2010-11-08T00:01:46 *** _demerzel <_demerzel!488039c1@gateway/web/freenode/ip.72.128.57.193> has quit IRC (Quit: Page closed) 2010-11-08T00:15:35 yay 2010-11-08T00:15:45 going to upload a working bot 2010-11-08T00:15:47 huzzah 2010-11-08T00:15:48 Top 10 players: bocsimacko(4305), shangas(4178), bix0r4ever(4137), rsergio(4034), george(4034), GreenTea(3992), dmj111(3988), Accoun(3983), fglider(3980), Xrillo(3977) 2010-11-08T00:15:49 * Prillicy dances. 2010-11-08T00:19:30 *** virdo has quit IRC (Ping timeout: 276 seconds) 2010-11-08T00:23:59 well, that went badly, heh 2010-11-08T00:25:14 *** Frontier has left #aichallenge 2010-11-08T00:25:37 *** Frontier has joined #aichallenge 2010-11-08T00:27:31 how do I watch the tcp battles? 2010-11-08T00:34:10 @tcp 2010-11-08T00:34:11 antimatroid: tcp = (#1) http://www.benzedrine.cx/planetwars/, or (#2) http://72.44.46.68/ 2010-11-08T00:34:21 go to the second link, click on rankings and find your username 2010-11-08T00:34:30 alternatively, use jBotManager and watch them live 2010-11-08T00:40:56 http://72.44.46.68/canvas?game_id=905052 <---it doesn't show aanything? 2010-11-08T00:42:56 *** decipher has joined #aichallenge 2010-11-08T00:44:58 hey guys i have some strategy questions if anybody wants to help i have a game tree with "funneling" i guess its called attach to it, i have no problem funneling ships to the closest and having no triangles happening and its clean but i do know how many leaders i should have.. the front guys that attack is it based on enemy owned planets, my owned planets, or hand picked, or map relative? 2010-11-08T00:45:59 I don't funnel personally 2010-11-08T00:46:11 I treat each planet like an individual entity until all neutrals are gone 2010-11-08T00:48:08 i didn't at first but it seems like its the top strat but first place kinda doesn't use it 2010-11-08T00:48:47 at first i had a game tree 1 ply but i wasn't getting anywhere wasn't fast enoug 2010-11-08T00:49:20 i top out at 550 rank now i have this new technique hybrid game tree and im not getting much further im just more consistant 2010-11-08T00:50:54 *** mega1 has quit IRC (Ping timeout: 240 seconds) 2010-11-08T00:51:11 is there a way to specifically challenge another person's bot? 2010-11-08T00:52:22 if its a jar .. you can load them up against each other 2010-11-08T00:52:31 whats your rank pril? 2010-11-08T00:59:38 very low 2010-11-08T00:59:44 22 hundred something 2010-11-08T00:59:48 I'm going to bed 2010-11-08T00:59:55 *** Prillicy has quit IRC (Quit: g'night) 2010-11-08T01:08:33 gnight 2010-11-08T01:15:05 *** davidjliu has quit IRC (Quit: Page closed) 2010-11-08T01:15:48 Top 10 players: bocsimacko(4333), shangas(4208), bix0r4ever(4158), rsergio(4078), george(4076), GreenTea(4048), Accoun(4021), Xrillo(4013), dmj111(4010), Hazard(3974) 2010-11-08T01:29:58 *** sigh_ has quit IRC (Remote host closed the connection) 2010-11-08T01:30:14 *** tapwater has quit IRC (Quit: tapwater) 2010-11-08T01:32:10 *** McLeopold has joined #aichallenge 2010-11-08T01:48:27 *** phxtrack has joined #aichallenge 2010-11-08T01:50:36 *** TheMacca has joined #aichallenge 2010-11-08T01:56:41 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-11-08T02:01:20 *** hellman has joined #aichallenge 2010-11-08T02:03:20 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-08T02:09:22 *** bhasker has quit IRC (Quit: bhasker) 2010-11-08T02:15:49 Top 10 players: bocsimacko(4318), shangas(4204), bix0r4ever(4151), george(4071), rsergio(4068), GreenTea(4039), Accoun(4009), Xrillo(4003), dmj111(3997), Hazard(3958) 2010-11-08T02:23:24 *** Appleman1234 has joined #aichallenge 2010-11-08T02:30:26 *** Ice_Harley has joined #aichallenge 2010-11-08T02:32:19 *** sigh has joined #aichallenge 2010-11-08T02:36:11 Is bocsimacko's jitter strategy key to his success or is it something else? Do the other bots 'lose' their target when he randomly reinforces his planets? 2010-11-08T02:36:31 sigh: What would you say? ^^ 2010-11-08T02:37:07 I don't know, I'm not that good at analysing these things 2010-11-08T02:37:19 ask McLeopold... he will have an opinion :) 2010-11-08T02:38:49 maybe he just wants to make us think this is important and we all end up with retarded RadomBot clones 2010-11-08T02:40:18 :) 2010-11-08T02:47:41 is anyone here scraping the contest site at the moment? and if so could you please turn it down a little 2010-11-08T02:48:19 *** jaspervdj has joined #aichallenge 2010-11-08T02:49:39 "accoun_stop_running_so_many_bots" who it? 2010-11-08T02:49:48 I let my front page reload every 15 min, reset it to 'web site default' 2010-11-08T02:50:32 one hit every 15 minutes should be fine :) 2010-11-08T02:50:57 Some people keep SQL databases of all games and refresh it every now and then, I think that is what you had in mind, right? 2010-11-08T02:51:16 something like that anyway 2010-11-08T02:51:36 the load average was up over 10.5 2010-11-08T02:52:18 The site is quite responsive actually 2010-11-08T02:54:10 yeah, the load average is back down now 2010-11-08T02:55:09 although watching the log someone is still obviously scraping all the game data right now 2010-11-08T02:55:24 at a rate of about 2 games per second 2010-11-08T02:56:57 whoever is scraping it would be better for you and us if you used game_info.php instead of visualizer.php 2010-11-08T03:13:13 *** p4p4 has joined #aichallenge 2010-11-08T03:15:48 Top 10 players: bocsimacko(4281), shangas(4182), bix0r4ever(4135), george(4068), rsergio(4047), Accoun(4011), GreenTea(4007), Xrillo(3983), dmj111(3968), fglider(3957) 2010-11-08T03:20:51 oh man.. i'm reading about django, this is epic 2010-11-08T03:45:38 *** vall has joined #aichallenge 2010-11-08T03:59:09 *** Florian95 has joined #aichallenge 2010-11-08T04:15:03 * antimatroid doesn't do any kind of logging of games 2010-11-08T04:15:48 Top 10 players: bocsimacko(4301), shangas(4162), bix0r4ever(4133), george(4068), rsergio(4046), Accoun(4008), GreenTea(4006), Xrillo(3977), dmj111(3967), fglider(3955) 2010-11-08T04:17:03 *** antimatroid1 has joined #aichallenge 2010-11-08T04:20:00 *** antimatroid has quit IRC (Ping timeout: 272 seconds) 2010-11-08T04:20:59 amstan: do you know of any users being deleted from the database? 2010-11-08T04:21:32 *** hellman has quit IRC (Quit: buffer overflow) 2010-11-08T04:21:43 janzert: no 2010-11-08T04:22:00 janzert: why? anyone having problems? 2010-11-08T04:22:04 *** hellman has joined #aichallenge 2010-11-08T04:22:14 did we work out whether it was a false alarm about elo rankings earlier or not? 2010-11-08T04:22:28 http://ai-contest.com/profile.php?user_id=7165 just ran into this 2010-11-08T04:22:43 it's a user_id with a valid submission but no user entry 2010-11-08T04:23:08 the submission was made back on 9-18 2010-11-08T04:23:16 where should i look? 2010-11-08T04:23:26 i see v_nsmv 2010-11-08T04:23:45 oops, wrong page 2010-11-08T04:23:46 http://ai-contest.com/profile.php?user_id=7164 2010-11-08T04:24:02 that was the user after it, there is a valid user for the id's before and after 2010-11-08T04:24:13 the record for 7164 just doesn't exist 2010-11-08T04:24:31 @sampledb 2010-11-08T04:24:32 amstan: sampledb = http://csclub.uwaterloo.ca/~amstan/aichallenge/sample.sql 2010-11-08T04:25:00 it looks to me like someone deleted the user but didn't mark the submission as inactive 2010-11-08T04:25:07 *** Naktibalda has joined #aichallenge 2010-11-08T04:25:13 janzert: was the elo ranking question asked to you? 2010-11-08T04:25:24 antimatroid1: not that I heard 2010-11-08T04:25:38 janzert: i did not delete any user, the policy is to deny most of these requests 2010-11-08T04:25:50 because it might cause exactly these problem 2010-11-08T04:26:08 *** TheMacca has quit IRC (Quit: Page closed) 2010-11-08T04:26:08 i can't find 7164 in the sampledb either 2010-11-08T04:26:09 someone was worrying that player 1 was being given a handicap as though it was a chess game where the first player is considered to have an advantage 2010-11-08T04:26:24 which was made when i was doing the irc bot 2010-11-08T04:26:40 i remember jeff banning at least one person at some point 2010-11-08T04:26:43 but it was a while ago 2010-11-08T04:26:47 janzert: Dump completed on 2010-09-10 5:32:51 2010-11-08T04:26:51 antimatroid1: oh! 2010-11-08T04:27:04 amstan: yeah, I'm thinking maybe danielvf was trying to check something possibly 2010-11-08T04:27:20 looks like sampledb is all in < 4000 user_id range 2010-11-08T04:27:23 janzert: that was it 2010-11-08T04:27:32 An organizer might want to sent him an email. 2010-11-08T04:27:34 http://ai-contest.com/profile.php?user_id=7164 2010-11-08T04:27:35 http://ai-contest.com/profile.php?user_id=7052 2010-11-08T04:27:53 looks like he deleted the other one too 2010-11-08T04:28:07 man.. he should have just changed the pass 2010-11-08T04:28:09 it was danielv? 2010-11-08T04:28:14 nah, jeff 2010-11-08T04:28:19 ahh 2010-11-08T04:28:26 "I just gave him the 'ol banhammer." 2010-11-08T04:28:36 ugh 2010-11-08T04:28:49 well I'll mark the submissions as inactive 2010-11-08T04:29:59 antimatroid1: no player1 isn't given any advantage or disadvantage 2010-11-08T04:30:36 I checked that and actually it caused a bit of confusion because the way it was disabled was by actually modifying the source code of bayeselo 2010-11-08T04:31:25 at some point I should probably change it and disable it in the normal way by just giving bayeselo the correct command 2010-11-08T04:33:18 janzert: thanks for checking, i will say as such when asked again 2010-11-08T04:33:31 *** perror has joined #aichallenge 2010-11-08T04:34:15 *** hellman has quit IRC (Remote host closed the connection) 2010-11-08T04:34:50 another thing that would have effected it is that bayeselo is always fed with the winner as player1 2010-11-08T04:35:25 I double checked and did a test randomizing the winner to player1 or player2 and the ranking result was the same 2010-11-08T04:36:04 <3 2010-11-08T04:36:31 i'll be free from undergrad in 15 hours, then i can get back to my bot 2010-11-08T04:36:40 the big thing limiting rank accuracy right now is only the past 3 days of games are used because of database performance limitations 2010-11-08T04:36:48 I'm going to work on solving that soon 2010-11-08T04:37:07 yeah okay, can you make sure the elo's are reset at the start of the final contest this time? 2010-11-08T04:37:28 hmm, they weren't last time? 2010-11-08T04:37:34 and i was wondering earlier, does it cause problems with an elo system and not completely randomising pairings? 2010-11-08T04:37:41 but either way yeah, I agree with that 2010-11-08T04:37:42 i'm pretty sure they weren't in th eend 2010-11-08T04:38:04 in that, those who win their first game, get to play better better, then on average they end up higher maybe? 