2010-10-23T00:09:34 *** jesionaj has joined #aichallenge 2010-10-23T00:11:17 Top 10 players: Hazard(4119), asavis(4098), rsergio(4081), felixcoto(4037), cfaftw(4011), bocsimacko(3957), dmj111(3943), sin_sun(3936), MariusCG(3910), RebelXT(3885) 2010-10-23T00:16:39 *** hellman has joined #aichallenge 2010-10-23T00:18:29 Zannick: if you're around; your compile_cleanup branch is fully merged now right? 2010-10-23T00:19:01 yep 2010-10-23T00:19:33 ok, I'll probably get rid of it unless you need for some more changes then 2010-10-23T00:21:10 i doubt there's anything left to do for it 2010-10-23T00:21:38 yeah, just thought I'd double check with you 2010-10-23T00:26:43 *** seisatsu has joined #aichallenge 2010-10-23T00:27:49 anyone know a better way than trial and error to find out what the maximum arguments length sudo will allow for a command? apparently the maximum length xargs calculates is longer than sudo allows 2010-10-23T00:28:23 janzert: can you sudo xargs? 2010-10-23T00:28:40 janzert: if it's so long, why not encapsulate it in a script file and then sudo run that script 2010-10-23T00:29:10 i imagine the limit is somewehre around 513 characters 2010-10-23T00:29:52 foucist: we could but it's a command run pretty frequently by the tournament 2010-10-23T00:30:17 *** rebelxt_ has quit IRC (Quit: Page closed) 2010-10-23T00:30:19 hoping to avoid the overhead of a script, but maybe I'm worried for nothing 2010-10-23T00:30:37 janzert: all the more reason to encapsulate it ? what overhead of a script? 2010-10-23T00:30:46 janzert: you running the comman within your code? 2010-10-23T00:31:28 yes, it is to clean up after a bot runs a game basically 2010-10-23T00:32:14 it is doing a straight sudo -u botuser rm -rf /botdirectory/* basically 2010-10-23T00:32:41 but if the bot has too many files sudo errors out 2010-10-23T00:33:14 try sudo -u botuser rm -rf /botdirectory/\* 2010-10-23T00:33:44 \* stops bash/etc from loading all the files into the args 2010-10-23T00:33:56 *** Accoun has quit IRC (Ping timeout: 265 seconds) 2010-10-23T00:34:04 but it will still behave the same 2010-10-23T00:34:13 hmm, yeah I assume something will still do the expansion after the sudo 2010-10-23T00:34:28 sounds like a much better fix 2010-10-23T00:35:53 bah, nope doesn't look like anything expands the glob then 2010-10-23T00:36:26 put an sh in? sudo -u botuser "sh rm -rf /botdirectory/\*" 2010-10-23T00:36:59 #repo 2010-10-23T00:37:09 contestbot: reoi 2010-10-23T00:37:10 amstan: I do not know about 'reoi', but I do know about these similar topics: 'repo' 2010-10-23T00:37:11 contestbot: repo 2010-10-23T00:37:12 amstan: repo = http://code.google.com/p/ai-contest/ 2010-10-23T00:40:39 *** Mjothvitnir has joined #aichallenge 2010-10-23T00:40:47 found it sudo sh -c "rm -rf botdirectory/*" :) 2010-10-23T00:41:15 janzert: what is that? 2010-10-23T00:41:56 a paraphrased fix for line 112 of user_sadbox.py in the cloud branch 2010-10-23T00:42:28 janzert: i assume this is for you: http://ai-contest.com/forum/viewtopic.php?f=18&t=877&start=30 2010-10-23T00:42:32 some bots leave enough files around that the current command has an error instead of cleaning up 2010-10-23T00:42:56 :) thanks 2010-10-23T00:43:09 yeah, they seem all happy 2010-10-23T00:43:11 I'm trying to keep at least two running 2010-10-23T00:43:15 it was all dark and gloomy for a while 2010-10-23T00:44:01 running my own when Jeff's go down, which unfortunately is about 6 hours after they get started lately 2010-10-23T00:44:48 janzert: so.. why do they go down? 2010-10-23T00:45:21 instability of ubuntu 8.04 on ec2 :( 2010-10-23T00:45:50 can't a script be done that does this? 2010-10-23T00:45:51 unfortunately I don't think he can give me access to restart them properly without completely opening his AWS account to me 2010-10-23T00:46:14 yes, but jeff will have to run it somewhere that he is comfortable putting his credentials 2010-10-23T00:46:32 for i in $(cat serverlist); do ssh $i hostname; done 2010-10-23T00:47:01 i can probably do it on hypertriangle 2010-10-23T00:47:02 when they crash the only way to get them back is a reboot from amazon interface itself 2010-10-23T00:47:06 he can kill me if i take them away 2010-10-23T00:47:12 there is no ssh or anything 2010-10-23T00:47:20 yeah, but at least you know they crashed.. 2010-10-23T00:47:23 or.. 2010-10-23T00:47:25 do we already know that? 2010-10-23T00:47:53 well, a glance at game_stats.php shows what is up 2010-10-23T00:47:59 i see 2010-10-23T00:48:13 well, you have to communicate with jeff, see what can happen 2010-10-23T00:48:30 *** rwa has joined #aichallenge 2010-10-23T00:48:31 i can potentially be a trustworthy source that can handle those details 2010-10-23T00:48:38 and i'm online most of the time 2010-10-23T00:48:57 yeah. I am hoping to create a script that will do it so it can watch my instances 2010-10-23T00:49:03 then he could use it as well 2010-10-23T00:49:14 yeah 2010-10-23T00:49:23 there's a few people that want golang support 2010-10-23T00:49:31 got a bit distracted fixing other bugs in our code 2010-10-23T00:49:32 as far as I saw, compile anything already has it 2010-10-23T00:49:35 man this competition is getting fierce 2010-10-23T00:49:49 janzert: all that's left to do is actually install golang 2010-10-23T00:50:12 install it on main server and add it to the instance setup script 2010-10-23T00:50:23 just trying to get instances sort of stable first 2010-10-23T00:50:42 well and actually I should probably try and sort out c# as well 2010-10-23T00:50:57 since it's suppose to be one of our primary languages 2010-10-23T00:51:02 yeah 2010-10-23T00:53:08 if jeff will give you access to his AWS account so you could reboot instances that would be great 2010-10-23T00:53:27 yeah, that might be possible 2010-10-23T00:53:46 one of the things that I feel bad about it is that since they are still running he is still getting charged even though they're doing nothing usefull 2010-10-23T00:55:37 i still don't get why we picked amazon clouds 2010-10-23T00:55:45 we have perfectly good comps in so many other places 2010-10-23T00:55:53 that really are spinning uselessly 2010-10-23T00:55:55 and they're free 2010-10-23T00:56:32 if we can get clean installs of ubuntu 8.04 and root access it would be easy to set them up now 2010-10-23T00:57:01 it's not completely automatic but I've done locally 2010-10-23T00:57:29 we just need a virtualbox image 2010-10-23T00:57:48 and we can deploy it on as many comps as possible 2010-10-23T00:57:52 @repo 2010-10-23T00:57:53 amstan: repo = http://code.google.com/p/ai-contest/ 2010-10-23T00:58:10 here were my ideas: http://code.google.com/p/ai-contest/wiki/Decentralization 2010-10-23T00:58:58 yeah, running just the bots remotely runs into lag issues though 2010-10-23T00:59:21 well.. w/e, how it's done now is fine too 2010-10-23T00:59:34 i just think it should be more modular 2010-10-23T00:59:57 there's not really anything in the current set up tied to amazon 2010-10-23T01:00:18 so why can't we do this on more servers? 2010-10-23T01:00:33 i have 3 comps doing nothing 2010-10-23T01:01:00 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-10-23T01:01:03 we can :) with the above restrictions at least and assuming enough power 2010-10-23T01:01:26 i just need an ubuntu image to deploy, and instructions on how to do stuff 2010-10-23T01:02:07 ok, I'll try write up step by step instructions from a fresh 8.04 install tomorrow 2010-10-23T01:02:50 janzert: while you do those, can you actually do them in a virtualbox vm? 2010-10-23T01:02:59 it would be much easier to deploy 2010-10-23T01:03:44 janzert: also.. what happens when a cloud server goes down? 2010-10-23T01:03:49 janzert: score wise 2010-10-23T01:03:57 do the bots loose? or is the match just ignored? 2010-10-23T01:04:01 it won't report any game result 2010-10-23T01:04:04 good 2010-10-23T01:04:10 so yes ignored basically 2010-10-23T01:04:17 *** hellman has quit IRC (Remote host closed the connection) 2010-10-23T01:04:23 and i assume they're plug and play, right, as in you can go online and offline with the servers as much as you want 2010-10-23T01:04:44 yeah, should be able to 2010-10-23T01:04:48 great 2010-10-23T01:05:02 so.. 2010-10-23T01:05:17 please do that, we have so many sources of comps, and we're not using them 2010-10-23T01:05:49 something we will need soon is a way to update cloud servers rather than just set them up from scratch 2010-10-23T01:05:54 we got donations from portland, at greghaynes' university 2010-10-23T01:06:05 janzert: update? 2010-10-23T01:06:07 if this takes off with many sources running them 2010-10-23T01:06:08 what do you mean? 2010-10-23T01:06:16 oh.. well 2010-10-23T01:06:18 update to the latest code 2010-10-23T01:06:34 you'll notice the code base isn't really static :) 2010-10-23T01:06:36 those ubuntu vms, get them setup so you can ssh into them as soon as they're deployed 2010-10-23T01:06:51 isn't that enough? 2010-10-23T01:06:58 one other problem I noticed this morning 2010-10-23T01:07:18 yeah, technically at least. doing manual updates over many servers will be painful 2010-10-23T01:07:44 the main server was getting overloaded with 5 instances requesting games 2010-10-23T01:08:06 of course once we have stable instances we can shut the games on the main server off 2010-10-23T01:08:15 which should help a lot 2010-10-23T01:08:15 yeah 2010-10-23T01:08:47 janzert: tell me if you need help with anything, i can do stuff, as long as i know what's going on 2010-10-23T01:09:14 janzert: also.. another thing.. will you support rewriting the codebase from scratch for the next contest? 