[FINAL RESULTS] Not that anyone cares, but I thought I should probably actually commit these. Congratulations Peter N Lewis See the mailing list archives for more information.
[PRELIMINARY ROUND 2] Finally. BLERGH.
[PATCH] Handle case where an AI sends an invalid message and then crashes If an AI sent a bad move and then crashed, a SIGPIPE would be reported rather than the AI's bad move. NOTE: There is an issue with Program::Running, because the segfaulted AI still returns 'true', which leads to confusion in handling SIGPIPE. Should probably handle segfaulting AIs better. Then again, people should probably not make segfaulting AIs. [SZM]
Committing results from Preliminary Round 1 (Rounds 1, 2, 3) Just in case...
[PRELIMINARY ROUND 1] Credits: -------- celcius - David Gow (sulix) ramen - John Hodge (tpg) peternlewis - Peter N Lewis Samples: basic_python basic_cpp (NOT COMPETING) basic_java (NOT COMPETING) asmodeus vixen hunter All samples are provided by Sam Moore The preliminary round is worth 10%; scores in simulate.py have been weighted accordingly Good luck!
[PATCH] Remove arguments, patch vixen, add hunter AI The arguments stuff has been causing errors for everyone. No one uses arguments for their AI, so I removed it. sulix now beats vixen, so I have patched it to not make illegal moves. hunter uses the same sort of idea as vixen, but seems to do slightly better. Enjoy [SZM]
Merge branch 'networking' of mussel.ucc.asn.au:progcomp2012 Conflicts: agents/basic_java/Makefile Stupid java AI broke everything
Minor rule changes simulate.py scores SURRENDER as 0 points for the surrendering AI, 3 points for its opponent. SURRENDER can be sent instead of a normal move. AIs must now place all 40 pieces during setup