2010-11-08T04:38:22 better people* (i have lots of sleep to catch up on sorry) 2010-11-08T04:38:37 antimatroid1: depends how they are skewed, when players can choose there opponent it can be a real problem 2010-11-08T04:39:25 remember we aren't using a 'true' elo system to find the elo score 2010-11-08T04:39:53 bayeselo handles it quite a bit better actually than if you tried to just use the basic elo system 2010-11-08T04:40:26 yeah, but i'm still curious whether the bots that win their first game are expected to end up with a higher ranking than an exactly equivalent bot that loses its first game 2010-11-08T04:41:07 I don't think it would last long, no 2010-11-08T04:41:22 but I don't have any hard data on it 2010-11-08T04:41:43 yep, it's just something that was nagging me before, not a huge issue 2010-11-08T04:42:24 speaking of ratings... my bot hovered around 250 for a week and then jumped to 150 without me doing anything... http://ai-contest.com/profile.php?user_id=10821 2010-11-08T04:42:48 lavalamp_: we're never going to get perfect rankings 2010-11-08T04:42:49 btw, working on matchup rewrite right now that will use a previous submission's rank for a new submission's pairing 2010-11-08T04:42:57 *** Appleman1234 has quit IRC (Ping timeout: 255 seconds) 2010-11-08T04:43:00 my issues come in trying to avoid any undesirable advantages 2010-11-08T04:43:17 lavalamp_: see above re only 3 days of games used 2010-11-08T04:43:35 antimatroid1: yeah I know 2010-11-08T04:44:14 janzert: ah, I guess that could explain it, sorta. 100 places just seems like an awful lot of variability 2010-11-08T04:45:08 my bot hovered around 40 then dropped to 95 \o/ 2010-11-08T04:45:11 lavalamp_: there are only probably reasonable attempts at bots from about 500 up, and i suspect there isn't huge amounts of difference between them where you are, more luck of the map and bot strength 2010-11-08T04:45:26 sigh: i'm pretty sure top 50 just got more competitive in the last week 2010-11-08T04:45:45 janzert: When that guy asked about the Elo thing, I glanced through the source code, and in the file http://code.google.com/p/ai-contest/source/browse/trunk/planet_wars/backend/process_ratings.py 2010-11-08T04:46:15 on line 36... should that really be an insert? 2010-11-08T04:46:18 I'm going to change some parameters and upload a slightly different bot 2010-11-08T04:46:35 does everything get dropped from that table beforehand? 2010-11-08T04:46:59 yes and no 2010-11-08T04:47:14 how else would you add a row? 2010-11-08T04:47:34 does the db keep a rating history? 2010-11-08T04:47:36 * janzert freely admits having minimal sql experience 2010-11-08T04:47:47 I expected it to be an update 2010-11-08T04:48:19 nope, each new set of rankings is added 2010-11-08T04:48:28 all the old ones are still there 2010-11-08T04:48:34 I'm fighting that a bit right now 2010-11-08T04:48:39 ok so it does keep a history and that's intentional 2010-11-08T04:48:54 since the rankings table currently has over 40 million rows 2010-11-08T04:49:33 LOL 2010-11-08T04:49:51 oh I see, leaderboard_id allows you to tell 2010-11-08T04:50:08 ok I hadn't noticed that 2010-11-08T04:50:11 right, each set of rankings is a leaderboard 2010-11-08T04:50:18 gotcha 2010-11-08T04:50:29 .... and this is done how many times per day? 2010-11-08T04:50:32 janzert: that's not something you have to "admit" like it's a problem :P 2010-11-08T04:51:12 sigh: still nothing in the way of a new bot then? 2010-11-08T04:51:13 lavalamp_: every 5 minutes 2010-11-08T04:51:37 ok I'm going back to my original position which is that that is a problem 2010-11-08T04:51:40 antimatroid1: nah... I'm trying something new though 2010-11-08T04:53:12 oohh... I haven't seen that state before "Ready to be compiled" 2010-11-08T04:53:18 i ahve 2010-11-08T04:53:19 :P 2010-11-08T04:55:13 janzert: if I were doing it, I would write a script to copy the last leaderboard from each day into a new table... then do stuff with the new, manageable table 2010-11-08T04:55:14 "entry created in db" what is this state? 2010-11-08T04:55:59 *** Appleman1234 has joined #aichallenge 2010-11-08T04:56:00 it's the inital state before processing the entry even starts 2010-11-08T04:56:17 rare and a bit wierd that you would see it 2010-11-08T04:56:32 *** decipher has quit IRC (Quit: Page closed) 2010-11-08T04:56:44 goooooooooooooooood mornin everybody 2010-11-08T04:56:46 janzert: this guy seems stuck in that state: http://ai-contest.com/profile.php?user_id=12315 2010-11-08T04:56:51 my first opponent 2010-11-08T04:57:46 what is your profile page? 2010-11-08T04:58:01 mine: http://ai-contest.com/profile.php?user_id=6041 2010-11-08T04:58:25 *** Gorovkhaia2 has quit IRC (Quit: leaving) 2010-11-08T04:58:57 awesome, I have rank 2051 2010-11-08T04:59:22 hehe 2010-11-08T05:00:14 *** Appleman1234 has quit IRC (Ping timeout: 240 seconds) 2010-11-08T05:00:40 awww, i miss botting it up 2010-11-08T05:00:45 so close, yet so far 2010-11-08T05:00:58 i only just started studying for this exam a couple of hours ago, going to be a LONG night 2010-11-08T05:01:41 bot it up now! study later :P 2010-11-08T05:02:29 hell, if it is an open book exam, then you can leave your studying until after the exam starts :D 2010-11-08T05:02:46 we don't get open book exams for maths 2010-11-08T05:03:03 although i get take home exams mostly next year :) 2010-11-08T05:03:13 heh 2010-11-08T05:04:59 *** jaspervdj has quit IRC (Ping timeout: 272 seconds) 2010-11-08T05:06:23 *** Accoun has quit IRC () 2010-11-08T05:10:36 *** jaspervdj has joined #aichallenge 2010-11-08T05:15:49 Top 10 players: bocsimacko(4272), shangas(4140), bix0r4ever(4097), george(4032), rsergio(4000), Accoun(3987), Polsky(3980), GreenTea(3977), Xrillo(3941), dmj111(3932) 2010-11-08T05:16:24 oh, dmj dropping 2010-11-08T05:17:01 *** Accoun has joined #aichallenge 2010-11-08T05:23:01 *** Appleman1234 has joined #aichallenge 2010-11-08T05:23:05 YEEEEEEEEEEEEEEEEEEEEEES its working 2010-11-08T05:49:46 Error323: :D 2010-11-08T05:49:49 what can it do? 2010-11-08T05:51:24 predict awesomeness :P 2010-11-08T05:53:10 :D 2010-11-08T05:53:20 can it beat bocksimacko? :D 2010-11-08T05:53:43 is this a heuristic one? 2010-11-08T05:53:46 not sure yet still in phase 1 =) 2010-11-08T05:53:50 nope! :) 2010-11-08T05:53:52 nice 2010-11-08T05:53:54 what is it? 2010-11-08T05:54:00 minimax hybrid? 2010-11-08T05:54:03 can't say yet :P 2010-11-08T05:54:17 ok :( 2010-11-08T05:56:07 *** Aaaarg has joined #aichallenge 2010-11-08T05:56:49 Aaaarg: from spring? 2010-11-08T05:57:55 huh? 2010-11-08T05:58:18 are you the pure developer? 2010-11-08T05:58:20 from spring 2010-11-08T05:58:33 he's called Aaargh iirc 2010-11-08T05:58:49 that doesn't sound like me 2010-11-08T05:59:00 heh ok :P nvm then =) 2010-11-08T06:15:33 sigh: yeah... i'm not making much progress, so that drop will most likely be permanent. :-/ 2010-11-08T06:15:51 Top 10 players: bocsimacko(4269), shangas(4122), bix0r4ever(4086), george(4013), rsergio(3989), Accoun(3978), Polsky(3972), GreenTea(3967), Xrillo(3938), dmj111(3924) 2010-11-08T06:16:15 dmj111: sounds like me... except better 2010-11-08T06:17:37 thanks :) I wish I could just give in, though.... but I keep thinking I have one more idea that might work. 2010-11-08T06:19:20 yeah... mostly I'm just coasting now... maybe I'll find something better but I'm not too fussy 2010-11-08T06:19:35 mostly I'm looking forward to finding out wtf bocsimako is doing 2010-11-08T06:19:56 yes, me too. 2010-11-08T06:25:10 dmj111: i've seen you say that so many times :P 2010-11-08T06:25:18 you're well and truly hooked, it's less than a month left 2010-11-08T06:25:26 i'm planning a full rewrite from tomorrow 2010-11-08T06:28:21 i am truly hooked. I hope to get something else working, but this time, it might be it :) I am working on _just one more_ rewrite. finding lots of bugs. I'll probably use it after it is debugged, but it doesn't look super great. 2010-11-08T06:36:18 ofrias's bot isn't bad 2010-11-08T06:36:20 http://www.ai-contest.com/visualizer.php?game_id=6798098 2010-11-08T06:45:02 *** boegel has joined #aichallenge 2010-11-08T06:54:34 *** Frontier has quit IRC (Ping timeout: 240 seconds) 2010-11-08T06:55:09 *** Migi32 has joined #aichallenge 2010-11-08T07:04:10 do any of you make use of the fact that the game is 2D? I personally only look at planet-planet distances so it could as well have been a 3D game but I was wondering if anyone else did. 2010-11-08T07:05:54 Migi32: i use distances too, so you are correct 2010-11-08T07:06:10 but to calculate the distances it needs to be 2d ;) 2010-11-08T07:06:49 actually i mean the format of data my program parses has to be 2d 2010-11-08T07:07:11 but with a few tweaks, it could easily be changed to 3d 2010-11-08T07:15:41 *** hellman has joined #aichallenge 2010-11-08T07:15:51 Top 10 players: bocsimacko(4242), shangas(4098), bix0r4ever(4081), george(4032), rsergio(3984), Polsky(3971), GreenTea(3967), Accoun(3957), Xrillo(3949), dmj111(3933) 2010-11-08T07:25:26 raedwulf, same here. If I were given 3 coordinates instead of 2 it would only be some very small changes that could make my bot play on those maps 2010-11-08T07:26:16 migi32: i treat it like a graph too 2010-11-08T07:26:27 well, it is a graph 2010-11-08T07:26:28 but i think it'd still be more complicated like that were it 3d 2010-11-08T07:26:39 in that the distances don't have the same constraints 2010-11-08T07:26:43 with respect to the geometry 2010-11-08T07:26:45 treating it as 2d can be useful for some heuristics though 2010-11-08T07:27:14 yea that's I was wondering if any of you calculated things that would make no sense in 3D, like battlefronts or "back" planets 2010-11-08T07:27:14 for example if you have some geometric model for your front lines 2010-11-08T07:27:20 :P 2010-11-08T07:27:25 :) 2010-11-08T07:27:36 am i right that you can't represent every 3d map in a 2d way such that the graphs are the same? 2010-11-08T07:27:51 you are right 2010-11-08T07:27:53 battlefront would certainly make sense in 3d 2010-11-08T07:28:04 well yes but they would be planes instead of lines 2010-11-08T07:28:08 redistribution in a way does my battlefront 2010-11-08T07:28:08 technically, this is not exactly 2d 2010-11-08T07:28:15 well, euclidean 2d 2010-11-08T07:28:18 no, the ceiling ruins it 2010-11-08T07:28:57 could that ever result in a map not being perfectly symmetric when it would under the euclidean metric? 2010-11-08T07:29:23 but for example in 3D you can have 4 planets at equal distances to each other, in 2D only 3 2010-11-08T07:29:27 I don't see how, so long as it is centered on (0,0) 2010-11-08T07:29:59 migi32: my point is the graph encapsulates that, in that you can't have a graph with 4 planets equal distance apart 2010-11-08T07:30:03 sigh, even that's not necessary. Same distance => same ceil(distance) 2010-11-08T07:30:20 i didn't think it's possible, just thought i'd pose the question 2010-11-08T07:30:30 yeah 2010-11-08T07:30:34 Migi32: yeah, your right.. my bad 2010-11-08T07:30:41 anyway, must keep studying, pdes is so gross and lame 2010-11-08T07:30:47 anywya, why restrict ourselves to 3d :P 2010-11-08T07:30:54 i want to do nd tron 2010-11-08T07:31:04 would that be interesting at all? 2010-11-08T07:31:05 even 3d would be cool 2010-11-08T07:31:16 really? 2010-11-08T07:31:21 because my brain can only visualise > 3D if I'm on LSD 2010-11-08T07:31:21 can't really watch, would be cool to do a couple of bots and play them against each other 2010-11-08T07:31:24 3d seems too sparse 2010-11-08T07:31:30 lsd is a wonderful thing 2010-11-08T07:32:02 sigh: the bots wouldn't be as intelligent, but you still only have max 6 moves per turn 2010-11-08T07:32:09 in 3d that is 2010-11-08T07:32:15 it's start blowing out as dimensions went up 2010-11-08T07:32:30 well, I was planning to make some video about planetwars and upload it to youtube, and while thinking about what to put in there I thought of a 2010-11-08T07:32:39 3d visualiser, and then 3d planetwars :) 2010-11-08T07:32:48 antimatroid1: my gut feeling is too much freedom 2010-11-08T07:33:02 sigh: the game has to end in finite time 2010-11-08T07:33:10 just use relatively small maps 2010-11-08T07:33:17 antimatroid1: sure, but that doesn't make it interesting 2010-11-08T07:33:33 you can run tron on a complete graph in finite time, but that game is trivial :P 2010-11-08T07:33:54 wait what? 2010-11-08T07:34:01 sigh, or on a tree 2010-11-08T07:34:07 did I say something stupid? 