2010-10-23T01:10:20 yeah, although I don't know how I'll feel at the end of this one about diving into another right away :} 2010-10-23T01:10:38 probably not right away 2010-10-23T01:11:08 *** delt0r has joined #aichallenge 2010-10-23T01:11:18 Top 10 players: Hazard(4112), rsergio(4095), asavis(4074), cfaftw(4051), felixcoto(4036), sin_sun(3964), bocsimacko(3962), MariusCG(3947), dmj111(3935), davidjliu(3900) 2010-10-23T01:11:47 *** delt0r___ has quit IRC (Ping timeout: 245 seconds) 2010-10-23T01:14:06 janzert: but at least a very long period for planning 2010-10-23T01:15:43 yeah, I probably will want to wait 4-6 months before doing another. which means you almost have to wait till next fall to do another 2010-10-23T01:20:00 i support a full rewriter :) 2010-10-23T01:20:05 rewrite * 2010-10-23T01:31:21 By support are you referring to $/hr or lip service?? 2010-10-23T01:31:31 'cause if you are talking about lipservice then I'm right there with you. :) 2010-10-23T01:33:31 *** Mjothvitnir has left #aichallenge 2010-10-23T01:37:00 *** Appleman1234 has joined #aichallenge 2010-10-23T01:42:43 well, i plan to work on it 2010-10-23T01:42:52 not that you can see this message since you're quit 2010-10-23T01:44:47 nor could he see that :P 2010-10-23T01:52:44 yes 2010-10-23T01:52:46 * Zannick shrug 2010-10-23T02:01:30 *** seisatsu has quit IRC (Quit: Leaving) 2010-10-23T02:11:08 sometimes I get surprised by my recursive screen levels :/ 2010-10-23T02:11:19 Top 10 players: Hazard(4106), rsergio(4067), cfaftw(4054), felixcoto(4045), asavis(3986), bocsimacko(3981), sin_sun(3980), MariusCG(3944), dmj111(3932), davidjliu(3902) 2010-10-23T02:12:13 Zannick: there's logs 2010-10-23T02:12:19 true 2010-10-23T02:17:16 *** hornairs has joined #aichallenge 2010-10-23T02:23:22 wassup 2010-10-23T02:25:43 @rankings 2010-10-23T02:25:44 Mathnerd314: Top 10 players: Hazard(4118), rsergio(4081), cfaftw(4070), felixcoto(4060), asavis(4024), bocsimacko(3999), sin_sun(3996), MariusCG(3964), dmj111(3929), Killer(3927) 2010-10-23T02:25:52 Hazard is up 2010-10-23T02:25:53 *** Rommel has joined #aichallenge 2010-10-23T02:27:15 also, some of the game-playing servers are up 2010-10-23T02:32:11 ^ bhasker 2010-10-23T02:32:15 *** Sarin has joined #aichallenge 2010-10-23T02:32:28 nice 2010-10-23T02:33:02 I, however, should not be up, as it is past midnight, so good night. 2010-10-23T02:33:25 west coast? 2010-10-23T02:33:57 mountains 2010-10-23T02:34:21 ah k 2010-10-23T02:47:33 *** nann has joined #aichallenge 2010-10-23T02:51:04 *** Apophis has quit IRC (Ping timeout: 250 seconds) 2010-10-23T02:51:58 *** Sarin has quit IRC (Remote host closed the connection) 2010-10-23T02:55:52 *** Titankiller has quit IRC (Ping timeout: 255 seconds) 2010-10-23T02:56:26 *** jesionaj has quit IRC (Ping timeout: 264 seconds) 2010-10-23T02:57:09 bhasker, what's your latest bot's name? 2010-10-23T02:58:11 *** seisatsu has joined #aichallenge 2010-10-23T02:59:18 *** Sarin has joined #aichallenge 2010-10-23T03:07:13 *** george_ has joined #aichallenge 2010-10-23T03:07:48 *** ryxy has joined #aichallenge 2010-10-23T03:11:19 Top 10 players: felixcoto(4097), Hazard(4092), rsergio(4051), cfaftw(4043), asavis(3988), bocsimacko(3987), sin_sun(3982), MariusCG(3965), davidjliu(3916), Killer(3915) 2010-10-23T03:15:09 never die: http://72.44.46.68/canvas?game_id=643551 2010-10-23T03:15:13 *** Accoun has joined #aichallenge 2010-10-23T03:19:05 *** hellman has joined #aichallenge 2010-10-23T03:21:47 *** Appleman1234 has quit IRC (Read error: Connection reset by peer) 2010-10-23T03:23:59 *** Titankiller has joined #aichallenge 2010-10-23T03:24:18 *** dmj111 has quit IRC (Ping timeout: 240 seconds) 2010-10-23T03:24:35 baggy g++ optimizer 2010-10-23T03:25:41 sigsetjmp(timer_halt::programState,1); optimised to unworket condition 2010-10-23T03:28:27 *** McLeopold has quit IRC (Ping timeout: 245 seconds) 2010-10-23T03:38:45 no idia how avoid it 2010-10-23T03:39:43 oh, hi Accoun 2010-10-23T03:41:02 *** Appleman1234 has joined #aichallenge 2010-10-23T03:41:20 *** tapwater has quit IRC (Quit: tapwater) 2010-10-23T03:43:52 Accoun: it's because setjmp is usually avoided by most applications 2010-10-23T04:11:20 Top 10 players: felixcoto(4081), Hazard(4077), rsergio(4032), cfaftw(4022), asavis(3970), bocsimacko(3966), sin_sun(3953), MariusCG(3948), davidjliu(3912), dmj111(3901) 2010-10-23T04:17:15 *** mega1_ has joined #aichallenge 2010-10-23T04:17:35 *** mega1 has quit IRC (Ping timeout: 240 seconds) 2010-10-23T04:27:46 *** mega1_ has quit IRC (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) 2010-10-23T04:27:52 *** mega1 has joined #aichallenge 2010-10-23T04:36:42 *** Rommel has quit IRC (Quit: Page closed) 2010-10-23T04:36:52 *** lavalamp_ has joined #aichallenge 2010-10-23T04:37:20 amstan, if you're awake-- I just sent you a PM on the forum 2010-10-23T04:37:46 I sent you one if you aren't awake, too... 2010-10-23T04:38:41 hmm 2010-10-23T04:38:44 so.. 2010-10-23T04:39:07 i think we're getting golang tomorrow, janzert said he'll look into it 2010-10-23T04:39:13 same with C# 2010-10-23T04:40:22 lavalamp_: 2010-10-23T00:49:23 @ http://contestbot.hypertriangle.com/channellogs/%23aichallenge/%23aichallenge.2010-10-23.log 2010-10-23T04:41:00 oh, awesome! 2010-10-23T04:41:02 lavalamp_: i'm not exactly sure what andrew could do for us, he should contact Jeff about participating in administration 2010-10-23T04:42:09 amstan, I don't know if you could use money for more servers or anything... it might not hurt to ask if so. I don't know what resources he has. 2010-10-23T04:42:31 lavalamp_: what we lack is organization, and time 2010-10-23T04:42:43 And I'm still not convinced Jeff actually exists :) 2010-10-23T04:42:57 we have lots of problems with the design, we should of planned this much better before launching 2010-10-23T04:42:59 and yeah.. 2010-10-23T04:43:08 jeff didn't show signs of survivol in.. 2010-10-23T04:43:12 @seen j3camero 2010-10-23T04:43:12 amstan: j3camero was last seen in #aichallenge 2 weeks, 4 days, 6 hours, 23 minutes, and 26 seconds ago: amstan: yes, but no names, haha. 2010-10-23T04:43:20 yeah... 2010-10-23T04:43:40 @seen dstuff 2010-10-23T04:43:40 amstan: I have not seen dstuff. 2010-10-23T04:43:49 hmm 2010-10-23T04:44:03 @seen dstufft 2010-10-23T04:44:03 amstan: dstufft was last seen in #aichallenge 4 weeks, 2 days, 13 hours, 49 minutes, and 46 seconds ago: there's nothing random in your bots is there? 2010-10-23T04:44:07 another one.. 2010-10-23T04:44:17 @seen danielvf 2010-10-23T04:44:17 amstan: danielvf was last seen in #aichallenge 2 weeks, 4 days, 7 hours, 3 minutes, and 16 seconds ago: * DanielVF out. :( 2010-10-23T04:44:23 OK I will respond to him then. I guess if janzert needs any assitance setting golang up I can put them in touch. 2010-10-23T04:44:31 it's me and janzert 2010-10-23T04:44:46 it's not hard... it's just installing it 2010-10-23T04:44:53 it's so odd that they seem to have skipped out on you like that 2010-10-23T04:45:03 for not the cloud servers are going down for no reason at all 2010-10-23T04:45:33 i would like to help, but i really don't like the organization, how it's svn, how i can't communicate to jeff and all that 2010-10-23T04:45:52 ah 2010-10-23T04:46:08 for this reason i don't feel like actually coding and stuff, since what i'll do will probably be useless 2010-10-23T04:46:29 Yeah I think I understand 2010-10-23T04:47:13 like it's an awesome contest.. but.. 2010-10-23T04:47:16 it's jeff's thing 2010-10-23T04:47:25 ...and he vanished 2010-10-23T04:47:53 we need a good talk to him 2010-10-23T04:48:14 but yeah, it's not about cpu time, or resources, or people to help 2010-10-23T04:48:25 ok, makes sense 2010-10-23T04:49:15 aw, come on! 2010-10-23T04:49:17 5 pm 2010-10-23T04:49:19 am* 2010-10-23T04:49:27 i should go zed, lol 2010-10-23T04:49:31 hahaha 2010-10-23T04:49:40 how the hell did the night go? 2010-10-23T04:49:46 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-10-23T04:49:57 *** Utkarsh has joined #aichallenge 2010-10-23T04:50:00 I've been to bed and up already 2010-10-23T04:50:14 you wake up at 5am? 2010-10-23T04:50:51 earlier, I sleep only a few hours at night and take a few naps during the day. 2010-10-23T04:51:01 lavalamp_: polyphasic? 2010-10-23T04:51:05 lavalamp_: how is that? 2010-10-23T04:51:38 amstan, yeah, it's interesting. Not easy to maintain a napping schedule, unfortunately. 2010-10-23T04:52:05 i'm not a schedule kinda guy, i don't think i could do it 2010-10-23T04:52:26 also, i'm not confident that it'll work during school and work 2010-10-23T04:52:33 neither was I, but the extra time is worth a lot of inconvienience to me... 2010-10-23T04:53:02 but most work places have this 8 hour block of working 2010-10-23T04:53:17 and polyphasic will require you to have a nap in the middle of that somewhere 2010-10-23T04:53:18 Yeah, I take a nap instead of lunch, it works for me but not everyone could pull that off, I'm sure... 2010-10-23T04:53:30 where do you nap? 2010-10-23T04:53:37 A supply closet :) 2010-10-23T04:54:16 what about physical rest? 2010-10-23T04:54:30 well, nvm, i guess this is not an issue with us computer types 2010-10-23T04:54:55 lol 2010-10-23T04:55:05 idk, i'm just scared 2010-10-23T04:55:40 I am still not convinced the more extreme schedules are actually doable, I certainly couldn't do them. 2010-10-23T04:56:27 Or actually, my current theory is that they are doable only by those who naturally have low SWS requirements. 