2011-06-09T00:46:24 *** delt0r_ has joined #aichallenge 2011-06-09T00:47:45 *** delt0r- has quit IRC (Ping timeout: 250 seconds) 2011-06-09T01:48:50 *** stocha has joined #aichallenge 2011-06-09T01:50:33 @later tell janzert it would be interesting to have an official test server, where multiple submissions are possible. For testing ideas and features against a variety of opponents. I think after all, it is better than the tcp thing for real condition testing. 2011-06-09T01:50:33 stocha: Aye, aye, sir 2011-06-09T01:51:35 well, how do you debug your code? 2011-06-09T01:54:15 *** Mathnerd314 has quit IRC (Ping timeout: 240 seconds) 2011-06-09T02:00:37 *** Palmik has joined #aichallenge 2011-06-09T02:04:11 rabidus: the usual way i guess. 2011-06-09T02:04:56 you spend 10 days building each test case, only to find out after a minor modification, no test case make sense anymore. usual stuff. 2011-06-09T02:05:53 the other strategy is to build your own implementation of the game server and graphical view, and test on those. (i did just that) 2011-06-09T02:07:14 who has the knowledge and access to the test serveur, and can help with all thoses bot that do not load properly ? 2011-06-09T02:14:00 *** stocha has quit IRC (Ping timeout: 252 seconds) 2011-06-09T02:32:31 there isn't some way or option to set the server to be more verbose ? 2011-06-09T02:40:55 *** xQuasar has joined #aichallenge 2011-06-09T02:45:08 *** Palmik has quit IRC (Remote host closed the connection) 2011-06-09T02:47:34 *** Palmik has joined #aichallenge 2011-06-09T02:48:05 *** Quasar has joined #aichallenge 2011-06-09T02:48:05 *** xQuasar has quit IRC (Ping timeout: 260 seconds) 2011-06-09T02:48:31 *** Quasar is now known as Guest8211 2011-06-09T02:52:27 I keep crashing on turn 0, what's wrong? 2011-06-09T02:52:30 Test Error: compiled, but failed test cases - Java 2011-06-09T02:52:32 turn 0 bot 0 crashed 2011-06-09T02:52:34 When I test it locally, there are no errors 2011-06-09T02:52:39 *** Guest8211 is now known as xQuasar 2011-06-09T02:52:49 *** Ice_Harley has joined #aichallenge 2011-06-09T02:54:48 xQuasar: are you testing it with the test_bot script locally? 2011-06-09T02:55:16 that should be trying to do the same test as the server 2011-06-09T02:55:51 did you tested it locally with the right java version ? 2011-06-09T02:57:58 dr-: What java version is being used? 2011-06-09T02:58:13 janzert: Do I have to get that off the git? I'm using the tools.zip from the test site 2011-06-09T02:58:14 openjdk 1.6 2011-06-09T02:58:33 dr-: i'm using sunJDK 1.6 2011-06-09T02:58:41 well, oracle's JDK 1.6 now 2011-06-09T02:58:51 dunno if it could do something wrong 2011-06-09T02:59:04 should be included in the tools.zip but I'm not positive 2011-06-09T02:59:08 dr-: I think they're mostly compatible - I don't think my code has any incompatible bits 2011-06-09T02:59:15 between oracle jdk and open jdk 2011-06-09T03:00:34 ok, where is the tools.zip on the site linked from? :) 2011-06-09T03:00:43 *** amstan has quit IRC (Remote host closed the connection) 2011-06-09T03:02:16 it was on the "Using the tools" webpage, but it's empty right now 2011-06-09T03:02:31 ok just got it directly. yes, the tools.zip from the site does include the test_bot scripts 2011-06-09T03:03:46 *** xQuasar has quit IRC (Ping timeout: 258 seconds) 2011-06-09T03:04:57 *** antimatroid has joined #aichallenge 2011-06-09T03:07:05 *** xQuasar has joined #aichallenge 2011-06-09T03:13:37 *** xQuasar has quit IRC (Ping timeout: 244 seconds) 2011-06-09T03:13:48 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-06-09T03:19:19 *** Eruonen has joined #aichallenge 2011-06-09T03:19:42 *** aerique has joined #aichallenge 2011-06-09T03:21:38 *** xQuasar has joined #aichallenge 2011-06-09T03:26:01 *** xQuasar has quit IRC (Ping timeout: 244 seconds) 2011-06-09T03:26:48 *** xQuasar has joined #aichallenge 2011-06-09T03:29:23 3rd on current ranking ... like it would not occur again, let's drink champagne 2011-06-09T03:29:49 *** smellyhippy has joined #aichallenge 2011-06-09T03:29:50 *** smellyhippy has joined #aichallenge 2011-06-09T03:30:58 dr-: :) 2011-06-09T03:35:54 *** xQuasar has quit IRC (Ping timeout: 255 seconds) 2011-06-09T03:37:33 *** xQuasar has joined #aichallenge 2011-06-09T03:38:24 antimatroid can you help me with my problem? my bot keeps crashing on turn 1 with no other information given (not jvm initiation error) 2011-06-09T03:38:40 on the main server or locally too? 2011-06-09T03:38:52 last time i uploaded my prelim bot to beta it wouldn't even pass the test 2011-06-09T03:38:57 i haven't looked since then 2011-06-09T03:39:38 It works locally, crashes turn 1 on server 2011-06-09T03:39:43 turn 0* 2011-06-09T03:42:58 not sure then sorry 2011-06-09T03:43:04 was is working previously? 2011-06-09T03:43:12 i've had troubles in the past with the server and my bots 2011-06-09T03:43:34 xQuasar: did you see that the test_bot scripts are included in tools.zip? 2011-06-09T03:43:35 i usually fix it by either going backwards and reprogramming stuff or just messing around till it works :\ so not much help :P 2011-06-09T03:44:00 *** xQuasar has quit IRC (Ping timeout: 255 seconds) 2011-06-09T03:44:28 heh, everytime I try and tell him he's already in the midst of a timeout :/ 2011-06-09T03:49:07 :P 2011-06-09T03:49:39 *** qacek has quit IRC (Quit: WeeChat 0.3.5) 2011-06-09T03:51:07 aichallenge: janzert epsilon * r91b8042 / worker/jailguard.py : New secure jail signaling and communication method with a script running inside the jail - http://bit.ly/lwrlz7 2011-06-09T03:51:07 aichallenge: janzert epsilon * r38ed0ad / (3 files): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://bit.ly/kZVshV 2011-06-09T03:57:05 aichallenge: aerique epsilon * r28c98f0 / ants/dist/starter_bots/common_lisp/README.md : Add note about redirecting *error-output*. - http://bit.ly/jGvL7R 2011-06-09T03:58:13 *** xQuasar has joined #aichallenge 2011-06-09T03:58:19 *** Eruonen has quit IRC () 2011-06-09T03:59:20 aerique: there isn't a simple way to disable warnings? 2011-06-09T03:59:52 simply redirecting stderr is going to hide genuine errors as well 2011-06-09T04:03:18 janzert: yeah, you're right. 2011-06-09T04:03:38 i kinda copied the blunt method we used for planet wars 2011-06-09T04:04:26 i'm always fighting with disabling the right warnings in sbcl, but i guess i'll have to dive into it 2011-06-09T04:04:26 and maybe it really is the best option, it'd be nice if it wasn't though ;) 2011-06-09T04:05:06 is there a simple command line switch to disable all of them? 2011-06-09T04:05:12 no :( 2011-06-09T04:05:20 too bad :( 2011-06-09T04:05:57 i'm not actually running into it with the starter bot, i added the note at mcstar's request. i think he build his own package 2011-06-09T04:06:15 * janzert nods 2011-06-09T04:06:17 i'm using a build system which avoids most of the problems 2011-06-09T04:07:17 i keep the note for now. i'll start testing my own bot soon and see what i'll run into and fix it properly 2011-06-09T04:07:33 maybe include in the note the downside and mention that if the build fails it would probably be best to resubmit with the redirection disabled so there is the possibility of getting a helpful error message 2011-06-09T04:08:02 will do 2011-06-09T04:09:24 *** Naktibalda has joined #aichallenge 2011-06-09T04:10:38 *** Quasar has joined #aichallenge 2011-06-09T04:11:04 *** Quasar is now known as Guest8983 2011-06-09T04:11:13 contestbot: later tell stocha a problem with an official test server is that necessarily any resources spent on it are resources not spent on the main server and we haven't had a situation yet where too many games were happening on the main server 2011-06-09T04:11:13 janzert: Ready to serve. 2011-06-09T04:12:29 *** xQuasar has quit IRC (Disconnected by services) 2011-06-09T04:12:35 *** Guest8983 is now known as xQuasar 2011-06-09T04:13:08 Who should I contact to ask about the errors behind my bot continuously crashing on turn 1 on the beta server? 