2010-11-08T07:34:22 but you can run tron on any finite graph in finite time 2010-11-08T07:34:49 no, i just don't know what you mean 2010-11-08T07:35:00 a complete graph is the extreme case of too much freedom 2010-11-08T07:35:07 which is why I used that example 2010-11-08T07:35:24 complete meaning every node is connected to every other node? 2010-11-08T07:35:28 yeah 2010-11-08T07:35:37 atm complete to me means every cauchy sequence converges in the space :p 2010-11-08T07:36:04 :) 2010-11-08T07:36:19 so why are you running tron on a graph? 2010-11-08T07:36:27 minimax works really well for tron 2010-11-08T07:36:32 well, with small maps 2010-11-08T07:36:55 I'm modelling the board as a graph 2010-11-08T07:37:05 because that is the obvious generalisation 2010-11-08T07:37:14 a million bucks if you can find play tron optimally on a complete graph with the constraint that you have to end up on your starting point 2010-11-08T07:37:31 haha 2010-11-08T07:38:03 what's the graph? min dist path to each square? 2010-11-08T07:38:10 what if that's not the optimal path? 2010-11-08T07:38:48 the graph is the distances between squares. But the squares are not necessarily on a rectangular grid :) 2010-11-08T07:38:50 edges connect any two nodes that can be reach in one move from each other 2010-11-08T07:38:56 more like the planets in planetwars 2010-11-08T07:39:07 so it's not actually a complete graph? 2010-11-08T07:39:20 no... not for normal tron 2010-11-08T07:39:44 gah, i'm pulling out paper and pencil 2010-11-08T07:39:54 I generalised tron to arbitrary graphs, then constructed a silly example 2010-11-08T07:40:24 regular tron on a complete graph would be a 2x1 (or 1x1) playing field 2010-11-08T07:40:56 normal tron is a square lattice... corner nodes have 2 edges, edge nodes have 3 and the rest have 4. 2010-11-08T07:40:59 hmm, i wish i had thought about tron like this earlier in the year 2010-11-08T07:41:11 sigh: they weren't all open maps 2010-11-08T07:41:12 each move you may move to an adjacent node which has never been occupied 2010-11-08T07:41:24 and after first move, max number of turns was 3 2010-11-08T07:41:28 antimatroid1: then you are just removing skme edges 2010-11-08T07:41:30 *sme 2010-11-08T07:41:42 or "pre-visiting" some nodes 2010-11-08T07:41:45 oh man, i want to do tron all over now 2010-11-08T07:41:48 can think of it either way 2010-11-08T07:42:26 well, they are the same way... you are just removing the node from the graph 2010-11-08T07:42:29 we need to get tcp servers for these older games going 2010-11-08T07:42:32 haha 2010-11-08T07:43:01 yeah I've been thinking about possible future AI Contests. How about tron with N "snakes" instead of 1? :P 2010-11-08T07:43:18 nah, n player is way too hard 2010-11-08T07:43:29 you're going to need to dive into coalitions then 2010-11-08T07:43:30 re-hashing old contests is not as fun either 2010-11-08T07:43:33 and the maths for that is messy 2010-11-08T07:43:36 no, 2 player, N snakes per player 2010-11-08T07:43:44 no, i would just like a tcp like thing for tron 2010-11-08T07:43:50 oh, okay 2010-11-08T07:44:01 too much of a repeat still 2010-11-08T07:44:14 i like the idea of doing a full on 3d world game 2010-11-08T07:44:30 i'm not much of a gamer, so fps doesn't really appeal 2010-11-08T07:44:56 fps with bots that have perfect accuracy is pretty lame as well 2010-11-08T07:45:18 i guess idea would be to restrict movement perhaps 2010-11-08T07:45:21 not to mention there's quite some research on the topic already, which is kind of a deal-breaker for me 2010-11-08T07:45:35 so you can only rotate your line of vision at a given rate 2010-11-08T07:45:56 lol, most of the ideas being used in these contests are far from original 2010-11-08T07:46:01 it's just rehashing old concepts 2010-11-08T07:46:28 what I mean is things like chess and FPS games already have very strong AI's 2010-11-08T07:46:42 Galcon and Tron much less 2010-11-08T07:46:50 chess would be the worst choice :P 2010-11-08T07:46:53 i think dots and boxes would make for an intersting game 2010-11-08T07:47:04 possibly, yeah 2010-11-08T07:47:11 barrier for entry is higher 2010-11-08T07:47:18 I don't know, a bit too abstract for me, I guess. 2010-11-08T07:47:23 the danger is that someone actually solves it 2010-11-08T07:47:41 I would like to see go-moku/connect-5 2010-11-08T07:47:44 with PlanetWars the first thing I though was "I can do this.", with Tron as well. 2010-11-08T07:47:50 with dots and boxes, meh. 2010-11-08T07:47:54 i'd be down for a contest on playing large 2 player non-zero sum games, but it'd just be who can approximate msne best 2010-11-08T07:48:22 you can't really choose traditional games for the competition, because there's far too much influence from existing theory 2010-11-08T07:48:31 what about nxn tic tac toe? 2010-11-08T07:48:41 that's been researched t death, no? 2010-11-08T07:48:46 lol, true 2010-11-08T07:48:49 though Go might be interesting because it might spark so bright ideas ^^ 2010-11-08T07:48:59 Go is way to high barrier for entry 2010-11-08T07:48:59 some* 2010-11-08T07:49:00 Go is a beast from what i've heard 2010-11-08T07:49:04 yea 2010-11-08T07:49:09 i don't know know how to play it myself 2010-11-08T07:49:15 neither 2010-11-08T07:49:29 I know how to play, and I don't know where to start 2010-11-08T07:49:33 other than all the fuss abut UCT 2010-11-08T07:49:51 even figuring out where to finish is a bitch 2010-11-08T07:50:03 but i heard its hard and computers haven't a clue how to play it well :P 2010-11-08T07:50:04 lol, i came up with something random to test the graph thin on with tron 2010-11-08T07:50:13 and actually came up with a reasonably complicated decision 2010-11-08T07:50:23 raedwulf: that's a simplification 2010-11-08T07:50:28 i'll do a paint job, hang on 2010-11-08T07:51:15 sigh: i didn't wish to write a book on the topic ^^ 2010-11-08T07:51:21 on 5x5 go has been solved... 9x9 they are close to pro level... 19x19 they are amateur dan level which is stronger than a lot of people will be after a year of playing 2010-11-08T07:51:40 sigh: ah right cool 2010-11-08T07:51:42 19x19 is full-size 2010-11-08T07:52:00 i was thinking 19x19 2010-11-08T07:52:33 yeah... but I wouldn't say "haven't a clue how to play it well" 2010-11-08T07:52:51 ok :P 2010-11-08T07:52:54 can't compete against good players is more accurate 2010-11-08T07:53:41 * raedwulf sighs 2010-11-08T07:53:48 :) 2010-11-08T07:53:58 :P 2010-11-08T07:54:07 well, when developing a bot it's good to see why it's doing bad. That is so for planetwars and tron. But if my Go bot would be stronger than me, it would come down to tweaking numbers and testing it out on TCP 2010-11-08T07:54:44 http://imgur.com/9BrlP.png 2010-11-08T07:54:57 i think you just have to randomise your decision for that 2010-11-08T07:55:09 what is that? 2010-11-08T07:55:19 tron? 2010-11-08T07:55:26 a possible tron map, you start as player 1, other person starts as player 2 2010-11-08T07:55:48 antimatroid1, player 1 always wins, no? 2010-11-08T07:55:54 yeah, I agree 2010-11-08T07:56:01 just go right 2010-11-08T07:56:01 Migi32: no 2010-11-08T07:56:13 oh wait 2010-11-08T07:56:16 lol 2010-11-08T07:56:18 i didn't think of player 1 aha 2010-11-08T07:56:22 yeah nevermind 2010-11-08T07:56:23 ... 2010-11-08T07:57:12 for JBotManager users: any opinion/comment about this mod : http://i54.tinypic.com/jzyyir.jpg 2010-11-08T07:57:32 looks cool 2010-11-08T07:57:35 but they had a good rock-paper-scissor map I think it went like this: 2010-11-08T07:57:38 ########## 2010-11-08T07:57:38 #________# 2010-11-08T07:57:38 ##1#2##### 2010-11-08T07:57:38 #________# 2010-11-08T07:57:38 ########## 2010-11-08T07:57:40 what i want more than anything is the ability to send exact fleet sizes 2010-11-08T07:57:57 if both go the same way, player 2 wins, otherwise player 1 wins 2010-11-08T07:58:20 yeah, there were plenty of sitations that were weird like that 2010-11-08T07:58:53 Nash equilibria would've done great for Tron, I think. Much more than for PlanetWars. 2010-11-08T07:59:41 nullkuhl, what's that slider there at the bottom? Don't remember if it's also there for the original JBotManager 2010-11-08T07:59:44 it was really easy to do a solid evaluation function on game states 2010-11-08T07:59:52 Migi32: it is, planet sizes scaling 2010-11-08T07:59:58 it was then just a matter of searching and maximising path the death once separated 2010-11-08T08:00:09 hence why c++ did so well 2010-11-08T08:01:28 antimatroid1: am not sure if its easy to type in a number and send to a target planet while the game is running in JBotManager 2010-11-08T08:01:56 when one clicks on a planet, input string, when they click on dest, convert string to num and send fleets 2010-11-08T08:08:57 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-11-08T08:08:58 thinking about Tron AI just gave me quite a good idea. PlanetWars and Tron are actually more similar than you may think :) 2010-11-08T08:10:59 Migi32: you want to re-create wormbot? 2010-11-08T08:11:12 antimatroid1: yes but that means you wont be able to see current planet ships while ur are typing the no of ships you want to send 2010-11-08T08:11:16 oh, no not in that way :) 2010-11-08T08:11:30 :P 2010-11-08T08:12:38 it's just one word but I'm going to test out this idea first and if it doesn't work well then I'm going to say it :) 2010-11-08T08:12:49 it's way different o tron 2010-11-08T08:13:34 maybe leaf node evaluation isn't ;) 2010-11-08T08:15:51 Top 10 players: bocsimacko(4256), shangas(4111), bix0r4ever(4094), george(4047), rsergio(3997), GreenTea(3981), Polsky(3977), Accoun(3972), Xrillo(3961), dmj111(3945) 2010-11-08T08:16:59 *** Utkarsh has joined #aichallenge 2010-11-08T08:19:52 antimatroid1, did you participate in that Tron contest? I don't see your name in the rankings. 2010-11-08T08:20:11 antimatroid1 = youwillfail 2010-11-08T08:20:48 he was in the tron contest, dunno if that was his name then as well 2010-11-08T08:21:30 nope, no youwillfail in the rankings either 2010-11-08T08:22:08 *** phreeza has joined #aichallenge 2010-11-08T08:22:16 oh, also he said he got disqualified for timeout 2010-11-08T08:22:42 huh? 2010-11-08T08:22:52 if you time out once you lose not that game but the whole contest? 2010-11-08T08:23:08 in tron, apparently that was the case 2010-11-08T08:23:14 wow 2010-11-08T08:23:39 O_O 2010-11-08T08:23:54 Migi32: i did, got disqualified for timing out in the final contest 2010-11-08T08:23:58 was sitting around 10th at the time 2010-11-08T08:24:20 ow, I feel sorry for you 2010-11-08T08:24:22 i didn't even know that was the rule till it happened, extremely devestating 2010-11-08T08:24:31 would be worse for this, that was only a few weeks of work 2010-11-08T08:25:06 i didn't shut up until that was changed for this contest :) 2010-11-08T08:25:12 haha :D 2010-11-08T08:25:36 was a silly constraint anyway 2010-11-08T08:25:43 yes 2010-11-08T08:25:47 yep, the reason given was bots that timeout cost resources 2010-11-08T08:25:56 but really they'd have on average faster games than others 2010-11-08T08:26:00 cause they lose all the time 2010-11-08T08:26:22 if someone wanted to spam the server, they'd upload a lot of half decent bots that always use their time limit 2010-11-08T08:26:40 but was it a timeout as in you took 1.01s for a move or did it get stuck in an infinite loop? 2010-11-08T08:26:55 and they had to kill it themselves? 