2010-10-23T04:56:35 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-10-23T04:56:39 Anyway, go to bed :0 2010-10-23T04:56:45 SWS? 2010-10-23T04:56:52 slow-wave-sleep 2010-10-23T04:56:56 oh 2010-10-23T04:57:09 stage 3 (used to be stage 4) sleep 2010-10-23T04:58:09 cool 2010-10-23T04:58:14 well, i'm going to zed now 2010-10-23T04:59:16 zed's dead 2010-10-23T05:11:21 Top 10 players: felixcoto(4068), Hazard(4047), rsergio(4018), cfaftw(3994), MariusCG(3954), asavis(3947), bocsimacko(3911), davidjliu(3909), dmj111(3908), sin_sun(3895) 2010-10-23T05:11:57 *** fio175 has joined #aichallenge 2010-10-23T05:12:13 *** Apophis has joined #aichallenge 2010-10-23T05:13:01 *** nann has quit IRC (Ping timeout: 252 seconds) 2010-10-23T05:33:11 i'm curious as to whether anyone is correctly doing future states with future moves from both players 2010-10-23T05:33:16 i have no idea how the fuck to do it 2010-10-23T05:33:36 its impossible 2010-10-23T05:33:52 antimatroid: thats the idea of minimax 2010-10-23T05:34:01 yup 2010-10-23T05:34:08 but the branching factor is massive 2010-10-23T05:34:12 which causes loads of problems 2010-10-23T05:34:14 yes, but i need it to do minimax :P 2010-10-23T05:34:20 it makes Go look simpler 2010-10-23T05:34:42 minimax on Go has been pretty much dismissed 2010-10-23T05:35:06 *** wvdschel has joined #aichallenge 2010-10-23T05:35:06 *** wvdschel has joined #aichallenge 2010-10-23T05:35:14 the only way minimax could work is if you prune the search tree early enough and not lose the benefits of minimax 2010-10-23T05:35:46 and its *very easy to lose* the benefits if you do prune early 2010-10-23T05:35:48 the pruning is the problem 2010-10-23T05:35:58 because you need to prune correctly 2010-10-23T05:36:00 hehe 2010-10-23T05:36:08 =] 2010-10-23T05:36:52 no doubt ill try a minimax approach after this one 2010-10-23T05:36:56 which im still doing ugh 2010-10-23T05:37:01 wish i could code fast 2010-10-23T05:37:03 faster* 2010-10-23T05:37:06 i need louder music 2010-10-23T05:37:23 thats better 2010-10-23T05:37:34 mmm coding with music is distracting 2010-10-23T05:37:35 =) 2010-10-23T05:37:39 nein 2010-10-23T05:37:53 helps me focus 2010-10-23T05:38:04 even if you want to do coordinated attacks on minimax 2010-10-23T05:38:12 it breaks unless you can do future moves in your data 2010-10-23T05:38:24 is it actually not possible? or just really fucking hard 2010-10-23T05:38:37 what do you mean *future moves in your data* 2010-10-23T05:38:44 im not entirely sure whats the problem 2010-10-23T05:38:56 say i want to send fleets from 5 planets to planet j, not all the same distance and arrive at the same time 2010-10-23T05:39:01 i have to make some of those moves in the future 2010-10-23T05:39:22 but then dependent on future moves that get made, that planet may not have that many ships or even be owned by the player 2010-10-23T05:39:34 oh 2010-10-23T05:39:36 one can try and add outgoing fleet information all over the place, but it still breaks down i think 2010-10-23T05:39:59 im not going to do that, i recalculate every move 2010-10-23T05:40:04 i move and *forget* 2010-10-23T05:40:08 when i get a new state 2010-10-23T05:40:10 yeah so do i 2010-10-23T05:40:24 IMO, keeping ideas what you're going to do in the future is wrong 2010-10-23T05:40:34 i don't want to keep them 2010-10-23T05:40:40 this is not chess 2010-10-23T05:40:40 but i want to make a tree, with players alternating turns 2010-10-23T05:40:50 if i make a future turn, that can be rendered useless, it causes issues 2010-10-23T05:41:01 each turn i recreate my information 2010-10-23T05:41:01 yea, don't create future moves 2010-10-23T05:41:16 how else do you make sure it works? 2010-10-23T05:41:43 say you need to send ships to the planet, but they can't arrive any time but the same time 2010-10-23T05:42:04 then you must send some in the future at a point in time 2010-10-23T05:42:08 it's unavoidable 2010-10-23T05:42:20 oh easy, basically you see where your other fleets are heading and see if they need help 2010-10-23T05:42:39 and send them off in that turn if you exactly that distance away 2010-10-23T05:42:48 you=that planet 2010-10-23T05:43:39 that doesn't make any sense 2010-10-23T05:43:45 it doesn't? 2010-10-23T05:44:16 an example that makes sense, you want to take a neutral just after the opponent, you can't send some of the ships to arrive early, otherwise you just make it cheaper for the opponent to take the planet 2010-10-23T05:44:37 but if you add those moves into the future to information, you run into the problems i mention above and it all falls to pieces 2010-10-23T05:44:37 yes 2010-10-23T05:45:13 well it really depends on your implementation 2010-10-23T05:45:21 but one way is to store future moves 2010-10-23T05:45:26 *** Palmik has joined #aichallenge 2010-10-23T05:45:37 yeah, but then finding out when to delete them is just messy 2010-10-23T05:45:39 and add constraints that if planet sends x ships away, do not violate future move 2010-10-23T05:45:43 and i'm not sure how i'd do it even works well 2010-10-23T05:45:50 if future move is violated, remove future move 2010-10-23T05:46:17 i.e. if the enemy attacks your planets and you don't have enough resources for that future move 2010-10-23T05:46:33 but there's better ways of doing this 2010-10-23T05:46:58 have you actually implemented something like this yet? 2010-10-23T05:47:05 it seemed easier to me before i tried doing it 2010-10-23T05:48:07 antimatroid: I'm solving this by considering future moves but not actually adding them to my state 2010-10-23T05:48:33 i've considered that, but then how do you evaluate your final state? 2010-10-23T05:48:39 antimatroid: im implementing it differently 2010-10-23T05:48:45 so that situation would not crop up 2010-10-23T05:48:56 i only consider what i see now, and not what i don't need to see :P 2010-10-23T05:49:16 yeah, well, my bot is terrible currently XP 2010-10-23T05:49:34 at least mine is theoretical 2010-10-23T05:49:38 nothing has beaten it yet 2010-10-23T05:49:40 :D 2010-10-23T05:50:01 undefeated... theoretical 2010-10-23T05:50:05 just about nothing hasn't beaten mine... http://72.44.46.68/getplayer?player=lavalamp.forfun.1 2010-10-23T05:50:26 *** fio175 has quit IRC (Ping timeout: 265 seconds) 2010-10-23T05:50:50 hehe 2010-10-23T05:51:00 *** Palmik has quit IRC (Ping timeout: 252 seconds) 2010-10-23T05:51:25 lavalamp_: interesting game.. http://72.44.46.68/canvas?game_id=645550 2010-10-23T05:52:18 :P 2010-10-23T05:52:22 raedwulf, sniping obviously is not working, haha 2010-10-23T05:52:57 well no doubt my first bot version is going to fail ^^ 2010-10-23T05:55:36 at least i've written lots of debugging code 2010-10-23T05:55:58 so i can list *why* exactly my bot choses the move it chose 2010-10-23T05:56:04 chooses* 2010-10-23T06:00:10 *** Titankiller has quit IRC (Ping timeout: 240 seconds) 2010-10-23T06:00:22 *** Palmik has joined #aichallenge 2010-10-23T06:04:34 *** Frontier has joined #aichallenge 2010-10-23T06:08:27 *** Bobng has joined #aichallenge 2010-10-23T06:11:22 Top 10 players: felixcoto(4060), rsergio(4022), Hazard(3999), cfaftw(3968), bocsimacko(3938), MariusCG(3919), asavis(3913), dmj111(3905), Killer(3900), davidjliu(3898) 2010-10-23T06:14:18 *** hornairs has quit IRC (Ping timeout: 250 seconds) 2010-10-23T06:20:32 *** mega1 has quit IRC (Ping timeout: 245 seconds) 2010-10-23T06:30:28 *** Frontier has quit IRC (Ping timeout: 276 seconds) 2010-10-23T06:40:09 *** Titankiller has joined #aichallenge 2010-10-23T06:46:54 *** ryxy has left #aichallenge 2010-10-23T06:54:09 *** JensT1 has joined #aichallenge 2010-10-23T07:11:16 *** Titankiller has quit IRC (Ping timeout: 276 seconds) 2010-10-23T07:11:21 Top 10 players: felixcoto(4067), rsergio(4017), Hazard(4012), cfaftw(3963), sin_sun(3906), bocsimacko(3901), dmj111(3896), davidjliu(3892), MariusCG(3885), Killer(3852) 2010-10-23T07:13:45 people have become a lot more competiitive in the last week 2010-10-23T07:17:05 *** abcd123 has joined #aichallenge 2010-10-23T07:21:35 *** yasith has quit IRC (Ping timeout: 240 seconds) 2010-10-23T07:21:54 *** area has quit IRC (Ping timeout: 240 seconds) 2010-10-23T07:24:06 *** abcd123 has quit IRC (Read error: Connection reset by peer) 2010-10-23T07:29:54 *** Utkarsh has joined #aichallenge 2010-10-23T07:31:17 *** mega1 has joined #aichallenge 2010-10-23T07:35:03 *** yasith has joined #aichallenge 2010-10-23T07:36:50 *** mceier has joined #aichallenge 2010-10-23T07:38:31 *** JensT1 has left #aichallenge 2010-10-23T07:49:24 morning mega1 2010-10-23T07:49:31 good afternoon 2010-10-23T07:49:35 afternoon here, but anyways :P 2010-10-23T07:49:41 oh ok, where do you hail from? 2010-10-23T07:50:20 hungary 2010-10-23T07:50:39 oh right ^^ 2010-10-23T07:50:45 << UK 2010-10-23T07:52:22 the bots are weird, its like they have a massive winning streak then suddenly fail with a losing streak 2010-10-23T07:52:54 well not exactly a losing *streak* but enough loses to shift the ranking again 2010-10-23T07:54:17 interesting, a miscalculation?: http://www.ai-contest.com/visualizer.php?game_id=6068264 2010-10-23T07:56:08 I think that's how the rankings work, you'll face people worse than you until you move up and start facing people better than you 2010-10-23T07:57:18 oh ok 2010-10-23T07:58:42 Just a guess though, I haven't read up much on how ELO works, spending all my free time trying to get something worth submitting 2010-10-23T07:59:27 *** hellman has quit IRC (Remote host closed the connection) 2010-10-23T08:00:20 *** edcba has joined #aichallenge 2010-10-23T08:01:57 accoun is the bane of my existence 2010-10-23T08:02:38 hehe 2010-10-23T08:05:26 on the tcp server 100 elo point advantage translates to 75% chance to win, right? 2010-10-23T08:08:06 * raedwulf has no idea 2010-10-23T08:10:05 I have about 200 points on him but lose more often than not, or maybe even. 2010-10-23T08:10:48 crap, just lost to an opponent with -288 elo :-o 2010-10-23T08:11:00 lol 2010-10-23T08:11:22 Top 10 players: felixcoto(4036), Hazard(4009), rsergio(4008), cfaftw(3928), bocsimacko(3906), sin_sun(3904), dmj111(3893), davidjliu(3863), Killer(3823), MariusCG(3821) 2010-10-23T08:25:20 *** Mekanik has joined #aichallenge 2010-10-23T08:33:21 *** iNtERrUpT has joined #aichallenge 2010-10-23T08:41:46 *** wvdschel has quit IRC (Quit: Leaving) 2010-10-23T08:42:45 *** wvdschel has joined #aichallenge 2010-10-23T08:42:46 *** wvdschel has joined #aichallenge 2010-10-23T08:44:14 *** krokkrok has joined #aichallenge 2010-10-23T08:58:14 *** p4p4 has joined #aichallenge 2010-10-23T08:58:21 wrahh cloud offline again >_> 2010-10-23T09:02:06 *** Meatkat has joined #aichallenge 2010-10-23T09:02:55 *** RainCT has joined #aichallenge 2010-10-23T09:05:34 anyone know who bocsimacko-b0.22 is on tcp? 2010-10-23T09:06:20 ah, it's probably bocsimacko 2010-10-23T09:06:29 ... >.