2011-06-09T04:13:13 turn 0* 2011-06-09T04:13:59 It only says turn "0 bot 0 crashed" - no other information. When I test locally, my bot works fine 2011-06-09T04:14:09 "turn 0 bot 0 crashed" ** 2011-06-09T04:17:25 the tools page (http://aichallengebeta.hypertriangle.com/using_the_tools.php) is empty. is this intentional? 2011-06-09T04:22:51 janzert: quick! :P 2011-06-09T04:23:07 aichallenge: janzert epsilon * r02a7c4f / (setup/worker_setup.py worker/sandbox.py): Correct version of new sandbox code and update worker_setup to install jailguard - http://bit.ly/iLTSVX 2011-06-09T04:23:15 ? 2011-06-09T04:23:27 antimatroid: yes? 2011-06-09T04:23:51 (17:43:34) janzert: xQuasar: did you see that the test_bot scripts are included in tools.zip? 2011-06-09T04:23:58 ahh 2011-06-09T04:24:20 aichallenge: aerique epsilon * r9ca282d / ants/dist/starter_bots/common_lisp/README.md : Add warning about redirecting *error-output*. - http://bit.ly/jDFVwP 2011-06-09T04:30:09 Oh - is the problem that I'm redirection SYstem.err ? 2011-06-09T04:30:55 xQuasar: this is for the lisp package but are you doing that? 2011-06-09T04:31:25 because i just talked this through with janzert and you will not get useful error output on your profile page if you are :) 2011-06-09T04:31:41 *** locutus2 has joined #aichallenge 2011-06-09T04:32:05 aerique: I'm redirection System.err to a log file, I'm using java 2011-06-09T04:32:27 I guess I'll just have to make a switch for it when I'm submitting to the contest site then :) 2011-06-09T04:32:34 xQuasar: yup :) 2011-06-09T04:33:03 having it automated will save you a lot of trouble 2011-06-09T04:33:05 I only redirected system.err because there's no way to view errors when testing locally :o 2011-06-09T04:33:18 aerique: what do you mean automated? 2011-06-09T04:33:58 Runtime error: java.lang.NumberFormatException: For input string: "3738924287407908045" 2011-06-09T04:34:01 ^ What the hell??? 2011-06-09T04:34:13 Why is there such a freakin massive parameter?? 2011-06-09T04:36:43 Oh, it's a test case. Why is there such a retarded test case? 2011-06-09T04:39:21 don't know how you'd get that from the server test case, the test map is only 17 by 19 2011-06-09T04:39:50 as I was trying to say earlier though you can run the server test case locally with the test_bot scripts 2011-06-09T04:39:57 xQuasar: i mean automatically redirecting stderr when you're testing locally instead of setting a flag by hand (since you're bound to forget to do that now and then and it also takes extra time) 2011-06-09T04:43:55 xQuasar, just add an option, a flag, like --local, which won't be put on the server 2011-06-09T04:44:00 *** boegel has joined #aichallenge 2011-06-09T04:44:48 oh ok. 2011-06-09T04:44:58 Anyways, my other question...? 2011-06-09T04:45:18 Why is there such a large number in the test case? It's stupid to have something like that because nothing in game will get anywhere near that 2011-06-09T04:48:49 *** vall has joined #aichallenge 2011-06-09T04:51:22 *** akranis has joined #aichallenge 2011-06-09T04:52:52 xQuasar: you don't know, maybe the goal of this contest is to make a real simulation of ants' colonies and such numbers could be possible :p 2011-06-09T04:53:58 dr-: That's just dumb. == 2011-06-09T04:55:20 *** vall has quit IRC (Ping timeout: 252 seconds) 2011-06-09T04:59:45 *** Eruonen has joined #aichallenge 2011-06-09T05:01:42 *** mceier has joined #aichallenge 2011-06-09T05:12:26 *** guyverthree has joined #aichallenge 2011-06-09T05:12:29 good morning all 2011-06-09T05:28:33 *** berak has joined #aichallenge 2011-06-09T05:32:51 *** stocha has joined #aichallenge 2011-06-09T05:49:32 *** guyverthree has quit IRC (Quit: Page closed) 2011-06-09T05:51:15 *** stocha has quit IRC (Quit: Page closed) 2011-06-09T05:53:11 *** antimatroid has quit IRC (Ping timeout: 268 seconds) 2011-06-09T05:56:50 *** Eruonen has quit IRC () 2011-06-09T06:07:00 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-09T06:07:46 *** sigh has joined #aichallenge 2011-06-09T06:09:56 *** antimatroid has joined #aichallenge 2011-06-09T06:18:15 *** Frontier has joined #aichallenge 2011-06-09T06:32:48 *** FireFly has joined #aichallenge 2011-06-09T06:35:44 *** sir_macelon has joined #aichallenge 2011-06-09T06:35:50 hi 2011-06-09T06:35:56 evening 2011-06-09T06:36:33 i am not happy with removing the 'r'... 