2010-11-08T08:26:56 you don't know, it cuts you off after a second? 2010-11-08T08:27:01 i never even got to see the game that killed me 2010-11-08T08:27:29 http://csclub.uwaterloo.ca/contest/profile.php?user_id=2788 2010-11-08T08:27:32 that was my tron profile 2010-11-08T08:27:47 lol 2010-11-08T08:27:48 a pretty good record 2010-11-08T08:27:54 Win win win win win ... :P 2010-11-08T08:28:07 i was consistently hitting 1st place with under a week to go 2010-11-08T08:28:14 i was expecting to be in the top 10 tbh 2010-11-08T08:29:37 in fact, that record looked better than second place if it had kept going :( 2010-11-08T08:30:17 well, a1k0n would definitely have beaten you on this map: http://csclub.uwaterloo.ca/contest/visualizer.php?game_id=4127685 2010-11-08T08:30:22 but still, impressive bot! 2010-11-08T08:30:48 you can understand my devestation when it was disabled and they wouldn't reenable it :P 2010-11-08T08:30:57 yea I can :) 2010-11-08T08:31:46 hence why i'm determined to get top 10 for this 2010-11-08T08:31:52 lol, this game is funny: http://csclub.uwaterloo.ca/contest/visualizer.php?game_id=4121221 2010-11-08T08:32:17 the end is so... anticlimactic :P 2010-11-08T08:32:21 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-11-08T08:32:44 antimatroid1, let me check you out on TCP. Are you running your latest/best bot there? 2010-11-08T08:32:55 oh, apparently not 2010-11-08T08:33:09 i had an exam saturday, one today and another tomorrow, no bot atm :P 2010-11-08T08:33:20 oh ok 2010-11-08T08:33:22 i'm doing a complete rewrite starting tomorrow 2010-11-08T08:33:26 I will let you study then :P 2010-11-08T08:35:49 antimatroid1: you almost lost here ? http://csclub.uwaterloo.ca/contest/visualizer.php?game_id=4127689 2010-11-08T08:36:31 nullkuhl: it knew it'd win 2010-11-08T08:36:42 because its the turn of the other player ? 2010-11-08T08:36:45 me evalutation function picked the fastest ensurable win 2010-11-08T08:36:50 my* 2010-11-08T08:36:56 it made it a mean mo fo 2010-11-08T08:37:26 you'll see that it even set it up :P 2010-11-08T08:37:35 that's the power of minimax with a fairly small search space for you 2010-11-08T08:37:48 ah, the little hole there :P 2010-11-08T08:37:50 nice 2010-11-08T08:38:11 *** Utkarsh has joined #aichallenge 2010-11-08T08:38:22 I gtg now 2010-11-08T08:38:25 see ya 2010-11-08T08:38:26 this remeniscing isn't doing good for my devestation levels :P 2010-11-08T08:38:28 see ya 2010-11-08T08:38:29 me too 2010-11-08T08:38:30 sleep 2010-11-08T08:38:34 and good luck on your exam antimatroid1 :) 2010-11-08T08:38:38 *** Migi32 has quit IRC (Quit: http://boinc.berkeley.edu) 2010-11-08T08:38:39 :) 2010-11-08T08:38:45 yah, good luck 2010-11-08T08:38:49 *** sigh has quit IRC (Quit: Leaving...) 2010-11-08T08:39:23 hmmm, the bot could have won that game a lot earlier 2010-11-08T08:39:40 yea 2010-11-08T08:40:01 that would have been some kind of bug 2010-11-08T08:40:01 *** Ice_Harley has quit IRC (Ping timeout: 265 seconds) 2010-11-08T08:40:06 no way it should have missed that 2010-11-08T08:40:14 it was like one move away in the tree 2010-11-08T08:56:45 *** bhasker has joined #aichallenge 2010-11-08T09:03:23 *** Rubicon-|-Cross has quit IRC (Read error: Connection reset by peer) 2010-11-08T09:06:36 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-11-08T09:09:49 *** bhasker_ has joined #aichallenge 2010-11-08T09:09:50 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-11-08T09:13:01 *** Utkarsh has joined #aichallenge 2010-11-08T09:14:48 *** bhasker has joined #aichallenge 2010-11-08T09:14:49 *** bhasker_ has quit IRC (Read error: Connection reset by peer) 2010-11-08T09:15:52 Top 10 players: bocsimacko(4220), bix0r4ever(4064), shangas(4057), george(4038), rsergio(3981), GreenTea(3971), Accoun(3956), Polsky(3949), Xrillo(3942), dmj111(3922) 2010-11-08T09:24:40 *** mk has joined #aichallenge 2010-11-08T09:25:07 *** mk is now known as Guest13825 2010-11-08T09:33:33 *** demerzel_ has joined #aichallenge 2010-11-08T09:34:34 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-08T09:39:25 *** Utkarsh has joined #aichallenge 2010-11-08T09:40:27 *** Azrathud has quit IRC (Ping timeout: 276 seconds) 2010-11-08T09:44:00 *** bhasker_ has joined #aichallenge 2010-11-08T09:44:01 *** bhasker has quit IRC (Read error: Connection reset by peer) 2010-11-08T09:44:49 *** yasith_ has quit IRC (Ping timeout: 265 seconds) 2010-11-08T09:51:04 *** ananda_ has joined #aichallenge 2010-11-08T09:51:30 Is there a way we can call JBotmanager.jar via commandline and receive game winner via commandline? 2010-11-08T09:52:27 *** mega1 has joined #aichallenge 2010-11-08T09:52:57 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-11-08T09:55:26 *** Rubicon-|-Cross has joined #aichallenge 2010-11-08T09:57:09 *** Utkarsh has joined #aichallenge 2010-11-08T09:59:51 *** yasith has joined #aichallenge 2010-11-08T10:00:31 you probably want the playgame.jar rewrite instead - iirc it just does command line resolution 2010-11-08T10:01:12 hey, just wondering if there's a way to find matches between a set of players, without scraping the site. It's a small set so I guess actually it wouldn't take a lot of scraping.. 2010-11-08T10:02:37 well you'd have to parse each player's 'played game' list 2010-11-08T10:02:49 beyond that not much 2010-11-08T10:04:12 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-11-08T10:05:51 Is there a way we can call JBotmanager.jar via commandline and receive game winner via commandline? 2010-11-08T10:08:52 ananda_: you probably want the playgame.jar rewrite instead - iirc it just does command line resolution 2010-11-08T10:09:37 *** Utkarsh has joined #aichallenge 2010-11-08T10:09:44 *** superflit has quit IRC (Quit: superflit) 2010-11-08T10:10:24 *** tapwater has joined #aichallenge 2010-11-08T10:11:19 demerzel_: How can I get the source files of playgame.jar? 2010-11-08T10:12:07 @contest 2010-11-08T10:12:08 raedwulf: You have no gotten any error messages recently, so here's a random one just to let you know that we care. 2010-11-08T10:12:14 hmmm what was the source agian :/ 2010-11-08T10:12:14 *** yasith_ has joined #aichallenge 2010-11-08T10:12:45 http://code.google.com/p/ai-contest/ 2010-11-08T10:13:49 *** yasith has quit IRC (Ping timeout: 265 seconds) 2010-11-08T10:14:51 ananda_:the jar has a src folder iirc - why do you want source? somebody on the forum actually rewrote it 2010-11-08T10:15:37 to fix bugs etc 2010-11-08T10:15:48 the one in the starter pack is a bit buggy 2010-11-08T10:15:53 Top 10 players: bocsimacko(4237), shangas(4056), bix0r4ever(4053), george(4041), rsergio(3982), GreenTea(3971), Accoun(3955), Polsky(3948), Xrillo(3940), dmj111(3928) 2010-11-08T10:16:03 oh yes! 2010-11-08T10:16:27 demerzel_: thanks a lot for the info. So I've to download a source that is called "iirc" right? 2010-11-08T10:16:34 demerzel_: Thanks! I'll do it 2010-11-08T10:16:48 oh iirc = acronym for 'if i remember correctly' 2010-11-08T10:17:02 there is a folder named "src" in the jar itself 2010-11-08T10:17:06 jars are just zip files 2010-11-08T10:17:22 IF i remember correctly if not i'm sure it's on the googlecode site 2010-11-08T10:17:32 Oh sorry. Thanks for the clarification 2010-11-08T10:17:38 but i would start with the rewritten version if possible 2010-11-08T10:17:47 Sure, thanks a lot! 2010-11-08T10:17:50 np 2010-11-08T10:21:23 demerzel_: I extracted the jar file and it shows many files named name.class. I actually want a jar package that does not check for wall time. Because I'm running on a ThinClient, this is required 2010-11-08T10:22:01 demerzel_: Could you tell me if anybody changed the Playgame.jar to stop checking for timout from wall time? 2010-11-08T10:22:07 *** JamesMG has joined #aichallenge 2010-11-08T10:27:08 ananda_: check this thread - http://ai-contest.com/forum/viewtopic.php?f=18&t=534 also i know there's an option to turn off timeouts but i haven't used it in a long time 2010-11-08T10:29:27 *** Frontier has joined #aichallenge 2010-11-08T10:31:40 *** ananda_ has quit IRC (Ping timeout: 265 seconds) 2010-11-08T10:41:14 haha, who is "accoun_stop_running_so_many_bots" on TCP? 2010-11-08T10:43:18 *** superflit has joined #aichallenge 2010-11-08T10:43:38 lol :D 2010-11-08T10:45:12 isn't accoun usually fairly mild? 2010-11-08T10:45:18 admitidly i haven't been watching for a few days 2010-11-08T10:45:20 mild in terms of what 2010-11-08T10:45:30 compared to some others 2010-11-08T10:45:39 he's been really active the past day 2010-11-08T10:45:47 http://72.44.46.68/getplayer?player=JamesMG.fun 2010-11-08T10:45:49 see this for evidence 2010-11-08T10:45:50 oh you mean in terms of instancing bot versions? 2010-11-08T10:45:50 wtf? are you talking about me mother fucker huh? 2010-11-08T10:45:58 just kidding 2010-11-08T10:46:38 i guess my bot isn't totally retarded it lasted 200 turns with him 2010-11-08T10:46:44 James, at least you win against XYZ :D 2010-11-08T10:46:45 i'm debating how much restriction i put on moves 2010-11-08T10:47:01 ahh okay, well i'll be stupidly active in a day or so till the end of the contest 2010-11-08T10:47:03 that's my aborted minimax attempt Frontier 2010-11-08T10:47:05 oh wait no it's still retarded haha 2010-11-08T10:47:13 i just had a 0 growth 2010-11-08T10:47:48 I keep it running because it is kind of entertaining, and I plan to get back to it eventually 2010-11-08T10:47:52 demerzel_: My bot gets a divide by 0 when it sees a 0-growth, so i must attack it last 2010-11-08T10:48:24 damn that's weird 2010-11-08T10:48:25 http://72.44.46.68/canvas?game_id=911846 2010-11-08T10:48:32 he has a 20 ships advantage at turn 51 2010-11-08T10:48:42 i am using a pure greed 4-ply move look 2010-11-08T10:48:46 how the hell 2010-11-08T10:49:02 we haven't touched each other yet 2010-11-08T10:49:19 pure greed should be the optimal on just neutral acq 2010-11-08T10:49:22 just another bot who doesnt like 0 growth ^^ 2010-11-08T10:49:31 Hey, Accoun, could you please not have that many bots on the TCP server.. it's kinda boring when everybody plays your bot all the time. :) 2010-11-08T10:49:50 hey Xrillo i don't know who you are but can you not tell other people your opinion on what they're doing unless someone cares 2010-11-08T10:50:16 lol 2010-11-08T10:50:19 well, I though people care. :) 2010-11-08T10:50:35 no, it's rather rude and presumptous 2010-11-08T10:50:36 :) 2010-11-08T10:50:55 whatever. :) 2010-11-08T10:50:56 what is a pure greed algorithm? 2010-11-08T10:51:07 where you don't care what the enemy does 2010-11-08T10:51:10 guessing it's just shooting for highest ship count by turn 200? 2010-11-08T10:51:12 did you watch Wallstreet 2? 2010-11-08T10:51:12 ie no level for enemy move 2010-11-08T10:51:45 ok, so - what i do. but turn 200? seriously? 