< 2010-10-23T09:06:34 *** dmj111 has joined #aichallenge 2010-10-23T09:06:48 Meatkat: he uses a different name on irc 2010-10-23T09:06:51 i forget who 2010-10-23T09:06:53 but he's here 2010-10-23T09:09:23 right-o, thanks 2010-10-23T09:09:26 *** Naktibalda has joined #aichallenge 2010-10-23T09:09:57 check the irc logs for yesterday 2010-10-23T09:10:15 do a search for bocksimacko 2010-10-23T09:10:20 mega1 2010-10-23T09:10:27 dmj111: hello 2010-10-23T09:10:30 *** delt0r___ has joined #aichallenge 2010-10-23T09:10:30 ah it is mega1 :P, if forgot lol xD 2010-10-23T09:10:40 *** JensT1 has joined #aichallenge 2010-10-23T09:10:43 hello. sorry mega1, just speaking up. 2010-10-23T09:10:47 ah, I see 2010-10-23T09:11:17 good to see your bot working well on the official server :) 2010-10-23T09:11:23 Top 10 players: felixcoto(4005), rsergio(3983), Hazard(3980), cfaftw(3904), bocsimacko(3888), sin_sun(3882), dmj111(3864), davidjliu(3834), Killer(3816), evja08(3804) 2010-10-23T09:11:35 *** delt0r has quit IRC (Ping timeout: 240 seconds) 2010-10-23T09:14:24 it's hard to tell at this rate 2010-10-23T09:16:41 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-10-23T09:24:09 *** krokkrok has quit IRC (Quit: Page closed) 2010-10-23T09:27:36 *** hellman has joined #aichallenge 2010-10-23T09:30:16 mega1: you have a point there... the ratings are not all the meaningful yet. Better than win/loss. At least you have worked around the lisp issues. 2010-10-23T09:30:17 later. 2010-10-23T09:32:41 *** seisatsu has quit IRC (Quit: Leaving) 2010-10-23T09:43:11 *** hellman has quit IRC (Remote host closed the connection) 2010-10-23T10:02:35 *** hellman has joined #aichallenge 2010-10-23T10:04:59 *** davidjliu has joined #aichallenge 2010-10-23T10:05:52 *** rebelxt_ has joined #aichallenge 2010-10-23T10:06:02 *** yasith_ has joined #aichallenge 2010-10-23T10:07:18 *** AJC has joined #aichallenge 2010-10-23T10:08:52 *** yasith has quit IRC (Ping timeout: 276 seconds) 2010-10-23T10:11:23 Top 10 players: felixcoto(3992), rsergio(3970), Hazard(3967), cfaftw(3892), bocsimacko(3875), sin_sun(3869), dmj111(3851), davidjliu(3822), Killer(3803), evja08(3790) 2010-10-23T10:19:21 *** Bobng has quit IRC (Remote host closed the connection) 2010-10-23T10:19:46 *** Bobng has joined #aichallenge 2010-10-23T10:23:12 *** Bobng has quit IRC (Read error: Connection reset by peer) 2010-10-23T10:25:00 *** Bobng has joined #aichallenge 2010-10-23T10:26:51 *** rebelxt_ has quit IRC (Quit: Page closed) 2010-10-23T10:33:26 *** abcd123 has joined #aichallenge 2010-10-23T10:34:23 *** abcd123 is now known as Titankiller 2010-10-23T10:34:53 *** Titankiller has joined #aichallenge 2010-10-23T10:36:24 *** Titankiller has joined #aichallenge 2010-10-23T10:37:37 *** Frontier has joined #aichallenge 2010-10-23T10:40:17 *** BtbN has quit IRC (Quit: leaving) 2010-10-23T10:42:06 *** Titankiller has quit IRC (Ping timeout: 250 seconds) 2010-10-23T10:44:56 *** Titankiller has joined #aichallenge 2010-10-23T10:46:07 *** iNtERrUpT has quit IRC (Remote host closed the connection) 2010-10-23T10:51:10 *** Meatkat has quit IRC (Quit: Leaving) 2010-10-23T11:03:54 morning 2010-10-23T11:05:11 *** Bobng has quit IRC (Quit: Leaving) 2010-10-23T11:11:24 Top 10 players: felixcoto(3999), Hazard(3981), rsergio(3972), bocsimacko(3889), cfaftw(3880), dmj111(3864), sin_sun(3863), davidjliu(3804), evja08(3800), Killer(3788) 2010-10-23T11:18:02 *** BtbN has joined #aichallenge 2010-10-23T11:20:55 *** Zalma has joined #aichallenge 2010-10-23T11:23:08 *** Florian95 has joined #aichallenge 2010-10-23T11:24:10 Hi. I'm really noob on coding but how do I compile my bot done with c++? =) 2010-10-23T11:25:12 I'm currenly using visual studio 2010 pro... 2010-10-23T11:30:04 Zalma, look for the build command 2010-10-23T11:30:30 I'm on an older version at work so I can't say exactly where it is 2010-10-23T11:31:26 well I cant even debug.. even that is gray 2010-10-23T11:33:20 you may need to set up a project 2010-10-23T11:33:45 try making a new command line app and adding your files 2010-10-23T11:34:00 ...if you already did that, ignore me 2010-10-23T11:34:41 use gcc instead :) 2010-10-23T11:35:55 no I haven't done that.. 2010-10-23T11:36:05 how do I actually add the files to project? =) 2010-10-23T11:37:19 I have used c++ last time in dos age ;) 2010-10-23T11:37:54 *** rebelxt has joined #aichallenge 2010-10-23T11:39:25 Zalma: mingw gcc is another compiler - thats similar to the one used on the server. Of course, it might take some familiarity to set up. 2010-10-23T11:39:38 have you any experience with unix? 2010-10-23T11:39:45 nope =) 2010-10-23T11:39:54 ok 2010-10-23T11:40:12 I've been banging my head today for 10 hours already 2010-10-23T11:40:45 technically you can use any C compiler (although the msvc c compiler doesn't support modern standards of C as well) 2010-10-23T11:40:52 I wish visual basic would work.. :P 2010-10-23T11:40:56 hehe 2010-10-23T11:41:04 you could put in a request 2010-10-23T11:41:08 they have c# 2010-10-23T11:41:13 *** rebelxt has left #aichallenge 2010-10-23T11:41:21 if you ask nicely, they could add visual basic .net 2010-10-23T11:41:58 Thou visual basic .net would be quite similar.. I'm more familiar on old visual basic (without .net) =P 2010-10-23T11:41:59 note they use the mono version of .net, which is compatible with an older version of .net 2010-10-23T11:42:05 oh ok 2010-10-23T11:42:24 atleast I did read doing C# isn't worth doing.. compiling is hard or something 2010-10-23T11:42:43 yes, its because the age of the version of mono being used 2010-10-23T11:42:58 if you're using msvs.net its a bitch to get it to work with the server 2010-10-23T11:43:31 a lot of people are using python and java 2010-10-23T11:43:59 lua works too, but no one has written a starter package for that 2010-10-23T11:44:26 I'm just trying to get even some language to work 2010-10-23T11:44:28 i stand corrected, http://ai-contest.com/forum/viewtopic.php?f=18&t=615 2010-10-23T11:44:49 got like 50x error while compiling submission today already 2010-10-23T11:44:59 only ones that really work were started packets.. 2010-10-23T11:45:18 I have already tried java, C# and C++ 2010-10-23T11:45:27 getting pretty much frustrated 2010-10-23T11:45:41 especially java versions work well locally 2010-10-23T11:45:49 and compile without errors 2010-10-23T11:46:21 sorry you're using C++ (msvc 2010 support for c++ is excellent) heh 2010-10-23T11:46:31 and finally when I got it to compile it crashed at first turn.. (even thou it did work well at locally) 2010-10-23T11:46:48 im not very familiar with java 2010-10-23T11:46:55 i am quite familiar with C# though 2010-10-23T11:47:19 but i prefer C... i like being close to metal :P 2010-10-23T11:47:20 Well it doesn't matter what I get working easiest... I 2010-10-23T11:47:24 I'm bad in them all 2010-10-23T11:47:25 Zalma: can't you get someone else's java/c++ program that doesn't crash 2010-10-23T11:47:27 and modify? 2010-10-23T11:47:51 well as soon as I start to modify something it doesn't work anymore in server ;) 2010-10-23T11:47:58 well he's on windows and unfamiliar with C++, so it doesn't plug and play 2010-10-23T11:49:07 raed: ehm? 2010-10-23T11:49:51 well, the start package is only useful if you can compile it on a unix environment; you can compile it on msvc 2010 as well, but you need to create the project files yourself 2010-10-23T11:49:55 and a solution file 2010-10-23T11:50:20 on unix, its extract, and run "make" 2010-10-23T11:50:22 and it works 2010-10-23T11:50:24 :P 2010-10-23T11:50:54 Well I do have ubuntu on my virtual box.. 2010-10-23T11:51:02 if compiling with that is easier 2010-10-23T11:51:08 in that case, yep it would 2010-10-23T11:51:19 basically you can extract it on there 2010-10-23T11:51:23 open a terminal 2010-10-23T11:51:32 into that directory - then run "make" 2010-10-23T11:51:42 it should compile to give you "MyBot" 2010-10-23T11:51:45 which you can then run 2010-10-23T11:51:58 well using the java tool 2010-10-23T11:54:00 need to learn using that virtual box first :P 2010-10-23T11:54:15 well atleast that I can access to my files 2010-10-23T11:54:47 *** tapwater has joined #aichallenge 2010-10-23T11:54:55 *** AJC has left #aichallenge ("Leaving") 2010-10-23T11:59:49 *** JensT1 has left #aichallenge 2010-10-23T12:00:43 hehe 2010-10-23T12:06:14 meatkat around? 2010-10-23T12:06:54 his bot is quite interesting 2010-10-23T12:07:01 if it can't hold a planet it runs with the ships on it 2010-10-23T12:07:01 http://72.44.46.68/canvas?game_id=650096 2010-10-23T12:07:09 pretty interesting to watch 2010-10-23T12:08:02 sorta 2010-10-23T12:08:21 oh the 72 was already on the way 2010-10-23T12:08:34 my bot will abandon planets now and then too 2010-10-23T12:10:08 *** Florian95 has quit IRC (Quit: Leaving...) 2010-10-23T12:11:24 Top 10 players: felixcoto(3999), rsergio(3965), Hazard(3949), cfaftw(3887), sin_sun(3866), bocsimacko(3863), dmj111(3851), davidjliu(3826), evja08(3796), Killer(3783) 2010-10-23T12:20:57 *** Frontier has quit IRC (Ping timeout: 252 seconds) 2010-10-23T12:28:26 *** narnach has quit IRC (Quit: Leaving.) 