2011-06-09T06:37:07 now ants will tend to forget about food when they lose vision over it? 2011-06-09T06:38:13 this introduces some new shitty calculations to keep that information... 2011-06-09T06:40:06 i will have to update the java starter bot as well 2011-06-09T06:41:12 sir_macelon: i made these arguments, apparently it's not that much work :\ 2011-06-09T06:41:18 the r is easy to compute for the bot, and it was really confusing to explain to people 2011-06-09T06:41:35 the second point was the main thing 2011-06-09T06:41:50 just make the starter bots forget food information once they go out of view though 2011-06-09T06:41:52 *** locutus2 has joined #aichallenge 2011-06-09T06:41:54 let the programmer deal with the messyness 2011-06-09T06:48:17 *** antimatroid1 has joined #aichallenge 2011-06-09T06:50:19 *** antimatroid has quit IRC (Ping timeout: 258 seconds) 2011-06-09T06:50:58 hopefully i can keep the food which i don't see 2011-06-09T06:51:25 but can i rely on that 'a' lines are always send before 'f' lines? 2011-06-09T06:51:38 *sent 2011-06-09T06:52:37 what do you mean? 2011-06-09T06:53:45 i can calulate the vision at the time of parsing then 2011-06-09T06:54:56 when i read the information about all ants then i know which of them doesn't see the food and then keep it 2011-06-09T06:55:12 before starting processing the food information 2011-06-09T06:55:15 oh, i wouldn't count on it 2011-06-09T06:55:23 read all data before doing anything 2011-06-09T06:55:41 it's messy to do well 2011-06-09T06:56:09 i didn't like moving to the format partially because of that reason, came up with the 'r' lines to avoid it and lost the argument for keeping them :P 2011-06-09T06:57:38 i would have kept it... 2011-06-09T06:57:44 same :P 2011-06-09T06:58:01 but, it's not the worst thing that could have happened 2011-06-09T06:58:10 once you've handled it, it's done 2011-06-09T06:58:38 yeah, well :) 2011-06-09T06:59:07 wanted to focus on more fancy things :) 2011-06-09T06:59:15 So this weekend we are finally starting the contest? 2011-06-09T06:59:22 we are? 2011-06-09T06:59:30 are we? 2011-06-09T06:59:31 crap, i need to do something about the c++ timer 2011-06-09T06:59:40 bloody windows 2011-06-09T07:00:07 bloody non mingw using windows users* 2011-06-09T07:01:34 Well even under Linux & MacOS there are more accurate functions for timing now that aren't compatible with each other 2011-06-09T07:02:29 i'm tempted to just take the timer out and tell individual users to deal with it 2011-06-09T07:02:45 i could even just a couple of timers in the forums, one for mac/linux another for windows 2011-06-09T07:04:31 lol, would it be terrible if i defined a different Timer structure depending on whether windows32 was defined? 2011-06-09T07:04:42 and have each one with the same functions ? 2011-06-09T07:08:57 bloody c++ :) 2011-06-09T07:10:31 it's all fun and games until you worry about operator precedence, templates and cross platformability :P 2011-06-09T07:10:42 antimatroid1: That would not be terrible at all. That's what I would expect anyway 2011-06-09T07:11:04 alright, i have a new task for the night 2011-06-09T07:16:49 antimatroid1, please throw out the timer. and the buglog, too, since the engine is already logging stderr perfectly 2011-06-09T07:17:15 berak: it's easy to throw out yourself 2011-06-09T07:17:22 it's there to make life easier for people who probably couldn't write it themselves 2011-06-09T07:17:45 *** olexs has joined #aichallenge 2011-06-09T07:18:05 already done that ;) it gets very annoying, if 5 c++ bots try to use the same log file 2011-06-09T07:21:48 sorry, did not want to sound that harsh. 2011-06-09T07:22:28 s'all good 2011-06-09T07:22:32 it is annoying when they use the same file 2011-06-09T07:22:39 it was annoying in planet wars with just 2 of them :P 2011-06-09T07:30:36 antimatroid1: you posted some c++ game library to the forum, did you ever try to hook that up to a general game player ? 