2010-11-08T10:51:47 only thing i can think is he's either checking every individual node for highest payout or has a higher forecast horizon 2010-11-08T10:51:57 demerzel_: how many nodes do you have in your plys? 2010-11-08T10:52:04 for neutral acq it's a 4-ply 2010-11-08T10:52:12 because i don't have to look at his moves 2010-11-08T10:52:24 i am stumped on how he got a better setup than me 2010-11-08T10:52:51 wow I have some buggy code: http://72.44.46.68/canvas?game_id=911876 2010-11-08T10:52:55 4-ply? does that mean 4 turns? or 4 acquisitions? 2010-11-08T10:52:59 4 acq 2010-11-08T10:53:09 lol 2010-11-08T10:53:56 JamesMG: pro tip: if you don't get a planet - send 20% more next time 2010-11-08T10:54:02 damn i got some weird algorithms i guess 2010-11-08T10:54:17 one very unlikely possibility is that he solved this with >4ply offline and then has a lookup table for all the maps ;) 2010-11-08T10:54:33 i dunno what i'm doing wrong but bogusbrainpower destroyed it completely 2010-11-08T10:54:39 http://72.44.46.68/canvas?game_id=911900 turn 19 he has over 100 ship adv 2010-11-08T10:54:53 err 18 when we were both /32 2010-11-08T10:54:56 that bug is caused by my "how many ships can he possibly have at this planet" function only looking a few turns out when my planet is that close 2010-11-08T10:55:08 my algo must be wrong somehow but idk how it's looking at a modified gamestate 2010-11-08T10:55:14 pretty simple really 2010-11-08T10:56:06 maybe i need to look per node level or force it to followthrough on plans 2010-11-08T10:56:25 i figured if move #4 is optimal now if i do moves #1-3 and recalc next turn move #4 will still be optimal but apparently not 2010-11-08T10:57:10 *** Commandir has joined #aichallenge 2010-11-08T10:57:30 does it factor in ships gained by the neutrals it has attacked? 2010-11-08T10:57:37 yes absolutely 2010-11-08T10:57:42 that's the point of carrying around a modified gamestate 2010-11-08T10:57:53 That is one ugly situation, when your bot gives up on a target and gets really confused about what to attack every other turn 2010-11-08T10:58:16 i probably just have some weird restriction left over from testing or something 2010-11-08T10:58:27 *** boegel has quit IRC (Ping timeout: 276 seconds) 2010-11-08T10:58:29 there's no way i should be losing 100 ship advantage on neutrals - i expect to get destroyed after 2010-11-08T10:59:09 demerzel: rude and presumptuous to politely ask someone something? as opposed to what you did afterwards? ffs :D 2010-11-08T10:59:11 phreeza: i hope lookup tables really don't happen 2010-11-08T10:59:30 flufffluff1: he was being an arrogant ass 2010-11-08T10:59:35 not demerzel_ 2010-11-08T10:59:37 fluffluff1: when you run the tcp server you can ask people to stop playing bots, when you are just using it and ask a top 10 player to stop running bots you are being an ass 2010-11-08T10:59:45 lol 2010-11-08T11:00:04 *** tapwater_ has joined #aichallenge 2010-11-08T11:00:19 *looks at contestbots listing* 2010-11-08T11:00:19 !rank 10 2010-11-08T11:00:22 !top 10 2010-11-08T11:00:27 well.. anyway :D 2010-11-08T11:00:32 @rankings 2010-11-08T11:00:32 oh boy he might not be in there at this very moment - point disproved 2010-11-08T11:00:33 JamesMG: Top 10 players: bocsimacko(4227), shangas(4042), bix0r4ever(4030), george(4012), GreenTea(3956), rsergio(3955), Accoun(3937), Xrillo(3932), dmj111(3918), Polsky(3916) 2010-11-08T11:00:39 oh he is there haha 2010-11-08T11:00:48 so is Xrillo 2010-11-08T11:00:49 oh snap, he's one position ahead of me SO SORRY 2010-11-08T11:00:53 great 2010-11-08T11:00:55 congrats 2010-11-08T11:01:05 wtf are you guys talking about rankings for? 2010-11-08T11:01:10 it's entirely irrelevant 2010-11-08T11:01:14 maybe ask dhartmei if you can have a special pipe for not playing accoun for being special :) 2010-11-08T11:01:19 there weren't that many bots being run by him 2010-11-08T11:01:33 think of it as a chance to optimize your bot to beat his 2010-11-08T11:01:34 Xrillo: people have been running multiple bots daily the entire length of the contest 2010-11-08T11:02:08 accoun is pretty mild - i remember playing 25 versions of vortex 2010-11-08T11:02:12 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T11:02:16 yes, I know.. but it gets a bit frustrating when my bot spends all day playing one single bot 95% of the games. 2010-11-08T11:02:36 *** tapwater has quit IRC (Ping timeout: 245 seconds) 2010-11-08T11:02:37 *** tapwater_ is now known as tapwater 2010-11-08T11:02:44 if you are on autorepeat and he is too then you're likely to play against each other again 2010-11-08T11:02:45 Xrillo: agreed, should i put my old bot in the queue, just for fun? 2010-11-08T11:02:51 the alternative is he switched to one username but runs as many instances at once, you're still going to get paired as often 2010-11-08T11:04:12 well, sorry to disturb you guys, you're clearly not seeing the problem.. keep on discussing your bots again. 2010-11-08T11:05:03 the problem to me sounds like there wasn't many people running their bots 2010-11-08T11:05:10 ah wow Monty is top 10 now 2010-11-08T11:05:22 now i know who you are 2010-11-08T11:05:29 antimatroid: read the lines again. it's pretty apparent who the arrogant ass was 2010-11-08T11:05:37 Xrillo: I get matched with Accoun - what a surprise ^^ 2010-11-08T11:06:27 :) 2010-11-08T11:07:00 who is p4p4? http://72.44.46.68/getplayer?player=frontier 2010-11-08T11:07:28 front: he's on here occasionally under that name 2010-11-08T11:07:54 *** rebelxt has joined #aichallenge 2010-11-08T11:08:02 *** Utkarsh has joined #aichallenge 2010-11-08T11:08:19 oh, right now in fact :) 2010-11-08T11:08:28 hi 2010-11-08T11:09:07 Ok, I'll keep my but running for a while. It isn't very good, but ... hi p4p4 ... it makes for some diversity :) 2010-11-08T11:09:56 p4p4: I could have lost when you took my starting planet. Oo 2010-11-08T11:10:43 im getting used to be beaten up .. 2010-11-08T11:12:49 im use only 8 copies 2010-11-08T11:12:59 to math? 2010-11-08T11:14:08 please run more copies of u bots to get more games 2010-11-08T11:15:30 Accoun, are you running 8 bots at the same time or alternating? 2010-11-08T11:15:52 Top 10 players: bocsimacko(4220), shangas(4035), bix0r4ever(4024), george(4005), GreenTea(3950), rsergio(3949), Accoun(3931), Xrillo(3921), dmj111(3912), Polsky(3912) 2010-11-08T11:16:39 8 bots at the same time 2010-11-08T11:16:51 8 copies of bot 2010-11-08T11:17:08 wtf?! 2010-11-08T11:17:15 for fast get reiting 2010-11-08T11:17:37 to math? 2010-11-08T11:17:50 it was quite boring to to test on tcp because of that 2010-11-08T11:18:03 getting paired to one or another of those in 70% of all games totally wracked my elo. 2010-11-08T11:18:04 if all of would do that then fine 2010-11-08T11:18:12 Accoun, you may want to alternate.. we were getting way too many games against your bots for that reason 2010-11-08T11:18:13 run 8 copies like me 2010-11-08T11:18:14 *all of us 2010-11-08T11:18:23 Accoun: we can't 2010-11-08T11:18:31 or at least some of us can't 2010-11-08T11:18:33 nor can I. 2010-11-08T11:18:34 Accoun: actually i made a tiny java program that run X games of my bot in parallel 2010-11-08T11:18:34 3 copies can? 2010-11-08T11:18:38 and N games in total 2010-11-08T11:18:59 Accoun: I think 3 should be a good compromise 2010-11-08T11:19:03 this way, i play with different people and many games in short time 2010-11-08T11:19:06 3 copies would be much better.. 8 bots is an overkill for sure 2010-11-08T11:19:20 rebelxt: overkill for what ? server isnt that bad 2010-11-08T11:19:28 i run 10 bots in parallel and its all fine 2010-11-08T11:19:28 on 4 cores CPU anyone can run 3 copies 2010-11-08T11:19:52 it has nothing to do with cores 2010-11-08T11:19:55 nullkuhl: overkill from other people's perspective.. i want to get more games against other people, not just accoun's bots ;-) 2010-11-08T11:19:59 i have 2 cores only and i run 10 versions 2010-11-08T11:20:38 Ah, so I wasn't the only to think of it this way, eh? :) 2010-11-08T11:21:24 i think, there's just not enough other players there ! now: 25 connections, 8 of them being account 2010-11-08T11:22:34 maybe need to do some advertisement for tcp 2010-11-08T11:23:59 man my neutral selection blows but i have no idea why 2010-11-08T11:24:01 it's really easy to write a script which alternates games for all of your bots, i.e. v1, v2, v3, v1, v2, v3, etc this sounds very reasonable 2010-11-08T11:25:15 5 conections is ok? 2010-11-08T11:27:59 any number of concurrent connections is OK since TCP server owner (dhartmei) does not enforce any rules. i personally feel 3 connections is a good compromise 2010-11-08T11:28:02 even 8 are ok for me, we've just got to wake up those sleepers. 2010-11-08T11:28:31 *** Commandir has left #aichallenge 2010-11-08T11:29:55 what I would like to have on tcp is a way to prevent different versions of my bot playing each other 2010-11-08T11:30:08 *** Florian95 has left #aichallenge 2010-11-08T11:31:49 dhartmei is likely paying ~60-80$ per month for the EC2 server out of his pocket.. very nice of him! 2010-11-08T11:32:22 if somebody wants to write php i'd put up a version haha 2010-11-08T11:32:28 i do not want to dick with php if i can help it 2010-11-08T11:32:59 ~60-80$ per month? overprice 2010-11-08T11:33:13 *** bhasker_ has quit IRC (Quit: bhasker_) 2010-11-08T11:33:43 virtual server can enougth 2010-11-08T11:33:57 i got a vps with unmetered and 2gig ram but it's linux so gotta be php :( 2010-11-08T11:34:04 for ~25 connections 2010-11-08T11:35:49 Accoun: yeah, he is running a VPS on amazon EC2.. he might have switched to a reserved instance since it's much cheaper. this contest would not be fun without TCP server 2010-11-08T11:36:27 Accoun: TestingFromPiter is not one of your bots, is it? 2010-11-08T11:36:41 @seen dhartmei 2010-11-08T11:36:41 phreeza: dhartmei was last seen in #aichallenge 5 weeks, 4 days, 23 hours, 52 minutes, and 3 seconds ago: project selection problem? 2010-11-08T11:38:06 phreeza: i would not be surprised if dhartmei cranks out a strong bot right before the deadline 2010-11-08T11:38:35 did he do well in tron? 2010-11-08T11:39:46 he was in top100 for sure 2010-11-08T11:40:10 http://csclub.uwaterloo.ca/contest/rankings.php 2010-11-08T11:41:33 *** vall has quit IRC (Quit: Page closed) 2010-11-08T11:43:14 demerzel_: regarding neutral selection issues - what's your scoring function like? max growth or max ship count increase within a number of turns? or something more complex? 2010-11-08T11:46:18 the latter 2010-11-08T11:49:28 demerzel_: so what are the issues you are having? greedy neutral selection algorithm is supposed to work pretty well.. of course the ideal is enumerating all possible neutral capture sequences and picking the best sequence (similar to traveling salesman problem). 2010-11-08T11:50:09 yes i do that - i am not sure what the issue is - might be implementation, i have huge gaps in neutral acq vs enemy 2010-11-08T11:50:17 should be close to even at least i am digging through the code ;) 2010-11-08T11:51:48 yeah probably just a bug 2010-11-08T11:51:57 *** lavalamp__ has joined #aichallenge 2010-11-08T12:01:32 *** fawek has joined #aichallenge 2010-11-08T12:03:00 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-11-08T12:03:32 *** rebelxt has left #aichallenge 2010-11-08T12:09:06 *** Utkarsh has joined #aichallenge 2010-11-08T12:15:54 Top 10 players: bocsimacko(4220), shangas(4038), bix0r4ever(4027), george(4009), GreenTea(3955), rsergio(3952), Accoun(3936), Xrillo(3930), dmj111(3918), Polsky(3916) 2010-11-08T12:18:53 *** Bobng has joined #aichallenge 2010-11-08T12:23:42 mega1: I just have to ask.. you wouldn't happen to be a Go-programmer? I'm just getting that feeling.. ( not the language, the game ) 2010-11-08T12:24:21 Xrillo: Hex programmer, but I stopped years ago 2010-11-08T12:24:40 http://en.wikipedia.org/wiki/Hex_%28board_game%29 2010-11-08T12:24:48 a bit similar to go 2010-11-08T12:24:58 yea, Hex is also an interesting game. :) 2010-11-08T12:25:39 Xrillo, I play go and my bot is written in go... ;) 2010-11-08T12:25:51 :D 2010-11-08T12:26:09 fellow Go player, good.. :) 2010-11-08T12:31:04 *** mceier has joined #aichallenge 2010-11-08T12:32:31 hex has ai? isn't that the game where player 1 can always win by just copying player 2? 