2010-10-23T12:46:07 *** Mjothvitnir has joined #aichallenge 2010-10-23T12:51:17 *** BtbN has quit IRC (Quit: Verlassend) 2010-10-23T12:51:24 *** BtbN has joined #aichallenge 2010-10-23T12:58:41 32 game winning streak on tcp :-) 2010-10-23T12:58:57 Against the same person? :p 2010-10-23T12:59:10 *** hellman has quit IRC (Remote host closed the connection) 2010-10-23T12:59:52 *** Mjothvitnir has quit IRC (Quit: Page closed) 2010-10-23T13:01:49 mega1: awesome 2010-10-23T13:02:00 timed out :-( 2010-10-23T13:02:04 :S 2010-10-23T13:02:11 you spoke too soon! 2010-10-23T13:03:42 * raedwulf is close to getting a bot to test 2010-10-23T13:05:26 *** Olathe has joined #aichallenge 2010-10-23T13:11:25 Top 10 players: felixcoto(3993), rsergio(3991), Hazard(3973), cfaftw(3920), bocsimacko(3891), sin_sun(3876), dmj111(3874), davidjliu(3831), MariusCG(3788), evja08(3781) 2010-10-23T13:21:30 *** Mekanik has quit IRC (Ping timeout: 240 seconds) 2010-10-23T13:36:08 *** Mekanik has joined #aichallenge 2010-10-23T14:00:39 * raedwulf finds a bug in his code 2010-10-23T14:00:42 brilliant >.< 2010-10-23T14:01:48 fixed it... bugs BEGONE 2010-10-23T14:09:02 *** smellyhippy has quit IRC (Ping timeout: 264 seconds) 2010-10-23T14:11:24 Top 10 players: rsergio(3992), felixcoto(3990), Hazard(3970), cfaftw(3915), bocsimacko(3903), sin_sun(3870), dmj111(3869), davidjliu(3827), MariusCG(3783), jambachili(3777) 2010-10-23T14:14:14 *** smellyhippy has joined #aichallenge 2010-10-23T14:21:40 *** Mekanik has quit IRC (Quit: WeeChat 0.3.3) 2010-10-23T14:30:31 wow, lost to JamesMG.old at -171 elo 2010-10-23T14:30:41 http://72.44.46.68/canvas?game_id=651867 2010-10-23T14:30:55 he seems to wait it out well 2010-10-23T14:32:01 *** choas has joined #aichallenge 2010-10-23T14:32:46 *** fawek has joined #aichallenge 2010-10-23T14:33:11 *** mrflippy has joined #aichallenge 2010-10-23T14:45:05 *** askinAround has joined #aichallenge 2010-10-23T14:45:59 can I still register for this challenge? 2010-10-23T14:46:39 ok bye 2010-10-23T14:46:41 askinAround: yes, 2010-10-23T14:46:50 askinAround: it's not over until nov 21 or something like that 2010-10-23T14:47:01 sorry I thought no one was here 2010-10-23T14:47:08 thank you 2010-10-23T14:47:22 askinAround: there's 83 ppl in here, there's definitely ppl around.. :P 2010-10-23T14:47:33 are you a contestant or an admin? 2010-10-23T14:47:34 minus a few bots 2010-10-23T14:47:38 neither 2010-10-23T14:47:43 ? 2010-10-23T14:47:57 i might be a contestant if i got off my butt and did something 2010-10-23T14:48:04 but ive been busy with other stuff heh 2010-10-23T14:48:37 is registration broken again? 2010-10-23T14:48:43 haha I think I'll give it a try - I hope I could make a basic thing quick 2010-10-23T14:49:01 I need to register too 2010-10-23T14:49:04 let's see 2010-10-23T14:50:06 mega1: how long have you worked on your bot? 2010-10-23T14:50:46 a month now 2010-10-23T14:50:57 askinAround: you waited a total between 39 seconds between your first and second message... we don't respond immediate :P 2010-10-23T14:51:00 mega1: nice 2010-10-23T14:51:01 but I'm working full time 2010-10-23T14:51:14 a1k0n: I succeeded registering 2010-10-23T14:51:21 im a full-time student :P but it is start of term 2010-10-23T14:51:22 hehe 2010-10-23T14:51:52 lol :P I guess I was trying to find a quick way out 2010-10-23T14:51:57 lol 2010-10-23T14:53:16 Now I need to wait for my email 2010-10-23T14:54:40 *** davidjliu has quit IRC (Ping timeout: 265 seconds) 2010-10-23T14:55:09 *** george_ has quit IRC (Ping timeout: 265 seconds) 2010-10-23T14:56:25 any c++ boost programmers around 2010-10-23T15:01:50 *** RainCT has quit IRC (Ping timeout: 264 seconds) 2010-10-23T15:03:39 *** RainCT has joined #aichallenge 2010-10-23T15:10:20 janzert, I've used it a bit 2010-10-23T15:11:25 Top 10 players: rsergio(3957), Hazard(3951), felixcoto(3950), bocsimacko(3914), cfaftw(3899), sin_sun(3856), dmj111(3854), davidjliu(3799), jambachili(3768), evja08(3761) 2010-10-23T15:17:55 I sent the package as downloade and it said that my bot did not start 2010-10-23T15:18:06 shoud it do that? 2010-10-23T15:29:18 *** Accoun has quit IRC () 2010-10-23T15:32:21 *** gerald512kk has joined #aichallenge 2010-10-23T15:32:24 hi all :) 2010-10-23T15:33:49 askinAround: it /should/ work but there seems to be some errors around, try resubmitting and it will probably work that time 2010-10-23T15:34:55 lavalamp_: it seems boost is already installed on the servers so it seems like the only thing stopping it from being used is that it's not getting linked in 2010-10-23T15:35:32 janzert: good luck with that, I have only ever set it up on a windows machine 2010-10-23T15:35:34 lavalamp_: my question is what is the correct command line for linking with boost 2010-10-23T15:35:41 ok 2010-10-23T15:35:53 yeah sorry I don't know that :( 2010-10-23T15:36:33 my guess would be you just have to get the include paths right but boost is friggin' complicated so there might be something special you need to do... 2010-10-23T15:36:58 yeah, I was hoping it was as simple as adding -lboost or some such :) 2010-10-23T15:37:13 it's probably -lboost1.36.53.2358 2010-10-23T15:37:24 ahh, well I've got more pressing things anyway :/ 2010-10-23T15:37:25 depending on which of the many mutually-incompatible versions out in the wild you have installed 2010-10-23T15:37:54 it's the ubuntu hardy libboost-dev package if someone wants to investigate 2010-10-23T15:40:42 *** Accoun has joined #aichallenge 2010-10-23T15:45:30 yay my bot compiles again 2010-10-23T15:47:08 janzert, i think they're generating a makefile (prbly from python) based on the files you submit, so if it's that header-only part of boost, it might work with just adding them to your zip 2010-10-23T15:47:41 no makefile is generated in the compilation of your bots :) 2010-10-23T15:48:09 see compile_anything.py if you really want to see how it's done 2010-10-23T15:48:13 :) 2010-10-23T15:48:15 ;) 2010-10-23T15:48:40 where's it ? 2010-10-23T15:48:47 #repo 2010-10-23T15:48:53 @rep 2010-10-23T15:48:53 janzert: I do not know about 'rep', but I do know about these similar topics: 'repo', 'botrepo', 'repository' 2010-10-23T15:48:57 @repo 2010-10-23T15:48:58 janzert: repo = http://code.google.com/p/ai-contest/ 2010-10-23T15:49:05 I'll get there :P 2010-10-23T15:49:18 *** rebelxt has joined #aichallenge 2010-10-23T15:49:26 http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/compile_anything.py 2010-10-23T15:50:15 for c++ basically each file is compiled separately then one command to link 2010-10-23T15:50:15 *** mrflippy has quit IRC (Ping timeout: 265 seconds) 2010-10-23T15:50:17 *** Naktibalda_ has joined #aichallenge 2010-10-23T15:50:37 see line 157 and 158 for the commands used 2010-10-23T15:50:44 ahh. 2010-10-23T15:51:04 is there a reason it doesn't use a makefile? 2010-10-23T15:51:10 security 2010-10-23T15:51:18 ok 2010-10-23T15:51:22 so, it's just an implicit rule ? 2010-10-23T15:51:22 at least so I've been told :) 2010-10-23T15:51:25 *** Naktibalda has quit IRC (Ping timeout: 272 seconds) 2010-10-23T15:51:42 p4p4: not sure what you mean 2010-10-23T15:51:43 you could probably do some pretty wacky stuff with a makefile I guess 2010-10-23T15:52:10 yes, a makefile is basically turing complete and can run anything 2010-10-23T15:52:25 line 193 look like: take any cpp,cc,c file and make an o of it 2010-10-23T15:52:49 p4p4: yep 2010-10-23T15:53:33 then they are all linked with g++ -O2 -lm -o MyBot 2010-10-23T15:54:23 I think the go compiling code there should work fine, just needs to be installed... 2010-10-23T15:54:44 lavalamp_: yeah :/ 2010-10-23T15:55:13 earlier amstan said you might be able to do that, you'll be my best friend if you do :) 2010-10-23T15:55:22 when I run the bots on my computer the program says that the bots timed out - I follow the site - what should I do? 2010-10-23T15:56:13 askinAround: it probably crashed. feed it some test data over stdin and see what it does 2010-10-23T15:56:46 well, I let myself get distracted again with the boost thing. was hoping it would be a simple one line fix to enable it 2010-10-23T15:56:56 but why does it crash? it's the one I downloaded from the site and I didn't modify it at all 2010-10-23T15:57:17 hopefully soon though I can work on adding a few more languages, but I hate to promise anything right now 2010-10-23T15:57:23 askinAround, oh then I don't know. maybe it couldn't find the exe? 2010-10-23T15:57:24 shouldn't the site have a version that doesn't crash? 2010-10-23T15:57:40 janzert anything you can do is greatly appreciated :)) 2010-10-23T15:57:52 askinAround: if you run the bot directly does it start up 2010-10-23T15:57:59 askinAround, they don't, there's something wrong with your setup 2010-10-23T15:58:01 the starter packages are not compiled btw 2010-10-23T15:58:16 I compiled :D 2010-10-23T15:58:47 javac *.java like it says on the site 2010-10-23T15:59:02 and I have a .class file now 2010-10-23T15:59:16 *** rebelxt has left #aichallenge 2010-10-23T15:59:55 and I have this 2010-10-23T15:59:57 java -jar tools/PlayGame.jar maps/map7.txt 1000 1000 log.txt "java MyBot" "java MyBot" | java -jar tools/ShowGame.jar 2010-10-23T16:00:14 sorry it's not all 2010-10-23T16:00:34 C:\PlanetWars\java_starter_package>java -jar tools/PlayGame.jar maps/map7.txt 10 00 1000 log.txt "java MyBot" "java MyBot" | java -jar tools/ShowGame.jar WARNING: player 1 timed out. WARNING: player 2 timed out. Turn 1 Draw! 2010-10-23T16:01:29 *** JensT1 has joined #aichallenge 2010-10-23T16:01:46 *** RainCT has quit IRC (Ping timeout: 265 seconds) 2010-10-23T16:02:21 *** RainCT has joined #aichallenge 2010-10-23T16:03:26 but it runs fine when I send it on their website 2010-10-23T16:03:49 askinAround, ifyou just type "myBot" what happens? 