2011-06-09T07:31:35 berak: it's possible to use a State with comparison operations if one wanted, but it'd be much more efficient to rewrite your own specific game object 2011-06-09T07:31:44 it's more for general game theory stuff 2011-06-09T07:32:47 k, i see 2011-06-09T07:34:16 *** george1914 has joined #aichallenge 2011-06-09T07:34:55 berak: it's gone to be pretty hard to get anything out of game structures with the number of ants in this 2011-06-09T07:35:08 are you familiar with game theory? 2011-06-09T07:35:49 no, i just got interested during pw, and started collecting stuff 2011-06-09T07:37:58 remember the big black hole, after pw finished ? ;) 2011-06-09T07:38:43 Wow there must be 100 parodies on different topics over a scene from the movie "Der Untergang" on YouTube. I whish I couldn't understand what they talk and just rely on the sub-titles. 2011-06-09T07:39:05 ah okay, well feel free to play with my game theory library some time anyway :P 2011-06-09T07:39:22 i need to upload the most recent stuff, i added back in two templates for game structures 2011-06-09T07:41:10 bbl, restarting in windows 2011-06-09T07:41:23 crap, i don't think i even have codeblocks installed :\ 2011-06-09T07:42:10 *** antimatroid1 has quit IRC (Quit: Leaving.) 2011-06-09T07:48:41 *** locutus2 has quit IRC (Ping timeout: 240 seconds) 2011-06-09T08:12:17 *** smiley1983 has joined #aichallenge 2011-06-09T08:25:02 *** antimatroid has joined #aichallenge 2011-06-09T08:28:10 fixed, just going to the shop before they close, then i'll upload it 2011-06-09T08:40:07 *** locutus2 has joined #aichallenge 2011-06-09T08:46:22 *** delt0r- has joined #aichallenge 2011-06-09T08:47:21 *** delt0r_ has quit IRC (Read error: Operation timed out) 2011-06-09T08:48:52 *** xQuasar has quit IRC () 2011-06-09T08:55:12 *** delt0r- is now known as delt0r_ 2011-06-09T08:57:30 *** olexs has quit IRC (Quit: Leaving.) 2011-06-09T08:57:55 *** delt0r_ has joined #aichallenge 2011-06-09T08:58:03 aichallenge: Nick Ham epsilon * rf444f48 / ants/dist/starter_bots/cpp/Timer.h : Fixed the cpp bots timer - http://bit.ly/kyg5qH 2011-06-09T08:58:42 Frontier: https://github.com/aichallenge/aichallenge/blob/epsilon/ants/dist/starter_bots/cpp/Timer.h 2011-06-09T08:59:17 *** sigh has quit IRC (Ping timeout: 276 seconds) 2011-06-09T08:59:39 *** delt0r_ is now known as deltOr 2011-06-09T09:00:02 *** deltOr is now known as delt0r- 2011-06-09T09:00:16 *** delt0r- is now known as deltOr 2011-06-09T09:05:30 *** locutus2 has quit IRC (Ping timeout: 260 seconds) 2011-06-09T09:10:22 *** akranis has quit IRC (Remote host closed the connection) 2011-06-09T09:12:48 i can see a potential problem 2011-06-09T09:13:04 people on windows might get rid of the linux part, then not realise that on the contest server the timer will then be using clock time 2011-06-09T09:13:15 *** UncleVasya has joined #aichallenge 2011-06-09T09:16:26 UncleVasya: I encoutered this error today: http://aichallengebeta.hypertriangle.com/visualizer.php?game=18458 2011-06-09T09:16:28 (double)(currentTime - startTime) / CLOCKS_PER_SEC 2011-06-09T09:16:55 If you're still using the first version of the ocaml starter, your bot might be susceptible to this bug 2011-06-09T09:17:05 I can explain how to make it ignore the bad input 2011-06-09T09:18:36 Also, does anyone know: is player_seed going to be official, or is it temporary? Does not seem to be mentioned in the spec. 2011-06-09T09:19:09 in the parameters? 2011-06-09T09:19:13 that is a bug if it's still being sent to players 2011-06-09T09:19:19 Yes, first turn 2011-06-09T09:19:30 smiley1983: was that ^^ how you were referring to do the timer in windows? 