gnugo-announce
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[gnugo-announce] GNU Go 3.1.16


From: Daniel Bump
Subject: [gnugo-announce] GNU Go 3.1.16
Date: Sun, 9 Dec 2001 10:59:37 -0800

GNU Go 3.1.16 is now available at:

http://www.gnu.org/software/gnugo/devel.html

The CVS release tag is rel-3-1-16.

This includes Tristan's revisions to readconnect and a configure option
--enable-experimental-connection with corresponding runtime options.  At the
moment it does absolutely nothing. However I hope we can use it in
experimental code calling readconnect from the engine.  (I think my previous
comments about how this could be used were clearly off-base.)

The next step in the semeai revisions will be to try and get
owl_analyze_semeai() working right. 

There may also be some issues with inessential worms. For example
strategy test 37 is broke with the experimental code because at
move 84 of nicklas8.sgf, a semeai is found between B18 and
D19. The old semeai module had some code after line 338 in
analyze_semeai which filtered out this semeai position as
"tactical." However adding this code to the new semeai module
broke more regressions than it fixed.

The following patches are included.

readconnect_1_16.1. Connection revisions. 
arend_1_16.4. New function add_influence_source fixes an inconsistency in 
influence.c. 
semeai_1_16.6. Dead code removed. 
arend_1_16.3.tar.gz and arend_1_16.3a. New tests based on 13x13 games. 
gunnar_1_16.8. Revision of connect.tst. 
trevor_1_16.6. New test case. 
semeai_1_16.5. Semeai revision
trevor_1_16.3. VC compatibility patch. 
arend_1_16.2. Revised trace message. 
trevor_1_16.5. Removes a test. 
trevor_1_16.4. Tuning. Added with revisions. 
gunnar_1_16.7. Important bugfix. 
gunnar_1_16.6. Test revisions. 
trevor_1_16.1. VC updates. 
trevor_1_16.2. Test corrections and html enhancements. 
gunnar_1_16.5. More atari_atari revisions. 
arend_1_16.1. Moyo recomputation. 
gunnar_1_16.4. Atari_atari revisions. 
semeai_1_16.4. Bogus code removed from analyze_semeai(). 
gunnar_1_16.3. Bugfix and 1D revisions. 
inge_1_16.1. Atari_atari revisions. 
compute_eyes_1_16.2. Bugfix in compute_eyes_pessimistic()
semeai_1_16.3. Semeai revision
semeai_1_16.2. Semeai revisions. 
gunnar_1_16.2. Atari_atari restructured. 
semeai_1_16.1. Semeai stones are owl_lively.

- readconnect revised
- new board.c function breakchain3()
- dsp and config.vc updates
- new function influence_get_moyo_segmentation()
- new function add_influence_source()
- computations of moyo size surrounding a dragon changed
- confirm_safety() takes two new parameters to tell which stones were saved
- atari_atari_confirm_safety() revised
- compute_aa_status() revised
- new is_vital_string() in combinations.c tracks stones not to be sacrificed
- generation of attack moves in do_atari_atari uses patterns
- new pattern database pattern/aa_attackpats.db
- atari_atari_find_defense_moves() improved
- new functions get_saved_worms() and get_saved_dragons() in move_reasons.c
- new function move_reasons_confirm_safety()
- examine_move_safety() revised
- new function decide_combination() in sgfdecide.c
- new option --decide-combination
- new experimental-combination configure option, and command line options
- new gtp command aa_confirm_safety
- bugfix in trymove()
- 1D revisions and cleaning up
- new file: engine/movelist.c
- generalized handling of move lists and moved from worm.c
- attack_threats() enhanced
- AA_THREAT_DEPTH set to 3
- find_worm_threats() uses attack_threats more
- pessimistic_compute_eyes() tries harder to find a vital point
- new atari_atari_find_attack_moves() and atari_atari_find_defense_moves()
- do_atari_atari() restructured
- experimental-semeai configure option
- experimental_semeai command line option
- new function new_semeai()
- owl_analyze_semeai improvements
- split USAGE_DEBUG in half due to VC limitation
- removed final VC warning in dfa.c
- revised trace in value_move_reasons()
- regress.plx improvements
- speedup in gtp_same_dragon
- test revisions
- 13x13 tests

Dan



reply via email to

[Prev in Thread] Current Thread [Next in Thread]