2010-10-23T16:03:57 *MyBot 2010-10-23T16:04:27 where? what do you mean? 2010-10-23T16:04:31 wait actually 2010-10-23T16:04:38 10 00 2010-10-23T16:04:41 that's your problem 2010-10-23T16:04:44 lose the space 2010-10-23T16:04:51 oh 2010-10-23T16:05:23 there's no space in the cmd - just copy/paste problem 2010-10-23T16:05:40 what does java MyBot do? 2010-10-23T16:05:50 it should start and do nothing basically 2010-10-23T16:06:16 *** bhasker has quit IRC (Ping timeout: 265 seconds) 2010-10-23T16:06:17 if I send it on ai-challenge site it does this:http://ai-contest.com/visualizer.php?game_id=6088990 2010-10-23T16:07:00 so the code is fine, just something with your local setup needs tweaking 2010-10-23T16:08:07 *** RainCT has quit IRC (Remote host closed the connection) 2010-10-23T16:08:24 *** RainCT has joined #aichallenge 2010-10-23T16:08:39 had the same problem on windows, that is, could not start java MyBot, python MyBot.py, etc 2010-10-23T16:09:02 What does provide the go message to the bot on my machine when I run it as I said before? 2010-10-23T16:09:40 PlayGame.jar? 2010-10-23T16:11:26 Top 10 players: rsergio(3982), Hazard(3977), felixcoto(3971), bocsimacko(3934), cfaftw(3908), dmj111(3888), sin_sun(3863), davidjliu(3825), jambachili(3789), MariusCG(3770) 2010-10-23T16:12:43 *** davidjliu has joined #aichallenge 2010-10-23T16:13:01 *** bhasker has joined #aichallenge 2010-10-23T16:13:20 *** RainCT has quit IRC (Ping timeout: 240 seconds) 2010-10-23T16:14:28 *** RainCT has joined #aichallenge 2010-10-23T16:14:49 askinAround, yeah 2010-10-23T16:20:55 *** bhasker has quit IRC (Ping timeout: 240 seconds) 2010-10-23T16:25:27 *** wh1teside has quit IRC (Quit: leaving) 2010-10-23T16:25:56 thanks I'm leaving now - I'll try to play with it tomorow 2010-10-23T16:26:12 *** askinAround has quit IRC (Quit: Page closed) 2010-10-23T16:28:08 *** bhasker has joined #aichallenge 2010-10-23T16:29:20 *** Accoun has quit IRC (Ping timeout: 240 seconds) 2010-10-23T16:41:03 *** jorge_ has joined #aichallenge 2010-10-23T16:46:05 *** cfaftw has joined #aichallenge 2010-10-23T16:46:27 hey 2010-10-23T16:52:33 hi 2010-10-23T16:52:51 anyone knows if "pacotabaco" has change his nick? 2010-10-23T16:53:16 I don't see him here since many days now... 2010-10-23T16:53:42 emm... are you there? :) 2010-10-23T16:55:35 *** yasith_ has quit IRC (Ping timeout: 240 seconds) 2010-10-23T16:56:12 @last pacotabaco 2010-10-23T16:56:12 rabidus_: (last [--{from,in,on,with,without,regexp} ] [--nolimit]) -- Returns the last message matching the given criteria. --from requires a nick from whom the message came; --in requires a channel the message was sent to; --on requires a network the message was sent on; --with requires some string that had to be in the message; --regexp requires a regular expression the message must (1 more message) 2010-10-23T16:56:40 !seen pacotabaco 2010-10-23T16:56:46 @seen pacotabaco 2010-10-23T16:56:46 rabidus_: pacotabaco was last seen in #aichallenge 4 weeks, 3 days, 21 hours, 28 minutes, and 36 seconds ago: ok, tnks :) 2010-10-23T17:00:50 wow, that means this channel existed way before anyone could submit something ;) 2010-10-23T17:01:31 no, my first submission was sept 16 2010-10-23T17:02:03 k, read i wrong. 2010-10-23T17:02:16 oh 2010-10-23T17:02:19 *** Palmik has quit IRC (Remote host closed the connection) 2010-10-23T17:04:36 rabidus_: ok thanks :) 2010-10-23T17:05:55 *** Titankiller has quit IRC (Ping timeout: 240 seconds) 2010-10-23T17:08:04 *** davidjliu has quit IRC (Ping timeout: 265 seconds) 2010-10-23T17:08:12 *** Titankiller has joined #aichallenge 2010-10-23T17:09:38 *** yasith has joined #aichallenge 2010-10-23T17:10:41 *** jorge_ has quit IRC (Remote host closed the connection) 2010-10-23T17:11:27 Top 10 players: rsergio(3977), Hazard(3975), felixcoto(3965), bocsimacko(3919), cfaftw(3889), sin_sun(3849), dmj111(3836), davidjliu(3815), jambachili(3789), evja08(3776) 2010-10-23T17:12:12 *** delt0r___ has quit IRC (Ping timeout: 245 seconds) 2010-10-23T17:12:20 *** delt0r has joined #aichallenge 2010-10-23T17:12:38 *** p4p4 has quit IRC (Ping timeout: 245 seconds) 2010-10-23T17:13:55 wheee 2010-10-23T17:13:57 my bot is defending! 2010-10-23T17:14:03 :o 2010-10-23T17:14:34 yay i have a partially working bot now 2010-10-23T17:15:00 playing on TCP? 2010-10-23T17:15:09 no, on my own box. 2010-10-23T17:15:24 it has no attack code yet, it will perish pretty fast on tcp 2010-10-23T17:16:56 oh, yeah that's pretty important :P 2010-10-23T17:17:28 well, i deserve another break 2010-10-23T17:21:20 *** jkl__ has joined #aichallenge 2010-10-23T17:23:00 *** Accoun has joined #aichallenge 2010-10-23T17:24:54 *** Appleman1234 has joined #aichallenge 2010-10-23T17:28:34 *** Titankiller has quit IRC (Ping timeout: 255 seconds) 2010-10-23T17:30:41 mega1: bocsimacko is your bot? 2010-10-23T17:34:44 *** FlagCapper has joined #aichallenge 2010-10-23T17:35:07 Is anyone not able to upload their code to the official server because the upload button is missing? 2010-10-23T17:39:08 Regarding tracking future states with future moves for both players, there is simple equation based on the battle rules that tells you how many ships you must have on-planet on turn t - 1 if you want to have s free ships on turn t. 2010-10-23T17:40:58 s(t) = s(t-1) + g + f(t-1) + maxE(t-1) 2010-10-23T17:41:29 *** RainCT has quit IRC (Ping timeout: 265 seconds) 2010-10-23T17:42:02 where f(t) is your arriving fleets, and maxE(t) is the largest enemy fleet 2010-10-23T17:42:21 *** RainCT has joined #aichallenge 2010-10-23T17:43:03 that should be - maxE(t-1) not + 2010-10-23T17:44:01 you use that equation to keep track of how many free and how many reserved ships you have on each planet on each future turn 2010-10-23T17:44:19 *** Zalma has quit IRC (Ping timeout: 265 seconds) 2010-10-23T17:45:00 So if on turn 3 you send 100 ships from planet 1, you will have 10 free and 90 reserved ships on turn 1, 10 free and 95 reserved ships on turn 2, and 0 free and 0 reserved ships on turn 3. 2010-10-23T17:47:32 That should be 10 free and 0 reserved on 3. 2010-10-23T17:48:10 The problem is what happens when you make a future plan at this stage of the analysis, and then you consider an enemy counter-move that invalidates the plan. 2010-10-23T17:49:43 What you have to do is separate the planning from execution. 2010-10-23T17:50:10 So you have your current plan, and you want to add some more moves to it. 2010-10-23T17:50:22 Execute the plan, and examine the resulting game state. 2010-10-23T17:51:05 Now look for new opportunities, taking into account ships that ended up getting reserved when your plan was executed. 2010-10-23T17:52:28 Now take the enemy's side, and build a response. 2010-10-23T17:53:01 However, when you evaluate the game result, you have to start with a fresh game state 2010-10-23T17:53:36 So you have to sort all the moves by turn, and execute those which actually happen first, first 2010-10-23T17:54:13 If a move is invalid because it assumes a game state that does not hold, it cannot be executed 2010-10-23T17:57:13 I have found that minimax works well for actual combat but not so well for maintaining field position 2010-10-23T17:58:56 In other words, it finds good short-term tactical moves but does poorly with long-term strategy 2010-10-23T18:00:48 *** seisatsu has joined #aichallenge 2010-10-23T18:01:16 It's hard to model the effect of reinforcements in your game state evaluator 2010-10-23T18:01:39 It's much easier to evaluate combat 2010-10-23T18:04:14 So what I've tried is running minimax to find tactical moves, and then running a separate reinforcements heuristic to handle the leftover ships. 2010-10-23T18:04:40 *** seisatsu has quit IRC (Client Quit) 2010-10-23T18:06:28 FlagCapper: did you have/read a message where the upload button normally is? 2010-10-23T18:06:50 there's a new 10 minute delay after a successful upload now 2010-10-23T18:08:05 *** jesionaj has joined #aichallenge 2010-10-23T18:08:58 *** FlagCapper has quit IRC (Ping timeout: 265 seconds) 2010-10-23T18:11:27 Top 10 players: rsergio(3974), Hazard(3971), felixcoto(3944), bocsimacko(3915), cfaftw(3874), dmj111(3822), davidjliu(3818), sin_sun(3818), jambachili(3796), evja08(3755) 2010-10-23T18:14:26 *** narnach has joined #aichallenge 2010-10-23T18:14:56 *** TK_ has joined #aichallenge 2010-10-23T18:16:45 *** TK_ has left #aichallenge 2010-10-23T18:17:52 *** TK_ has joined #aichallenge 2010-10-23T18:19:39 *** TK_ has quit IRC (Client Quit) 2010-10-23T18:24:49 *** Azrathud has quit IRC (Ping timeout: 276 seconds) 2010-10-23T18:25:15 *** cfaftw has quit IRC (Quit: Leaving) 2010-10-23T18:26:34 *** Azrathud has joined #aichallenge 2010-10-23T18:31:27 *** seisatsu has joined #aichallenge 2010-10-23T18:33:12 *** dmj111 has quit IRC (Quit: ERC Version 5.3 (IRC client for Emacs)) 2010-10-23T18:37:05 *** amriedle has joined #aichallenge 2010-10-23T18:42:54 *** fawek has quit IRC (Remote host closed the connection) 2010-10-23T18:44:09 *** olwe has joined #aichallenge 2010-10-23T18:45:15 *** seisatsu has quit IRC (Quit: Leaving) 2010-10-23T18:45:39 how to run c++ MyBot with ShowGame.jar? 2010-10-23T18:45:58 *** yasith has quit IRC (Ping timeout: 252 seconds) 2010-10-23T18:47:44 *** choas has quit IRC (Quit: leaving) 2010-10-23T18:48:10 *** genericbob has quit IRC (Ping timeout: 252 seconds) 2010-10-23T18:50:29 should i write "cpp MyBoy" instead of "java MyBot" or what? 2010-10-23T18:51:42 the cmdline has to execute your code. 2010-10-23T18:52:32 so just type path to afile? 