2010-11-08T12:32:52 thought that was a nash thing 2010-11-08T12:33:22 hex uses a pie rule to defeat that strategy 2010-11-08T12:33:47 IOW player a makes the first move and then player b decides which color they want to play 2010-11-08T12:34:42 so why couldn't b just use the same advantage 2010-11-08T12:35:14 because A won't play in the center (the starting move which is necessary for that strategy) 2010-11-08T12:35:37 ah ok 2010-11-08T12:35:55 A has to find a move that's good but not TOO good :) 2010-11-08T12:35:57 i didn't know it was requisite to a certain move 2010-11-08T12:36:50 I've never played the game but that's the only way I can imagine that strategy working out 2010-11-08T12:38:48 I can't imagine why Nash thought Hex was superior to Go... 2010-11-08T12:40:04 i am trying to remember what game it is where you have a hexagon and you start out on one side and want to get to the other - is that hex? was called chinese stones or some crap 2010-11-08T12:40:12 fuzzy memory lol 2010-11-08T12:40:23 chinese checkers starts on a hexagonal star 2010-11-08T12:40:32 yup that's it 2010-11-08T12:40:50 *** bhasker has joined #aichallenge 2010-11-08T12:41:00 that'd be a cool game to ai for i wonder if it's been solved - i doubt it 2010-11-08T12:42:39 i might have to get into go after this, spent wayyy too much time learning about game theory etc to just drop it after it's over 2010-11-08T12:44:33 how about poker for the next contest 2010-11-08T12:45:14 demerzel_, learning to play or writing an AI for it? 2010-11-08T12:45:27 the latter 2010-11-08T12:45:33 i think everyone knows how to play it, probably not well ;) 2010-11-08T12:47:32 lavalamp__: may i ask what the krumm means in your bot name? 2010-11-08T12:47:40 german? 2010-11-08T12:47:56 phreeza: if you can figure out what "gromble" means you'll be mostly there :) 2010-11-08T12:48:04 not german, no ^^ 2010-11-08T12:48:08 hehe 2010-11-08T12:48:42 Aaah! Real Monsters 2010-11-08T12:48:44 lol 2010-11-08T12:48:47 ok 2010-11-08T12:48:58 demerzel_, most people haven't even heard of it... I play at a club at a coffee shop and explain it to a lot of passerbys... 2010-11-08T12:49:04 phreeza: yup 2010-11-08T12:49:30 I was wondering if anyone would ask me, hehe 2010-11-08T12:49:37 :) 2010-11-08T12:49:44 Krumm means crooked in gemran 2010-11-08T12:49:47 german* 2010-11-08T12:49:55 haha 2010-11-08T12:50:47 lavalamp_: most people know it as othello/reversi not as go - granted they are different but rules are similar 2010-11-08T12:51:07 *head explodes* 2010-11-08T12:51:19 very very different games... 2010-11-08T12:51:29 othello/reversi is NOTHING like go 2010-11-08T12:51:34 not flipping :P 2010-11-08T12:51:39 the goal is similar 2010-11-08T12:51:45 to win :) 2010-11-08T12:51:47 similar boards 2010-11-08T12:51:55 phreeza: no 2010-11-08T12:52:06 no? 2010-11-08T12:52:06 othello you play in the squares, go the intersections 2010-11-08T12:52:17 also a go board is 19x19 2010-11-08T12:52:22 othello is 8x8 2010-11-08T12:52:23 hehe does it matter? 2010-11-08T12:52:31 gomoku is more similar 2010-11-08T12:52:42 Naxtibalda: to which? 2010-11-08T12:52:46 if somebody doesn't know go i doubt they will know gomoku 2010-11-08T12:53:02 gomoku is more commonly known as five-in-a-row 2010-11-08T12:53:20 it at least really is played on the same board 2010-11-08T12:53:26 it's a square board with black/white pieces - one flips, the other captures -- all you need to know to get started 2010-11-08T12:53:34 in any of them 2010-11-08T12:53:49 ... 2010-11-08T12:54:24 what there are more complexities than that? i had no idea :P 2010-11-08T12:54:37 othello has no ko rule 2010-11-08T12:54:56 because it's not possible 2010-11-08T12:55:00 yes 2010-11-08T12:55:37 there are a ton of differences obviously i was referring to your 'nobody knows what it is' comment, what i would say to that person 2010-11-08T12:55:55 if you overwhelm them with the deep and rich history of ko rules and go and how it is DEFINITELY NOT like other games you will jsut turn them off lol 2010-11-08T12:56:36 ah yes actually I do say the goal is similar to othello 2010-11-08T12:57:07 I have it down to a few sentences because that's all most people have patience for 2010-11-08T12:57:19 but it seems to me othello is only slightly more like go than american checkers is like chess 2010-11-08T12:58:02 janzert: see my head exploding earlier 2010-11-08T12:58:56 ok people bbl 2010-11-08T13:01:34 bhasker: got new stuff yet? 2010-11-08T13:01:50 nope 2010-11-08T13:01:57 my current version is horrible 2010-11-08T13:01:58 :-\ 2010-11-08T13:02:00 mmm 2010-11-08T13:02:06 how come? 2010-11-08T13:04:47 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2010-11-08T13:06:40 reversi with single line choice would be interesting 2010-11-08T13:08:42 *** kjdf has joined #aichallenge 2010-11-08T13:09:35 would it be possible to change timout rule to 200s per game instead 1s per turn? 2010-11-08T13:09:42 much friendlier to garbage collected languages 2010-11-08T13:09:59 i've got a new version soon i hope which will be epicness 2010-11-08T13:11:31 *** bhasker_ has joined #aichallenge 2010-11-08T13:14:36 for the next contest, i want planetwars with >2 players and with diplomacy :D 2010-11-08T13:15:04 *** bhasker has quit IRC (Ping timeout: 265 seconds) 2010-11-08T13:15:06 *** Apophis has joined #aichallenge 2010-11-08T13:15:54 Top 10 players: bocsimacko(4214), shangas(4028), bix0r4ever(4025), george(3986), rsergio(3958), GreenTea(3948), Accoun(3915), dmj111(3903), Xrillo(3902), Polsky(3899) 2010-11-08T13:20:29 *** perror has quit IRC (Quit: Bye all !) 2010-11-08T13:22:56 meen 2010-11-08T13:23:03 im almost ready :P 2010-11-08T13:29:54 phreeza : speaking of which, 'diplomacy' is another good ai challenge. But Hasbro own the rights. 2010-11-08T13:32:01 Guest13825: yes, but a similar game would be great 2010-11-08T13:33:15 *** boegel has joined #aichallenge 2010-11-08T13:33:25 *** McLeopold has joined #aichallenge 2010-11-08T13:34:22 i just don't get it why people doesn't compete in ai-poker game 2010-11-08T13:34:31 is it somekind of ... 'banned' thing or something 2010-11-08T13:34:52 no one doesn't want to make a good poker bot or something.. 2010-11-08T13:36:11 rabidus: what gave you that idea?? 2010-11-08T13:36:47 theres a world championship for poker bots 2010-11-08T13:37:17 ??? what? 2010-11-08T13:37:21 somekind of playground? 2010-11-08T13:37:41 umm.. i mean somekind of 'tcp-server' a-like solution? 2010-11-08T13:38:07 yea there is a server like that 2010-11-08T13:38:12 i played with it for a while 2010-11-08T13:38:16 let me find it 2010-11-08T13:38:25 wow, i didn't knew that. 2010-11-08T13:38:36 i found myself some ... 'challenges' but they were all dead 2010-11-08T13:43:51 i can't find the public server i ran my bot on back then... but this looks good http://www.computerpokercompetition.org/ 2010-11-08T13:43:52 heh, if we were to have a pokerbot ai contest i wonder how many people would feel confident enough in their bots to gamble real money with them :P 2010-11-08T13:44:08 doubtful that it would even be legal in the US 2010-11-08T13:44:15 (stupid) 2010-11-08T13:44:20 *** boegel_ has joined #aichallenge 2010-11-08T13:44:29 hey, this is internet. run your bot on some other country 2010-11-08T13:44:30 :) 2010-11-08T13:44:32 *** boegel has quit IRC (Ping timeout: 252 seconds) 2010-11-08T13:44:41 borrow someones id :) 2010-11-08T13:46:46 I think a poker tournament would be quite good. 2010-11-08T13:49:06 too much luck involved 2010-11-08T13:52:58 mega1 are you still around? curious what your branching factor is with 2ply - gotta be nuts 2010-11-08T13:53:26 i am still fighting my stupid tree because it keeps hooking the same move :( 2010-11-08T13:53:34 not always though so not obvious to catch lol 2010-11-08T13:53:38 I think the branching factor is about a 100 2010-11-08T13:53:45 that's it? 2010-11-08T13:53:56 dunno 2010-11-08T13:53:57 damn i would have expected way more 2010-11-08T13:54:22 consider 10 moves apiece you've got 20 separate scenarios with 200 possibilities apiece 2010-11-08T13:54:22 depends on parameters 2010-11-08T13:54:40 obviously I don't generate all of them 2010-11-08T13:54:58 what do you consider a move if you don't mind me asking? i've limited mine to capture / defending 2010-11-08T13:55:11 still debating if i should put the restrictions i do on my captures on enemy projections 2010-11-08T13:56:02 a move is set of orders not necessarily to be carried out on the same turn 2010-11-08T13:56:06 *is a 2010-11-08T13:56:31 does your eval account for the turn or do you cap the horizon? 2010-11-08T13:56:45 *** fawek_ has joined #aichallenge 2010-11-08T13:57:07 *** rangzen has joined #aichallenge 2010-11-08T13:57:09 i see now why you don't mind talking it's insane to implement bug-free 2010-11-08T13:57:10 lol 2010-11-08T13:57:14 what does "account for the turn" mean? 2010-11-08T13:57:33 well a move that isn't available till turn 8 may not be the 'best move' if you're only looking to turn 20 2010-11-08T13:57:42 *** choas has joined #aichallenge 2010-11-08T13:57:49 so either you're capping it further out or putting some kind of turn modifier in 2010-11-08T13:57:53 yes, I cap the horizon 2010-11-08T13:58:25 there is no point in simulating a game for 200 turns when the uncertainty is too high 2010-11-08T13:58:33 your bot inspired me to do the full on tree search btw ;) 2010-11-08T13:58:39 yea that's true 2010-11-08T13:58:43 i think my horizon is too low 2010-11-08T13:59:04 I've been using the maximum distance on the board as the horizon 2010-11-08T13:59:09 same 2010-11-08T13:59:09 ... which can prevent you from getting expensive planets 2010-11-08T13:59:23 *** fawek has quit IRC (Ping timeout: 252 seconds) 2010-11-08T13:59:30 yea i kept passing up a 65/5 on a short map 2010-11-08T14:00:30 somebody theorized you were using book moves mega1 - is that true? 2010-11-08T14:00:53 for the record i doubt it but i am curious 2010-11-08T14:01:17 * mega1 considers misleading #aichallenge 2010-11-08T14:01:40 haha even if you spelled out your entire logic i doubt many of us could actually implement it 2010-11-08T14:01:42 lol 2010-11-08T14:01:42 no, I don't 2010-11-08T14:01:44 this tree stuff is nightmarish 2010-11-08T14:02:12 to debug that is, implement fairly easy haha 2010-11-08T14:02:23 book moves are worthless come the final tournament if there is no generalization 2010-11-08T14:02:29 hm I think I need to rethink my move generation 2010-11-08T14:02:52 mega1: yea that was what i was thinking too but also that i doubt they will make new maps 2010-11-08T14:03:04 i still doubt the usefulness of book moves though 2010-11-08T14:03:24 I hope they will. 2010-11-08T14:03:39 I think I've been using my tree to try and solve the wrong problems... 