2011-06-09T09:19:43 i am a little worried now about windows users stripping the second half out 2011-06-09T09:19:52 smiley1983: You can give me a link to the page with colored changes in the file. 2011-06-09T09:20:26 antimatroid: No, sorry, this was not connected to the timer problem 2011-06-09T09:20:44 i know, but a week or two ago you said something to me about timers yeah? 2011-06-09T09:20:52 i meant my previous link too sorry :P 2011-06-09T09:20:53 UncleVasya: it's a one line change, easier to give it here, but I can pastebin the section if you prefer 2011-06-09T09:21:09 ok, give it here 2011-06-09T09:21:27 Ants.ml, in function add_line 2011-06-09T09:21:48 near the end, there is a line "failwith (blah)" 2011-06-09T09:22:00 You said it's only in the first version so I thought this fix already on the github. 2011-06-09T09:22:23 Oh, yes, I'll go and find the line there. Sorry :) 2011-06-09T09:22:37 antimatroid: yes, it was just a source code implementation of a timer function 2011-06-09T09:22:41 failwith (Printf.sprintf "unable to parse '%s'" line)))) 2011-06-09T09:23:02 Yes, it should be "Ants.ddebug (blah); gstate" instead 2011-06-09T09:23:12 I'll go and find it on github 2011-06-09T09:23:37 Don't worry 2011-06-09T09:24:10 *** stocha has joined #aichallenge 2011-06-09T09:24:24 *** Zaphus has joined #aichallenge 2011-06-09T09:24:25 Okay 2011-06-09T09:24:34 I just turned on computer and was wondered about my 24st place. :) 2011-06-09T09:25:10 The one on github replaces the whole debug output with just "gstate" - that is, it silently ignores the input and returns the game state unmodified 2011-06-09T09:25:24 Yes, there have been some more uploads I think. :) It is good to see 2011-06-09T09:25:33 I thought that another bots became so good. 2011-06-09T09:26:47 antimatroid: I don't actually know if the link I posted was going to be of any use, it just seemed to be source code providing what you were asking for, but I'm not really a c++ person so I could have misunderstood the whole situation. 2011-06-09T09:26:55 But I've change my decision and came here to cry "rankings is broken!". And never thought to look at the bot and find an issue :( 2011-06-09T09:27:19 smiley1983: you are very fast, thanks. 2011-06-09T09:27:34 *** Dlayne has joined #aichallenge 2011-06-09T09:27:50 *** Dlayne is now known as InsaneMalkavian 2011-06-09T09:28:00 Sometimes, if I happen to be looking at the screen :) sometimes I appear to be here but actually am not 2011-06-09T09:29:07 when did the change to jskill happen ? ranking seems to oscillate a lot last 36 hrs 2011-06-09T09:29:20 who is antstar? 2011-06-09T09:29:34 smiley1983: yeah fair enough, i ended up just doing the same :P 2011-06-09T09:29:38 mine 2011-06-09T09:30:05 ahh nice ;) 2011-06-09T09:30:36 does it use A* ? 2011-06-09T09:30:42 but my local ranking is always: stocha cellular antstar hungry 2011-06-09T09:30:44 yes 2011-06-09T09:30:48 meh, i'm still out on A* 2011-06-09T09:30:49 :) 2011-06-09T09:31:49 I can see how it might be useful. My ants will follow a sub-optimal path if sent on a long journey using my current pathfinding system. 2011-06-09T09:32:25 none of my intended strategies involve a single target for something 2011-06-09T09:33:24 pretty good representation from antarctica so far :P 2011-06-09T09:33:46 the ranking are mostly random right now. 2011-06-09T09:35:40 berak : what is your local ranking ? 2011-06-09T09:35:51 how is it done 2011-06-09T09:36:30 haven't got trueskill working yet, so i just average scores 2011-06-09T09:36:47 but where the bots come from ? 2011-06-09T09:37:01 ? 2011-06-09T09:37:13 The ranking on the test server is broken, because a lot of bot randomly crash at the start of the game. 2011-06-09T09:37:22 Am I right in thinking it's okay to include open source library code with your bot? 2011-06-09T09:37:48 berak: are your running the bots on your own machine ? Are you using stats from the test server ? 2011-06-09T09:38:00 smiley1983: pretty sure 2011-06-09T09:38:04 on my machine 2011-06-09T09:38:08 smiley1983: it is as long as your bot is open source :) 2011-06-09T09:38:25 are cellular and hungry form git maybe ? 