2010-10-23T18:54:53 *** genericbob has joined #aichallenge 2010-10-23T18:57:40 *** olwe has quit IRC (Quit: Leaving) 2010-10-23T18:58:03 *** yasith has joined #aichallenge 2010-10-23T19:06:15 *** davidjliu has joined #aichallenge 2010-10-23T19:07:18 *** jesionaj` has joined #aichallenge 2010-10-23T19:07:26 *** jesionaj has quit IRC (Read error: Connection reset by peer) 2010-10-23T19:11:26 Top 10 players: Hazard(3956), rsergio(3953), bocsimacko(3932), felixcoto(3918), cfaftw(3846), dmj111(3816), davidjliu(3813), jambachili(3797), sin_sun(3789), savchenko(3736) 2010-10-23T19:13:34 http://ai-contest.com/visualizer.php?game_id=6098593 <= andy17682 crashed / did not start / timeout and WIN 2010-10-23T19:17:43 *** hornairs has joined #aichallenge 2010-10-23T19:22:38 *** Appleman1234 has quit IRC (Excess Flood) 2010-10-23T19:23:07 *** Appleman1234 has joined #aichallenge 2010-10-23T19:27:45 *** davidjliu has quit IRC (Ping timeout: 265 seconds) 2010-10-23T19:35:16 *** RainCT has quit IRC (Remote host closed the connection) 2010-10-23T19:42:04 *** mceier has quit IRC (Quit: leaving) 2010-10-23T19:44:26 *** seisatsu has joined #aichallenge 2010-10-23T19:44:50 Accoun: that is a bug, it always shows the first player's name, but the win/loss record is correct 2010-10-23T19:50:37 *** Titankiller has joined #aichallenge 2010-10-23T19:56:19 *** yasith has quit IRC (Ping timeout: 240 seconds) 2010-10-23T20:01:08 *** wh1teside has joined #aichallenge 2010-10-23T20:11:11 *** yasith has joined #aichallenge 2010-10-23T20:11:27 Top 10 players: Hazard(3969), bocsimacko(3946), rsergio(3938), felixcoto(3930), cfaftw(3869), davidjliu(3831), dmj111(3812), jambachili(3797), sin_sun(3787), savchenko(3748) 2010-10-23T20:13:18 *** Sarin has quit IRC (Remote host closed the connection) 2010-10-23T20:17:16 *** Accoun has quit IRC (Ping timeout: 252 seconds) 2010-10-23T20:17:50 *** Accoun has joined #aichallenge 2010-10-23T20:18:23 janzert? 2010-10-23T20:19:49 jkl_? 2010-10-23T20:21:41 yes? 2010-10-23T20:23:10 the 10 minute delay - what's the purpose of that? 2010-10-23T20:23:33 in order to stop constant resubmition 2010-10-23T20:23:47 makes sense 2010-10-23T20:24:47 and if the submission fails there is no delay so on a compilation error you can fix and resubmit right away 2010-10-23T20:25:33 before it was rational to keep resubmitting until your first game was against a decent bot if you expected a high rank 2010-10-23T20:26:26 actually even that probably wouldn't have triggered putting in the delay 2010-10-23T20:26:48 someone was just constantly resubmitting sometimes multiple times per minute 2010-10-23T20:27:12 me? :P 2010-10-23T20:27:24 hehe, don't think so 2010-10-23T20:27:30 i'm not quite that bad 2010-10-23T20:28:55 *** Azrathud has quit IRC (Read error: Connection reset by peer) 2010-10-23T20:29:10 antimatroid: was this guy: http://ai-contest.com/profile_submissions.php?user_id=10464&page=1 2010-10-23T20:29:28 janzert: seems like your stuff is working 2010-10-23T20:29:45 sorta 2010-10-23T20:29:48 janzert: can you make it per user? so we can throttle him even more, lol 2010-10-23T20:29:50 jk 2010-10-23T20:29:55 hehe 2010-10-23T20:30:01 it does work though.. 2010-10-23T20:30:02 ahh, you meant the delay stuff 2010-10-23T20:30:07 yeah 2010-10-23T20:30:15 it's either the email or your limit 2010-10-23T20:30:16 my mind is on the worker instances again/still 2010-10-23T20:30:50 I've got the build script about generalized so it should work on any bare ubuntu 8.04 install 2010-10-23T20:31:44 but it lead to seeing another error with the worker tournament manager and I need to track that down 2010-10-23T20:32:16 s/lead/led 2010-10-23T20:32:30 janzert: neat 2010-10-23T20:32:58 bah, trying to fix one bug I keep running into more 2010-10-23T20:34:32 *** rommel has joined #aichallenge 2010-10-23T20:36:40 Hey fellas, if anyones interested : https://wave.google.com/wave/waveref/googlewave.com/w+NkB4ALvPA -- Could be a better way to share ideas :) 2010-10-23T20:36:51 not in the stone age anymore lol 2010-10-23T20:37:22 I want to help with issue 189, which addresses the problem of incorrect error messages. Can anyone explain how the error table gets populated? I see api_record_game.php, but what uses it? 2010-10-23T20:37:36 hmm, I thought wave was shutting down end of september. guess I got the wrong date 2010-10-23T20:37:59 the cloud tournament_manager.py 2010-10-23T20:38:13 found in the 2010....cloud... branch :) 2010-10-23T20:38:30 games_in_the_cloud 2010-10-23T20:38:38 bbs .. stupid work haha 2010-10-23T20:39:50 OK, I see 20100929-games-in-the-cloud 2010-10-23T20:40:03 in the Web interface 2010-10-23T20:40:04 yep look in the backend code in there 2010-10-23T20:40:13 thanks 2010-10-23T20:40:18 that is the code running on the worker instances 2010-10-23T20:40:28 trunk is running on the main server 2010-10-23T20:41:00 it really should be merged back together, but that's something to do later :/ 2010-10-23T20:43:09 *** rommel has quit IRC (Ping timeout: 265 seconds) 2010-10-23T20:56:32 OK, so you have a try/except pair around engine.play_game, but all the except code does is some logging, it sleeps (?), and then it falls through to the code that handles the game result. Shouldn't the game result be abandoned? 2010-10-23T20:58:19 not super familiar with that code, let me look at it and see if anything different jumps out to me 2010-10-23T20:59:12 hmm, that does look pretty suspicious 2010-10-23T21:00:12 I don't think that is what is causing 189, but it may cause the tournament manager to crash 2010-10-23T21:00:30 yeah, I was just going to say it should cause the manager to crash 2010-10-23T21:00:32 outcome will be None 2010-10-23T21:00:51 actually worse than that 2010-10-23T21:00:58 read before assignment isn't it? 2010-10-23T21:01:07 either way crash 2010-10-23T21:01:25 You're probably right 2010-10-23T21:01:29 outcome will just be undefined 2010-10-23T21:01:37 an explicit sys.exit would be better though 2010-10-23T21:01:52 or probably can even just continue 2010-10-23T21:02:16 I'll submit a patch 2010-10-23T21:02:22 after I figure out 189 2010-10-23T21:02:46 *** rebelxt_ has joined #aichallenge 2010-10-23T21:02:56 *** Azrathud has joined #aichallenge 2010-10-23T21:03:01 I'll change the sleep to sys.exit unless you see something that is definitely better 2010-10-23T21:05:57 I don't know right now. I'm still trying to understand what is going on. 2010-10-23T21:06:28 mega1 is just killing us on the tcp server. 94% win 469 games wow 2010-10-23T21:06:50 ok, in all my test running I've never seen the tournament_manager crash with outcome undefined so if that code is getting triggered it's very rare 2010-10-23T21:07:27 why catch an error condition you can't handle? 2010-10-23T21:07:52 If the program is just going to fail, let it fail 2010-10-23T21:08:07 yeah, I was just wondering that as well unless he was logging sys.stdout at the time or something and wanted the traceback 2010-10-23T21:08:24 But wouldn't you get that anyway? 2010-10-23T21:08:25 but it does look almost like debugging code that got left 2010-10-23T21:08:36 not if he was throwing away stderr 2010-10-23T21:08:58 since many of the bots print tons of trash to stderr 2010-10-23T21:09:18 OK, that makes sense 2010-10-23T21:09:23 at this point though the way it is set up both stdout and stderr are just fed to /dev/null 2010-10-23T21:10:05 tournament_manager_runner.sh is used to run tournament_manager.py in a loop 2010-10-23T21:11:27 Top 10 players: Hazard(3933), bocsimacko(3930), rsergio(3913), felixcoto(3912), cfaftw(3861), davidjliu(3829), dmj111(3796), sin_sun(3777), jambachili(3772), luksian(3737) 2010-10-23T21:11:31 but yeah I think just getting rid of the whole try except clause at this point is probably best 2010-10-23T21:16:07 *** Mathnerd314 has quit IRC (Ping timeout: 252 seconds) 2010-10-23T21:17:44 er, that still needs to be fixed 2010-10-23T21:18:00 Zannick: what's that? 2010-10-23T21:18:07 the user_sadbox is where stdout and stderr are read, but stderr is not actually read 2010-10-23T21:18:42 *** JensT1 has quit IRC (Quit: Leaving.) 2010-10-23T21:19:02 hmm, when I'm running tournament_manager.py directly on worker instances I'm definitely seeing the bot stderr output on the console 2010-10-23T21:19:14 haven't actually looked at what the code is doing 2010-10-23T21:20:33 At one point I saw one go by where somebody apparently was a bit frustrated and printed an endless loop of swear words until they timed out. :P 2010-10-23T21:20:46 Unrelated to 189, but related to 202: how many virtual CPUs do the cloud instances have? 2010-10-23T21:21:06 2 for the amazon ec2 instances 2010-10-23T21:21:12 that are being used 2010-10-23T21:21:29 and I've yet to see the load average over 1.5 2010-10-23T21:22:50 currently server #56 is 0.64 2010-10-23T21:23:05 heh 2010-10-23T21:23:15 i guess something is reading it then 2010-10-23T21:26:20 afk for a while 2010-10-23T21:35:33 It's obvious 2010-10-23T21:35:53 trunk/planet_wars/www/game_info.php 2010-10-23T21:36:01 get_error_message() 2010-10-23T21:36:13 you are supposed to be reading all errors into $out 2010-10-23T21:36:23 there could be multiple errors in one game 2010-10-23T21:36:29 but what happens is 2010-10-23T21:36:38 $out[]=$result['username'].' '.$nice_error."."; 2010-10-23T21:37:25 never mind 2010-10-23T21:38:05 I misunderstood the PHP syntax 2010-10-23T21:41:20 *** davidjliu has joined #aichallenge 2010-10-23T21:43:15 how does my program work in the server ? 2010-10-23T21:43:35 *** adg has left #aichallenge 2010-10-23T21:43:39 is it like it's called after each move ? 2010-10-23T21:44:01 or does it stay in memory till the game's over ? 2010-10-23T21:45:08 Your bot and the server exchange data over your bot's stdin and stdout 2010-10-23T21:46:03 Your bot runs until the game ends. 2010-10-23T21:46:41 hmmmm 2010-10-23T21:46:56 so in java my constructor only runs once 2010-10-23T21:47:50 If you want to understand what the bot does, read PlanetWars.java. 2010-10-23T21:48:28 That's what handles I/O 2010-10-23T21:48:59 It contains the main method 2010-10-23T21:53:48 janzert: thanks a lot for your work getting EC2 servers going! 