Hi James, Whoa! This is weird. Merlin chooses Kli vs. Ori. But human chooses Rom. In a 1 vs. 1 vs. 1. The bot will prefer Rom vs. Kli, and start spawning Roms or Klis. Once there is a 4 vs. 4 game, it will start to wipe out the 3rd team. This is what was supposed to happen. I think what happened here is: There was already a Kli vs. Ori 3 vs. 3 or maybe even a 4 vs. 4. Human joins Rom (somehow!), the server will allow it if t-mode hasn't been set yet. Since Merlin still prefers a Rom vs. Kli game, it will start to spawn Roms. Since, the server does not restrict which teams the bots can join, Roms will come in. But since, there is a 4 vs. 4 game Kli vs. Ori, Merlin will subsequently wipe a Rom out. Playing his own little polish game. Starting a Rom, cuz Merlin prefers a Rom vs. Kli game (doesn't check for t-mode when starting players). And subsequently wiping that same Rom out when it does it's 3rd team robot scummer check. I knew this race condition was possible. But, you have to have good timing in clicking Roms, to make this happen! U R good! newbiebetter.patch doesn't fix this. You have to have good timing to make this race condition happen everytime. But, once it starts, it doesn't stop until all human players quit. Let me see what I can do to fix this. Jimmy --- James Cameron <quozl at us.netrek.org> wrote: > In the current code in the darcs repo, if Merlin > chooses Kli and Ori as > the teams, but the first human player chooses Rom, > then Merlin begins > spawning the remainder of the robots as Rom. Each > robot spawned is then > obliterated to make room for a human (according to > the message to ALL). > > Eventually the terminator arrives to take out the > lone human, but Merlin > continues to insist on starting robots as Rom. > > See the recordings: > > http://quozl.linux.org.au/tmp/cambot-1.pkt.gz > http://quozl.linux.org.au/tmp/cambot-2.pkt.gz > > These can be played by gunzip'ing them and then > using "netrek -F > cambot-1.pkt" > > Any ideas, Jimmy? Does this relate to > newbiebetter.dpatch? > > -- > James Cameron mailto:quozl at us.netrek.org > http://quozl.netrek.org/ > > _______________________________________________ > netrek-dev mailing list > netrek-dev at us.netrek.org > http://mailman.us.netrek.org/listinfo/netrek-dev >