2011-06-09T09:38:29 from 2011-06-09T09:38:57 smiley1983: so when you submit it, doesn't forget to join the source hehe :p 2011-06-09T09:39:16 no, mine, and yours is sto003_AdaptToContestbotLk , from the forum 2011-06-09T09:39:25 okay. 2011-06-09T09:39:44 Hm, that's a good point. GPL code might be problematic, fortunately the Mersenne library I want is LGPL :) 2011-06-09T09:39:49 at this point stocha bot and a few other always win (1 or 2) when they do not crash. 2011-06-09T09:40:07 and they crash randomly about 50% of the time. 2011-06-09T09:40:31 java bot mostly crash because the jvm doesn't want to launch. 2011-06-09T09:41:01 berak: what's your opinion on the scoring for the games in general? 2011-06-09T09:41:04 does it work/make sense? 2011-06-09T09:41:04 so, seems you have to do your own ranking calculatzion, excluding those 2011-06-09T09:41:12 Anyway, in the submission, we do join the code. because we submit the code. So open source is okay :) 2011-06-09T09:41:34 if i only knew ! 2011-06-09T09:41:42 scoring seems to make sense, in that it follow the number of ant most of time :p 2011-06-09T09:41:57 and the order of elimination also. 2011-06-09T09:42:48 berak : it's hard to compare java bots together (and other crashing bots) because they very rarely play each other without on crashing 2011-06-09T09:43:35 * antimatroid makes yet another random poke at java 2011-06-09T09:43:56 I think it's impossible to get a good sense of how well the ranking system works with so many crashes 2011-06-09T09:44:10 also, if i look at antstars latest games, it had a long streak of luck mostly against starterbots. 2011-06-09T09:44:33 Also, even the non-crashbots can be pretty random at the moment. 2011-06-09T09:44:51 pairing is coupled to ranking, some kind of feedback loop 2011-06-09T09:44:57 smiley1983: i'm afraid you are right. But for that it show that the ranking is hardly consistent when bots are randomly very strong/very weak :) 2011-06-09T09:44:58 some are at least 2011-06-09T09:46:49 what option do you launch the jvm with ? 2011-06-09T09:46:57 if somebody knows. 2011-06-09T09:47:08 *** Eruonen has joined #aichallenge 2011-06-09T09:49:14 my guess is the jvm is launch with asking too much memory. so it doesn't load. 2011-06-09T09:49:32 but for the first bots, wich have access to enough memory 2011-06-09T09:49:34 server has not enough ram 2011-06-09T09:49:42 known issue 2011-06-09T09:49:42 so one fix would be to reduced the asked memory for jvm. 2011-06-09T09:49:58 and then they would either not crash, or they would crash when allocating memory. 2011-06-09T09:50:13 i guess at this point few bot will use very much memory anyway. 2011-06-09T09:50:30 and if those crash, they will crash with consistency. wich is better. 2011-06-09T09:50:56 so make it the max jvm memory like 100 mb 2011-06-09T09:51:11 and then they will either work, or crash with consistency. 2011-06-09T09:51:39 i have no idea who can do this though. who is in charge of the server, and has access to it ? 2011-06-09T09:52:35 stocha: that has already been tried, in fact 100mb was exactly tried, it seems the jvm doesn't strictly obey it 2011-06-09T09:52:50 and refuses to start with less than 800mb anyway 2011-06-09T09:52:54 can i have a look at the params ? 2011-06-09T09:53:08 they're in compiler.py 2011-06-09T09:53:09 jvm silently ignore params if they are not right. 2011-06-09T09:53:23 i'll try to have a look 2011-06-09T09:53:47 janzert: do you have access to test server ? 2011-06-09T09:54:53 yes 2011-06-09T09:54:55 janzert, and the runtime settings are where ? 2011-06-09T09:55:36 berak: not sure which you mean, but server_info files 2011-06-09T09:55:56 is where the server configuration is stored 2011-06-09T09:57:09 because you invoke java once for compiling, and another time for running the bot in the sandbox, 2011-06-09T09:57:57 isn't compiler.py for compiling ? 