2010-10-23T22:00:42 *** Azrathud has quit IRC (Ping timeout: 240 seconds) 2010-10-23T22:00:43 *** Azrathud| has joined #aichallenge 2010-10-23T22:02:37 *** Titankiller has quit IRC (Ping timeout: 240 seconds) 2010-10-23T22:07:16 *** yasith has quit IRC (Ping timeout: 252 seconds) 2010-10-23T22:11:28 Top 10 players: Hazard(3935), bocsimacko(3933), rsergio(3915), felixcoto(3914), cfaftw(3864), davidjliu(3831), dmj111(3800), sin_sun(3780), jambachili(3773), savchenko(3737) 2010-10-23T22:12:29 I wish I could just make them stable 2010-10-23T22:12:31 fu 2010-10-23T22:15:08 Someone on the forums noticed that worker=55 is associated with game crashes. Is it possible there is something different about it? 2010-10-23T22:15:37 hmm, no there shouldn't be 2010-10-23T22:16:24 it just locked up and is in the process of rebooting now so I'll double check that it's running the svn code after it comes back up 2010-10-23T22:16:48 It locked up? 2010-10-23T22:17:07 they all do, that's the problem with keeping them up :/ 2010-10-23T22:17:22 they only stay up for about 6 hours at most 2010-10-23T22:17:37 it's a known bug with ubuntu hardy on ec2 :( 2010-10-23T22:17:52 What about Sarge? 2010-10-23T22:18:00 Oh I mean Lenny 2010-10-23T22:18:13 we're pretty much stuck with hardy for the moment 2010-10-23T22:18:20 10.04 does fix it yeah 2010-10-23T22:18:50 problem is that we need the same version running on the main server and the workers so the same language packages are being used 2010-10-23T22:18:56 why on earth do some maps have +0 planets on them?? 2010-10-23T22:19:05 just to bug you :) 2010-10-23T22:19:18 to make sure bots handle it correctly? 2010-10-23T22:19:28 I think I might write code to completely ignore them 2010-10-23T22:20:25 just make sure you take it over at the end of the game if it's the only one your opponent has left 2010-10-23T22:20:39 I remember others had that problem before 2010-10-23T22:21:07 I think some maps it works as a good staging planet as well 2010-10-23T22:22:43 *** yasith has joined #aichallenge 2010-10-23T22:25:14 I'm trying to see if my efforts raised my bot's -4xx Elo... first game it matches me with bocsimacko >.< 2010-10-23T22:27:26 *** Titankiller has joined #aichallenge 2010-10-23T22:40:34 Oh shit 2010-10-23T22:40:40 players = [ {"path" : player_one_path, "command" : player_one["command"], "submission_id": player_one["submission_id"]}, {"path" : player_two_path, "command" : player_two["command"], "submission_id": player_one["submission_id"]} ] 2010-10-23T22:40:52 player 2 gets player 1's submission id 2010-10-23T22:40:55 it's right there 2010-10-23T22:41:09 copy n paste error 2010-10-23T22:41:18 there's your bug 189 2010-10-23T22:41:21 where? 2010-10-23T22:41:38 line 121 2010-10-23T22:41:46 look at the players array 2010-10-23T22:41:49 tournament_manager? 2010-10-23T22:41:53 yes 2010-10-23T22:41:56 on the cloud branch 2010-10-23T22:42:10 {"path" : player_two_path, "command" : player_two["command"], "submission_id": player_one["submission_id"]} 2010-10-23T22:42:15 ouch 2010-10-23T22:42:19 player two's path, player 1's id 2010-10-23T22:42:38 serious bummer 2010-10-23T22:42:54 checking trunk 2010-10-23T22:42:57 so is it used for anything other than error reporting? 2010-10-23T22:43:06 I don't know 2010-10-23T22:43:16 there are i's and i+1's in a lot of places 2010-10-23T22:43:17 ok, looking now 2010-10-23T22:43:30 and you also have player_one and player_two 2010-10-23T22:43:38 lots of different ways to refer to the players 2010-10-23T22:43:54 *** narnach has quit IRC (Ping timeout: 240 seconds) 2010-10-23T22:44:46 trunk doesn't have that code 2010-10-23T22:44:59 it doesn't do any error reporting to the DB 2010-10-23T22:45:19 luckily the submission id there seems to only be used for the error reporting 2010-10-23T22:45:35 yeah, that's why worker 0 never has a error shown in the visualizer 2010-10-23T22:45:55 yay, thanks for finding it 2010-10-23T22:48:24 One down, two to go. 2010-10-23T22:48:57 you're jklandis on the tracker I take it? 2010-10-23T22:49:17 *** yasith has quit IRC (Ping timeout: 240 seconds) 2010-10-23T22:49:32 *** narnach has joined #aichallenge 2010-10-23T22:50:11 hope so since that's who I attributed it to in the commit message :) 2010-10-23T22:50:49 Yes 2010-10-23T22:51:05 I wrote a note about the fix on the bug tracker. 2010-10-23T22:51:30 Can you close it? 2010-10-23T22:51:50 yes, I will just getting the server updated 2010-10-23T22:53:17 Great. I wrote a note on 199 about what you just said about reporting. That explains why some games just end. 2010-10-23T22:55:49 yep 2010-10-23T22:56:06 I didn't realize 199 was specifically about worker 0 before 2010-10-23T22:56:31 I just noticed that myself. 2010-10-23T22:56:56 I'm satisfied at least knowing that it is expected behavior 2010-10-23T22:57:12 Even if it isn't perfect behavior 2010-10-23T22:57:25 yeah, it really does need to get fixed soon 2010-10-23T22:57:43 just wish there were a few more people active in the code :) 2010-10-23T22:57:49 thanks for taking a look at it 2010-10-23T22:58:12 So let's see 2010-10-23T22:58:18 how hard it would be 2010-10-23T22:58:28 *** Titankiller has quit IRC (Ping timeout: 276 seconds) 2010-10-23T22:58:54 I guess you would have to merge engine.py, since that is what actually tracks the errors 2010-10-23T22:59:04 Then you have to get the data into the db 2010-10-23T22:59:25 right and the main server tournament manager deals directly with the db of course 2010-10-23T23:00:16 *** iFire has quit IRC (Ping timeout: 240 seconds) 2010-10-23T23:00:27 So it's some Python and SQL work 2010-10-23T23:02:10 right 2010-10-23T23:04:56 *** yasith has joined #aichallenge 2010-10-23T23:04:56 I'm on it 2010-10-23T23:05:12 great 2010-10-23T23:06:07 *** iFire has joined #aichallenge 2010-10-23T23:06:31 The plan is to port just the error stuff from the cloud engine, and then copy the error-reporting SQL from the PHP submission relay into the trunk tournament manager. 2010-10-23T23:07:31 I haven't looked at the changes made to the cloud engine, but can it be brought back over whole? 2010-10-23T23:07:59 i.e. so trunk and the cloud branch will be back to the same engine.py? 2010-10-23T23:08:31 Maybe. Trunk is doing something with selectable I/O 2010-10-23T23:08:44 oh, yeah I remember that now. 2010-10-23T23:09:05 I saw that when I was bringing svn back into sync with the server 2010-10-23T23:09:10 I don't understand what is going on. There are threads, non-blocking reads, and now select 2010-10-23T23:09:38 I really think that is just a failed experiment actually but it is what was in use so I decided just to stick it into svn for now at least 2010-10-23T23:09:54 *** iNtERrUpT has joined #aichallenge 2010-10-23T23:10:00 argh argh argh 2010-10-23T23:10:11 *** rebelxt_ has quit IRC (Quit: Page closed) 2010-10-23T23:10:22 I'm still convinced that a scheduling issue is at the heart of 202, and overzealous polling may be part of the problem. But one bug at a time... 2010-10-23T23:11:14 I really want to rewrite it leave the threads and blocking reads in place and just have the i/o threads signal the main thread with an event 2010-10-23T23:11:28 seems like the easiest way to clean it up from where we are now 2010-10-23T23:11:29 Top 10 players: Hazard(3949), rsergio(3932), felixcoto(3927), bocsimacko(3916), cfaftw(3883), davidjliu(3858), dmj111(3811), sin_sun(3802), jambachili(3775), savchenko(3745) 2010-10-23T23:11:51 but yeah for now probably best just to merge the error reporting back in 2010-10-23T23:12:05 I was just about to change my mind. 2010-10-23T23:12:10 another case of trying to fix one bug and butting into another 2010-10-23T23:12:26 If the cloud and trunk run exactly the same code, it will be much easier to track down problems 2010-10-23T23:12:35 In the diff I see the error code and the selector code 2010-10-23T23:12:36 whichever you feel up to then 2010-10-23T23:12:59 also, outcome["engine_stats 2010-10-23T23:13:02 "] 2010-10-23T23:13:05 I'm going to try and finish up a fix in the worker submission download code 2010-10-23T23:13:39 OK 2010-10-23T23:14:00 then go back to figuring out why workers are sometimes requesting submissions that no longer exist :P 2010-10-23T23:14:26 I'm not sure how many layers of bugs deep I'm working on any more :( 2010-10-23T23:14:46 You're 3 levels deep. You could spend your whole life down there. 2010-10-23T23:14:57 Here's the kick 2010-10-23T23:24:17 *** rebelxt_ has joined #aichallenge 2010-10-23T23:25:18 threads? i thought threads weren't allowed? 2010-10-23T23:25:25 *** Azrathud| has quit IRC (Quit: Meh) 2010-10-23T23:32:56 Threads in the game server engine 2010-10-23T23:33:11 You have two bots running at the same time, and you have to multiplex I/O somehow 2010-10-23T23:33:24 one approach is to use blocking I/O in threads 2010-10-23T23:33:43 Another approach is to poll using non-blocking reads 2010-10-23T23:34:37 Right now the code does... both. 2010-10-23T23:36:14 ok, one more pain point in setting up test servers assuaged. the countries and user_status_code database info is now in svn 2010-10-23T23:36:40 I have the engine patch ready 2010-10-23T23:36:53 still working on the SQL / tournament manager 2010-10-23T23:38:10 *** Titankiller has joined #aichallenge 2010-10-23T23:38:43 ok 2010-10-23T23:39:20 *** amstan_ has joined #aichallenge 2010-10-23T23:39:20 *** ChanServ sets mode: +o amstan_ 2010-10-23T23:40:36 *** amstan__ has joined #aichallenge 2010-10-23T23:40:36 *** ChanServ sets mode: +o amstan__ 2010-10-23T23:43:02 *** amstan has quit IRC (Ping timeout: 245 seconds) 2010-10-23T23:44:03 *** amstan_ has quit IRC (Ping timeout: 252 seconds) 2010-10-23T23:46:59 *** amstan__ is now known as amstan 2010-10-23T23:48:13 *** iNtERrUpT has quit IRC (Remote host closed the connection) 2010-10-23T23:51:40 *** hellman has joined #aichallenge 2010-10-23T23:55:49 *** McLeopold has joined #aichallenge 2010-10-23T23:59:37 *** yasith has quit IRC (Ping timeout: 255 seconds)