2010-11-08T14:03:42 creating opening books is boring 2010-11-08T14:03:45 amstan what's the word on that 2010-11-08T14:04:01 re: creating new maps for final tournament to avoid book openings 2010-11-08T14:04:20 we will 2010-11-08T14:04:28 oh cool i didn't know it was official 2010-11-08T14:04:40 i guess i can unhook the database then haah 2010-11-08T14:05:01 amstan: and will you be using the same generation script? (i.e., always 23 planets?) 2010-11-08T14:05:12 or at the very least <=23 haha 2010-11-08T14:05:26 probably just changing the seed 2010-11-08T14:05:30 otherwise Demerzel crashed / timed out / did not start. 2010-11-08T14:05:30 amstan: would it be possible to change timout rule to 200s per game instead 1s per turn? this would make it much friendlier for garbage collected languages 2010-11-08T14:05:55 amstan ok that makes sense 2010-11-08T14:06:09 sounds good 2010-11-08T14:06:19 did you guys have any idea how much effort bots in this would be vs. the last contest? 2010-11-08T14:06:28 it's pretty large effort, much more than i originally anticipated 2010-11-08T14:06:31 kjdf: almost certainly won't happen for this tournament, also it causes other problems 2010-11-08T14:06:50 i just used static everything so gc wasn't an issue 2010-11-08T14:07:02 not like you need to destroy/recreat 2010-11-08T14:07:08 demerzel_ C doesn't even have GC 2010-11-08T14:07:12 kjdf: that just means a program can spin for 200 seconds without even giving a first turn 2010-11-08T14:07:19 yes my official entry is C# tho :P 2010-11-08T14:07:45 yes but then it will have to think fast or lose... 2010-11-08T14:07:46 until i get this damn tree debugged 2010-11-08T14:08:39 kjdf: no.. that means that if someone submits an infinite loop bot, they'll be given 200 seconds for nothing.. 2010-11-08T14:08:58 you could institute a hard limit per turn just higher 2010-11-08T14:09:03 ie 200s/5s per turn 2010-11-08T14:09:20 that just makes things unnecessarily complicated 2010-11-08T14:09:24 personally though i think people just need to learn to code without destroying objects until they're not needed any more ;) 2010-11-08T14:09:30 *** fawek_ is now known as fawek 2010-11-08T14:09:43 if you picked a garbage collected language, you have to realize the disadvantages of using them 2010-11-08T14:09:48 demerzel_: hard to do that in functional languages 2010-11-08T14:10:06 kjdf: functional like what? i didn't destroy a single object in my C# implementation 2010-11-08T14:10:14 like lisp 2010-11-08T14:10:27 ah well i don't know lisp but i'm sure mega1 can answer that haha 2010-11-08T14:10:42 amstan ok i get it :) I will just set harder timeout limits then 2010-11-08T14:13:29 *** RainCT has joined #aichallenge 2010-11-08T14:14:26 kjdf, you could try manually triggering your GC right after you finish your turn-- I plan to do that if it becomes an issue 2010-11-08T14:14:53 *** Migi32 has joined #aichallenge 2010-11-08T14:15:17 that might cause more problems than its worth 2010-11-08T14:15:22 lavalamp__: that seems like a good idea. 2010-11-08T14:15:30 not sure what lisp's gc model is but it costs to call gc 2010-11-08T14:15:50 demerzel_: hence you call it right after your turn. before it's your turn again. 2010-11-08T14:15:56 Top 10 players: bocsimacko(4211), shangas(4030), bix0r4ever(4012), george(3985), rsergio(3948), GreenTea(3936), Accoun(3899), dmj111(3895), Xrillo(3891), medrimonia(3876) 2010-11-08T14:16:06 brx but clock starts while you are doing gc 2010-11-08T14:16:13 and then you don't even know how much time you have left 2010-11-08T14:16:42 hm that might be a problem... but you can time the GC 2010-11-08T14:16:46 kjdf: that is possible. 2010-11-08T14:16:50 mh 2010-11-08T14:17:28 you can probably even get away with only running the GC every n turns 2010-11-08T14:17:33 how do you even know gc is a problem 2010-11-08T14:17:36 * brx nods at jmcarthur 2010-11-08T14:17:39 i would make sure it is before implementing anything like that 2010-11-08T14:17:46 yeah really 2010-11-08T14:17:54 demerzel_: gc can randomly spin up and throw you over the timeout threshold. 2010-11-08T14:18:01 theoretically yes 2010-11-08T14:18:08 but if you're dealing with a bunch of tiny ints it probably won't 2010-11-08T14:18:16 plus if it's just clearing tiny ints it's probably already as quick as it can be 2010-11-08T14:18:22 it depends on the allocation model 2010-11-08T14:18:59 i'm not saying it's not possible i'm just saying manually playing with the gc is a rare solution to a rare problem 2010-11-08T14:19:10 it's not rare 2010-11-08T14:19:10 i would make sure i have the problem first 2010-11-08T14:19:18 but i would still make sure i had the problem, as you say 2010-11-08T14:19:25 maybe not in lisp idk 2010-11-08T14:19:44 in lisp it shouldn't be a problem. 2010-11-08T14:19:46 it's not rare in most runtimes that i know of 2010-11-08T14:19:48 in java/.net you don't wanna mess with the gc 2010-11-08T14:19:56 just program more efficiently 2010-11-08T14:20:17 ah, if we exclude GC-unfriendly code then you're right, it's rare 2010-11-08T14:20:31 lol 2010-11-08T14:20:44 but GC-unfriendly code is quite common 2010-11-08T14:20:53 lern2code ;) 2010-11-08T14:21:20 honestly i actually had the same concern with .net 2010-11-08T14:21:23 i set up a gc run per turn 2010-11-08T14:21:30 ended up costing a ton more than it needed 2010-11-08T14:21:48 mark and sweep GC in an impure language really sucks when you have deep trees 2010-11-08T14:21:51 gc on first turn is an ok deal 2010-11-08T14:21:55 during warmup 2010-11-08T14:21:58 i wouldn't do it after that 2010-11-08T14:22:33 ah i went to the length of actually reusing static vars rather than local 2010-11-08T14:22:38 in .net 2010-11-08T14:22:45 yuck 2010-11-08T14:23:40 well in nodes it makes sense 2010-11-08T14:23:48 you're going to have a lot of cleanup needed if you don't reuse 2010-11-08T14:24:05 or byref 2010-11-08T14:24:47 "in nodes"? 2010-11-08T14:26:22 with nodes? 2010-11-08T14:27:08 what nodes? 2010-11-08T14:27:33 "the" nodes 2010-11-08T14:27:33 "node" is a very generic word and I think i'm not clear on the context 2010-11-08T14:27:40 with programming 2010-11-08T14:27:42 and nodes 2010-11-08T14:27:43 yes? 2010-11-08T14:27:50 like a language 2010-11-08T14:27:50 even in programming is pretty generic 2010-11-08T14:28:00 you make words and calculate 2010-11-08T14:28:02 a language called nodes? 2010-11-08T14:28:06 nodeOS 2010-11-08T14:28:25 .js 2010-11-08T14:28:37 that reminds me, antimatroid1, has the bayeselo question been cleared up already? 2010-11-08T14:28:47 brx: yea it has 2010-11-08T14:28:55 to what end? 2010-11-08T14:29:06 whatever version they got already had the source modified 2010-11-08T14:29:10 as in node.js? 2010-11-08T14:29:15 ah, good! 2010-11-08T14:29:25 jmcarthur: i am just screwing with you i don't even know what we're talking about 2010-11-08T14:29:31 *sigh* 2010-11-08T14:29:50 okay but what did you originally mean? 2010-11-08T14:29:57 "well in nodes it makes sense" 2010-11-08T14:30:12 if you are making a search tree and creating nodes locally and destroying them that is bad 2010-11-08T14:30:22 ah 2010-11-08T14:30:28 well, not necessarily 2010-11-08T14:30:39 sounds like something that would be bad in java or .net 2010-11-08T14:30:42 + if you want to avoid gc 2010-11-08T14:31:57 actually, even in java that doesn't sound like it would be very mean to the GC 2010-11-08T14:32:05 i'm not familiar enough with how .net's gc works 2010-11-08T14:32:41 my understanding is that java's gc should be very good with short-lived objects though 2010-11-08T14:32:56 true 2010-11-08T14:33:05 also, hi everyone 2010-11-08T14:33:15 it's a Migi32! 2010-11-08T14:33:56 Migi32 uses logic attack! It's super effective! :P 2010-11-08T14:35:25 but according to my understanding of garbage collection mark and sweep is equally effective if the tree is deep or not 2010-11-08T14:36:04 can't you just turn off the garbage collector? how much memory do you need per turn? it's only 200 turns 2010-11-08T14:36:12 memory leak till the end 2010-11-08T14:36:22 lol. The engineering solution ;) 2010-11-08T14:36:31 I tihnk it's a question of how big the tree is (i.e. how much garbage you make), not how deep it is 2010-11-08T14:36:31 hey.. it'll work won't it? 2010-11-08T14:37:03 amstan, depends on memory limits and how much garbage you make... 2010-11-08T14:37:50 i know in python there's a way to do manual garbage collection 2010-11-08T14:37:52 I think also "tiny ints" (or any POD's) aren't even garbage collected. Just reading through the logs... 2010-11-08T14:38:47 Migi32: yeah.. in python, it actually reserves a few objects and uses them for ints from -1 to 10 or something like that 2010-11-08T14:39:55 ah, right, Python doesn't even have such a thing as a stack 2010-11-08T14:40:12 I mean, to put local small variables on 2010-11-08T14:40:59 python mostly uses reference counting and only uses GC to break cycles 2010-11-08T14:41:32 at least cpython that is 2010-11-08T14:41:36 what actually happens if your bot doesn't listen to stdcin for a while after it's written to stdout? 2010-11-08T14:42:25 is it written to a buffer and your timer starts running? Or do you just not get the data at all? 2010-11-08T14:42:33 buffer 2010-11-08T14:42:35 ok 2010-11-08T14:43:40 but actually if you've only spent 1 second thinking, how likely is it that you will spend more than 0.1s garbage collecting? 2010-11-08T14:44:02 calling the gc is more expensive than the gc itself in .net idk about other languages 2010-11-08T14:44:30 it's not really made to handle direct calls tho 2010-11-08T14:44:37 i think they took it out completely in 3.5 or 4 2010-11-08T14:45:29 well the real solution here is to just swallow your pride, stop playing with toy languages and use C/C++ :P 2010-11-08T14:45:31 or lisp :P 2010-11-08T14:45:47 haha, "or lisp" 2010-11-08T14:46:01 lisp has gc apparently 2010-11-08T14:46:03 i did not know that 2010-11-08T14:46:17 holy crap my bot won - i am going to enjoy that for a minute before picking apart its actions 2010-11-08T14:47:09 ah there's shampoo to give me the defeat i need to get me going on debug 2010-11-08T14:50:30 i apparently broke defending oops 2010-11-08T14:57:27 *** Guest13825 has quit IRC (Quit: Page closed) 2010-11-08T15:00:17 lol JamesMG you like my RageBot mkVII? 2010-11-08T15:01:24 I was testing new turn 1 code, obviously it's dumb :( 2010-11-08T15:01:33 it went for planets it was never gonna get 2010-11-08T15:01:57 that was a pure greed no holds barred one 2010-11-08T15:02:11 i tried implementing auto-holdback based on defense need but it paralyzed it :( 2010-11-08T15:02:21 *** Accoun has quit IRC () 2010-11-08T15:02:37 it cleaned mine up quite well after mine screwed the start for sure 2010-11-08T15:02:40 *** hellman has quit IRC (Read error: Connection reset by peer) 2010-11-08T15:03:14 eh your start wasn't that bad - it does ok: http://72.44.46.68/canvas?game_id=914497 2010-11-08T15:03:51 it's totally screwed once it loses the upper hand tho 2010-11-08T15:04:18 how does it choose planets to attack, pure growth/numships? 