2011-06-09T09:58:16 ah i see berak is already inquiring :p 2011-06-09T09:58:23 so i'll leave it to berak :p 2011-06-09T09:58:42 compiler.py has the full language definitions 2011-06-09T09:59:00 including the command line to run the submission after compilation 2011-06-09T09:59:06 Does the ranking get recalculated for a bot even if it doesnt enter any matches ? mine hasnt seen a game for 11 hours and has dropped from 54 to 48 2011-06-09T10:00:53 'George' was in a similar position, no games for 15 hours and dropped 4 points in that time (has now played a game that scored 0.00 so can't really judge after that) 2011-06-09T10:04:06 Zaphus: yes, the sigma will gradually increase which will cause a decline in the displayed skill 2011-06-09T10:10:53 that's going to be annoying, even if correct - because you have no control over it. would love to see a matchmaking system that doesnt leave one bot with no games while others play 10 in the same time period 2011-06-09T10:11:59 especially when the ranking is not demonstrated not to be random yet. 2011-06-09T10:12:09 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-06-09T10:13:26 janzert: 800mb is a Lot i find doubfull that the jvm wouldn't start with less. 2011-06-09T10:13:48 berak: have you dig into it ? 2011-06-09T10:14:07 wait a sec 2011-06-09T10:14:51 janzert: what i would do, is try to launch a simple looping java program, and aknoledge how small the jvm process can be (trying 100m seems fine though) 2011-06-09T10:14:55 *** mceier has quit IRC (Quit: leaving) 2011-06-09T10:15:06 it also refuses to start with less that 11 or so processes 2011-06-09T10:15:11 and also trying without any argument. because it can hardly be worst than 800 mb :p 2011-06-09T10:15:11 berak: I just managed to do old food information fairly easily 2011-06-09T10:15:36 *** UncleVasya has joined #aichallenge 2011-06-09T10:15:40 if you add an "oldFood" vector/set, then after processing vision information you can just check those locations for whether they're visible and if not make that location food 2011-06-09T10:15:43 can i have an access to the server ? or someone that know jvm stuff well 2011-06-09T10:15:57 i'm not a master at the thing to say the least. 2011-06-09T10:16:10 smiley1983: If I'll need to use an A* i think I can use "when a free ant was in this cell last time" as a cost heuristic function. 2011-06-09T10:16:10 antimatroid: do not eat old food! ( what was that ? ) 2011-06-09T10:16:23 what? 2011-06-09T10:16:40 ah, seems you pinged the wrong one 2011-06-09T10:16:41 oldFood is my food vector/set from the last turn 2011-06-09T10:17:04 it's probably not the most efficient, but at the start of the turn it's fairly negligable 2011-06-09T10:18:06 stocha: also it only has actually about 13mb resident but it has over 700mb of virtual space 2011-06-09T10:18:30 lol virtual space with 1sec time limits? 2011-06-09T10:19:04 i think the engine should send the bots a time limit just under what it actually uses 2011-06-09T10:19:05 normally I'd say the virtual space is unimportant anyway but I guess for some reason with the virtualization method the server is running under it takes it all into account 2011-06-09T10:19:32 antimatroid: virtual space is basically the address space a program uses 2011-06-09T10:19:49 have you guys tried upping the start time limit for the beta server for java bots? 2011-06-09T10:20:03 or is jvm literally crashing/not having enough memory? 2011-06-09T10:20:08 antimatroid: it crashes instantally, more time wouldn't do much 2011-06-09T10:20:13 ah yeah okay 2011-06-09T10:20:32 instantly even 2011-06-09T10:21:03 who is antimatter? 2011-06-09T10:21:07 can you trying launching any looping java program (let's say a bot) without any parameter for the jvm ? 2011-06-09T10:21:23 and report how much memory the jvm process take ? 2011-06-09T10:21:31 then try to play a bit with parameter. 2011-06-09T10:22:10 i expect the single -Xmx100m single parameter to give a Jvm that take about 100m of real memory. 2011-06-09T10:22:20 or less. 2011-06-09T10:22:47 berak knocked out of first place :P 2011-06-09T10:22:53 then we can digin in why the tournament have it wrong, once we know the server default are okay. 2011-06-09T10:23:02 ranking is random. 2011-06-09T10:23:07 never deserved it anyway 2011-06-09T10:23:20 nut took a screenshot for mom! 2011-06-09T10:23:31 :) 2011-06-09T10:24:05