: Rename bot_ab to bot_opfor ------------------------------------------------------------ revno: 37 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-12-13 10:09:48 -0500 message: Encapsulate transition table handling. Change transition table replacement to greater depth only. Small bug fixes and cleanups. ------------------------------------------------------------ revno: 36 committer: janzert@janzert.com branch nick: D timestamp: Mon 2007-12-03 17:46:32 -0500 message: Fix one major and a couple of minor bugs in arimaa score function. Add arimaa score evaluation to bot_ab ------------------------------------------------------------ revno: 35 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-11-30 13:23:15 -0500 message: Split alphabeta search class out from bot_ab ------------------------------------------------------------ revno: 34 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-11-28 00:17:03 -0500 message: Initial capture move generator ------------------------------------------------------------ revno: 33 committer: janzert@janzert.com branch nick: D timestamp: Mon 2007-11-26 01:08:18 -0500 message: A few more board constants ------------------------------------------------------------ revno: 32 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-11-23 12:02:27 -0500 message: Correct bug allowing pulls after a push step ------------------------------------------------------------ revno: 31 committer: janzert@janzert.com branch nick: D timestamp: Mon 2007-11-19 00:26:48 -0500 message: Add goal search to alpha beta bot ------------------------------------------------------------ revno: 30 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-11-16 21:25:19 -0500 message: Beginnings of a goal search ------------------------------------------------------------ revno: 29 committer: janzert@janzert.com branch nick: D timestamp: Mon 2007-11-12 17:20:10 -0500 message: More positional evaluation work. Infinite analysis support in alpha-beta bot. ------------------------------------------------------------ revno: 28 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-11-09 15:27:33 -0500 message: Some positional evaluation. More various stats recorded. ------------------------------------------------------------ revno: 27 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-11-07 11:46:44 -0500 message: Add History heuristic. Do some top level move ordering. Make the search run incrementally. ------------------------------------------------------------ revno: 26 committer: janzert@janzert.com branch nick: D timestamp: Tue 2007-11-06 11:59:07 -0500 message: Add transposition table to alpha beta Create caching FAME eval fastFAME ------------------------------------------------------------ revno: 25 committer: janzert@janzert.com branch nick: D timestamp: Sun 2007-11-04 23:31:24 -0500 message: Minimally working alpha beta bot. Fix some memory leaks. ------------------------------------------------------------ revno: 24 committer: janzert@janzert.com branch nick: D timestamp: Sat 2007-11-03 02:48:36 -0400 message: Initial alpha beta bot work ------------------------------------------------------------ revno: 23 committer: janzert@janzert.com branch nick: D timestamp: Sat 2007-11-03 02:48:03 -0400 message: Higher initial size for StepList ------------------------------------------------------------ revno: 22 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-11-02 13:32:19 -0400 message: combine get_[single/double]_steps methods ------------------------------------------------------------ revno: 21 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-11-01 14:06:47 -0400 message: Fix bug in move generation, don't allow frozen pieces to push a piece. Add length limited playouts. ------------------------------------------------------------ revno: 20 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-10-04 21:40:49 -0400 message: Get rid of unused import. Get rid of unnecessary code in random playout. ------------------------------------------------------------ revno: 19 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-10-04 20:40:06 -0400 message: Added FAME scoring Various fixes. UCT still plays much better as gold. ------------------------------------------------------------ revno: 18 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-10-03 10:40:28 -0400 message: Add aei info message Fix short position string parsing Get an analysis mode working for bot_uct ------------------------------------------------------------ revno: 17 committer: janzert@janzert.com branch nick: D timestamp: Sat 2007-09-29 11:53:08 -0400 message: Further work on UCT ------------------------------------------------------------ revno: 16 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-09-27 13:49:29 -0400 message: UCT based bot ------------------------------------------------------------ revno: 15 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-09-27 13:48:58 -0400 message: Arimaa score p1 bot ------------------------------------------------------------ revno: 14 committer: janzert@janzert.com branch nick: D timestamp: Thu 2007-09-27 13:48:06 -0400 message: Correct UCB algorithm Add Step.set(Step) method ------------------------------------------------------------ revno: 13 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-09-21 20:00:04 -0400 message: UCB based bot ------------------------------------------------------------ revno: 12 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-09-19 23:22:53 -0400 message: Add trap steps to move string output ------------------------------------------------------------ revno: 11 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-09-19 12:57:51 -0400 message: Oops forgot to add ------------------------------------------------------------ revno: 10 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-09-19 12:51:09 -0400 message: add copy method to steps rename random move bot create random step bot ------------------------------------------------------------ revno: 9 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-09-19 11:21:39 -0400 message: Split out AEI classes. ------------------------------------------------------------ revno: 8 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-09-19 10:31:11 -0400 message: Remove dead code. ------------------------------------------------------------ revno: 7 committer: janzert@janzert.com branch nick: D timestamp: Wed 2007-09-19 10:05:02 -0400 message: Fix bug with incorrect chains in PosStore after deletion of item. Add to_move_str method to StepList Stop memory leaking with current gc in get_moves complete implementation of random move bot ------------------------------------------------------------ revno: 6 committer: janzert@janzert.com branch nick: D timestamp: Sun 2007-09-16 14:35:09 -0400 message: Some Step refactoring. Small changes to updatefrozen and do_step/step generation ------------------------------------------------------------ revno: 5 committer: janzert@janzert.com branch nick: D timestamp: Fri 2007-09-14 20:36:13 -0400 message: More work on creating aei bot. ------------------------------------------------------------ revno: 4 committer: janzert@janzert.com branch nick: D timestamp: Sat 2007-09-08 23:12:31 -0400 message: Start of a uci-like bot interface ------------------------------------------------------------ revno: 3 committer: janzert@janzert.com branch nick: D timestamp: Tue 2007-09-04 11:40:06 -0400 message: Show real percentage ------------------------------------------------------------ revno: 2 committer: janzert@janzert.com branch nick: D timestamp: Mon 2007-09-03 21:20:29 -0400 message: Change steps to always be a single step. Also ran across and fixed a bug in PosStore.del_pos that stopped the position from sometimes being removed from the store. ------------------------------------------------------------ revno: 1 committer: Brian branch nick: D timestamp: Fri 2007-08-31 16:23:29 -0400 message: Initial import ------------------------------------------------------------