2010-11-08T15:04:28 pretty much 2010-11-08T15:04:42 the swarming behaviour of it messes with my fairly rigid defense code 2010-11-08T15:04:43 it's a test of my neutral acq system really but if it has the upper hand it can be the entire game 2010-11-08T15:05:41 maybe i'll port this bot to official it will do ok 2010-11-08T15:05:56 just so i can have C and a bot less than 2 months old on there lol 2010-11-08T15:06:22 your attack pattern makes me want to revise my defense code 2010-11-08T15:07:07 haha holy shit it beat janstaunton too 2010-11-08T15:07:15 that'll do pig, that'll do 2010-11-08T15:07:29 *** Yellow655 has joined #aichallenge 2010-11-08T15:07:49 i'm gonna throw it up on official been waiting to see what compile errors i'll get lol 2010-11-08T15:08:38 hmm maybe i underestimated it: http://72.44.46.68/canvas?game_id=914520 it was definitely a comeback 2010-11-08T15:09:24 interesting that it took the 0 growth middle planet 2010-11-08T15:10:00 haha probably a bug 2010-11-08T15:10:04 at least it didn't crash it 2010-11-08T15:10:18 no i take that back 2010-11-08T15:10:28 was probably evaluated as 0 gain 2010-11-08T15:10:57 oh no it's auto-staging 2010-11-08T15:10:58 wow it works 2010-11-08T15:11:02 i forgot about that 2010-11-08T15:11:15 Is the main server by any chance down? 2010-11-08T15:11:57 you have staging ships as part of your tree? 2010-11-08T15:12:00 flag: looks like it 2010-11-08T15:12:07 d0ugal: okay, thanks 2010-11-08T15:12:28 It seems they have too many mysql connections 2010-11-08T15:12:35 Judging by the PHP error's 2010-11-08T15:15:23 should be back now, it was an attempt and a database backup gone awry 2010-11-08T15:15:34 Aha 2010-11-08T15:15:44 Working now for me 2010-11-08T15:15:55 Top 10 players: bocsimacko(4191), bix0r4ever(4028), shangas(4024), george(3973), rsergio(3943), GreenTea(3937), medrimonia(3936), Accoun(3895), dmj111(3891), Xrillo(3886) 2010-11-08T15:16:05 *** fawek has quit IRC (Read error: Connection reset by peer) 2010-11-08T15:20:08 *** Accoun has joined #aichallenge 2010-11-08T15:20:53 *** irchs has joined #aichallenge 2010-11-08T15:24:40 lol this guy's bot got matched vs. itself: http://ai-contest.com/forum/viewtopic.php?f=19&t=1072 2010-11-08T15:24:52 damn i forgot how to readline in C 2010-11-08T15:25:03 is there super quick way or am i one char at a timing it 2010-11-08T15:25:58 *** boegel_ has quit IRC (Quit: Leaving) 2010-11-08T15:26:16 *** Utkarsh has quit IRC (Ping timeout: 272 seconds) 2010-11-08T15:29:24 *** Yellow655 has quit IRC (Ping timeout: 265 seconds) 2010-11-08T15:29:29 *** McLeopold has quit IRC (Ping timeout: 240 seconds) 2010-11-08T15:30:02 ah fgets thanks! 2010-11-08T15:30:28 demerzel_, I'd getchar() it, readline requires a char* buffer, which has a finite size. That can get nasty. 2010-11-08T15:30:45 i'm not using actual readline 2010-11-08T15:30:55 if orders are exceeding 128 chars i'm probably screwed anyway 2010-11-08T15:31:09 (per line) 2010-11-08T15:31:30 demerzel_, quoting the specs: 2010-11-08T15:32:03 *** Utkarsh has joined #aichallenge 2010-11-08T15:32:28 ok I don't fin it in the specs 2010-11-08T15:32:45 lol 2010-11-08T15:32:52 find* 2010-11-08T15:33:05 if your opponent could 10000000000000... ships, shure you're doomed. 2010-11-08T15:33:05 fgets will work 2010-11-08T15:33:15 i have a char buffer of 128 chars so yea what p4p4 said 2010-11-08T15:33:20 but somewhere it said that it's unlikely that comments are unlikely but that you shouldn't count on there never ever being comments 2010-11-08T15:34:05 the only comment htey ever considered was a turn # 2010-11-08T15:34:29 considering an actual line of planet data is 32 chars if somebody puts a 90 char comment in there i'll be mad 2010-11-08T15:34:54 besides fgets has a safety on it 2010-11-08T15:35:06 ah yes true, good point. 2010-11-08T15:35:29 if it were a comment of > 128 chars it would just ignore those chars, I think 2010-11-08T15:35:42 it would try to parse the last 128+ chars 2010-11-08T15:36:06 would probably break me anyway 2010-11-08T15:36:07 lol 2010-11-08T15:36:45 well, just pray they don't add # helloooooooo demerzel_ to the communications :P 2010-11-08T15:37:22 indeed 2010-11-08T15:37:27 although that is ok 2010-11-08T15:37:29 :P 2010-11-08T15:37:50 * d0ugal adds random really long comments 2010-11-08T15:39:18 demerzel_, but why are you using c? c++ has some nice functions that don't have buffer limits (they just return a std::string) 2010-11-08T15:40:06 because they do have buffer limits it's just hidden away 2010-11-08T15:40:10 i don't need the overhead 2010-11-08T15:40:29 plus i don't use a single c++ library not gonna do it for that lol 2010-11-08T15:40:43 i do not like black boxes 2010-11-08T15:41:22 Anywhere listing if the server has any non-standard python libs? 2010-11-08T15:41:28 (like numpy) 2010-11-08T15:42:44 haha how sad is that my bot has never used stdin/out to this point 2010-11-08T15:42:45 demerzel_, oh your bot is entirely c? 2010-11-08T15:42:52 yes 2010-11-08T15:42:56 let's see what gcc has to say about it 2010-11-08T15:44:04 d0ugal: numpy is not available 2010-11-08T15:44:22 d0ugal: http://pastebin.com/ZBMtLbFm 2010-11-08T15:44:22 janzert: figured as much, just checkin' 2010-11-08T15:44:29 d0ugal: not sure if it covers everything 2010-11-08T15:44:39 That's helpful, thanks 2010-11-08T15:44:48 language unknown 2010-11-08T15:44:52 .c and .h files? 2010-11-08T15:45:41 you have MyBot.c 2010-11-08T15:45:54 mybot.c 2010-11-08T15:45:54 capitalization is important 2010-11-08T15:45:57 oh god 2010-11-08T15:45:59 really? 2010-11-08T15:46:02 sheesh 2010-11-08T15:46:05 this is unix 2010-11-08T15:46:31 now i gotta go through and rename my includes lol 2010-11-08T15:46:44 you really couldn't have had it look for mybot or Mybot or MyBot 2010-11-08T15:46:51 amstan: do you access to the root database user? 2010-11-08T15:47:17 *** bhasker has joined #aichallenge 2010-11-08T15:47:17 not the root, but i can read write all databases 2010-11-08T15:47:26 i do have sudo too 2010-11-08T15:47:30 so.. meh 2010-11-08T15:47:32 lol 2010-11-08T15:47:43 hmm, I need the contest user granted the FILE and RELOAD rights 2010-11-08T15:47:51 guess I'll have to get jeff to do it 2010-11-08T15:48:06 *** Rubicon-|-Cross has quit IRC (Quit: Leaving) 2010-11-08T15:48:22 demerzel_, you only need MyBot.c to be correctly capitalised. And do you even have a MyBot.h? 2010-11-08T15:48:29 i did try just naming the .c MyBot.c since everything else is including mybot.h 2010-11-08T15:48:32 see if that works 2010-11-08T15:48:39 yea i have a mybot.h 2010-11-08T15:49:09 ERROR 2010-11-08T15:49:25 at least it sees C :) i figured gcc would have a thing or two to say 2010-11-08T15:49:27 janzert: you could do it.. if you turn off mysql for a minute or 2 2010-11-08T15:49:29 *** bhasker_ has quit IRC (Ping timeout: 240 seconds) 2010-11-08T15:49:50 where's my email >:( 2010-11-08T15:49:55 hmm, maybe tonight 2010-11-08T15:49:58 demerzel_: spam? 2010-11-08T15:50:22 oh wow good call 2010-11-08T15:50:26 weird it never was spam before 2010-11-08T15:51:13 why are you using unroll-loops 2010-11-08T15:51:18 that's bad 2010-11-08T15:51:35 wraaa 2010-11-08T15:51:41 Time to build a bot with persistent state 2010-11-08T15:51:43 *** irchs is now known as janstaunton 2010-11-08T15:51:59 state space searching here we come 2010-11-08T15:52:12 I'm coming for you demerzel_ 2010-11-08T15:52:21 haha i beat you! 2010-11-08T15:52:23 :D 2010-11-08T15:52:29 with my bot i totally intended to lose 2010-11-08T15:52:32 oh no newline 2010-11-08T15:52:33 hmm 2010-11-08T15:52:33 :P 2010-11-08T15:52:33 demerzel_: probably because someone showed it to be an improvement for them in the tron contest and it has stayed ever since 2010-11-08T15:52:59 *** McLeopold has joined #aichallenge 2010-11-08T15:53:51 it's kinda bad 2010-11-08T15:54:07 i mean i guess if you want my ++turn < 201 to be 201 separate iterations that's fine 2010-11-08T15:54:19 it doesn't matter to me but it will make the prog bigger 2010-11-08T15:54:33 doubt it matters either way to my actual code 2010-11-08T15:59:53 try that agaiun 2010-11-08T16:00:51 Who is eXample? 2010-11-08T16:01:04 me and her/him be battlin' 2010-11-08T16:01:06 demerzel_, there's a limit to how much it will unroll I think 2010-11-08T16:01:09 what does this mean 2010-11-08T16:01:09 pod_globals.h:15: warning: 'struct gamestate' declared inside parameter list 2010-11-08T16:01:21 Missing a }?! 2010-11-08T16:01:23 or a ) 2010-11-08T16:01:23 ? 2010-11-08T16:01:26 or a ;? :) 2010-11-08T16:01:42 great 2010-11-08T16:01:45 lol 2010-11-08T16:02:05 does the server compile with more warnings than -Wall? 2010-11-08T16:02:20 and why does it compile with -Werror, lol 2010-11-08T16:04:47 i /think/ it was just upset about my order of include 2010-11-08T16:04:54 i hope so anyway cuz i have no idea otherwise 2010-11-08T16:05:20 Is there any word on the number of turns for the official tourney? 2010-11-08T16:05:23 Will it vary at all 2010-11-08T16:05:24 ? 2010-11-08T16:05:40 1000 turns, 100 planets! 2010-11-08T16:05:45 I think it staying at 200 is a pretty safe bet at this point 2010-11-08T16:05:54 janstaunton: I noticed your bot improved pretty drastically in the last week or so 2010-11-08T16:06:11 oh i know what it is - for some reason i still had POD_structs.h 2010-11-08T16:06:18 looking for pod_structs lol 2010-11-08T16:06:27 Jeff talked about making an official statement about a week ago but never did 2010-11-08T16:06:34 wraaaa I can now time travel into the future 2010-11-08T16:06:48 It's still a bit shit, I am making it infinitely better atm 2010-11-08T16:06:56 but yeah, it is improving :D 2010-11-08T16:08:24 My bot and yours seem to spar a lot 2010-11-08T16:08:29 demerzel_ at some point it becomes worth it to install virtual box, ubuntu 8, run the build script, and try your bot with gcc locally until it works... :P 2010-11-08T16:08:36 Which one is your current bot, the dev one? @JamesMG 2010-11-08T16:09:13 i am pretty close 2010-11-08T16:09:32 there aren't enough crappy bots on tcp any more... I should run some old ones 2010-11-08T16:09:32 some kind of bitching about a global void instead of being in mybot.h 2010-11-08T16:09:35 and something else 2010-11-08T16:09:42 i can run mine :P 2010-11-08T16:09:47 janstaunton: .old is the one on the official server, .dev is me trying to work out bugs and stuff 2010-11-08T16:09:52 so .dev is changing pretty often 2010-11-08T16:09:56 I see 2010-11-08T16:09:59 well I don't need it right right now 2010-11-08T16:10:09 but maybe I will just set mine up to stay on 2010-11-08T16:10:12 I can't remember what one is pinging against the tcp server atm 2010-11-08T16:10:16 It's a week or so old 2010-11-08T16:10:25 And my official server one is well old :) 2010-11-08T16:10:35 What's your username on the official server? 2010-11-08T16:10:38 JamesMG 2010-11-08T16:10:46 lavalamp 2010-11-08T16:10:55 it's still clawing it's way back up from 2000th place 2010-11-08T16:11:10 :o 2010-11-08T16:11:19 mine was 250's and then magically became 150's over the weekend... 2010-11-08T16:11:49 the server had me play someone in 2000th place when I resubmitted 2010-11-08T16:12:08 and it has crashed a couple times on it's way back up due to C#/mono 2010-11-08T16:13:27 uh wtf 2010-11-08T16:13:58 there are no errors here: http://pastebin.com/PQjGaMjg 2010-11-08T16:14:25 *** McLeopold has quit IRC (Ping timeout: 272 seconds) 2010-11-08T16:14:32 it may realize const void is redundant but so do i 2010-11-08T16:14:36 JamesMG: 2010-11-08T16:14:39 balls 2010-11-08T16:14:41 stupid client 2010-11-08T16:14:43 not an error 2010-11-08T16:14:47 no errors there? 2010-11-08T16:14:51 JamesMG: You using monodev or vs? 2010-11-08T16:15:57 Top 10 players: bocsimacko(4207), shangas(4051), bix0r4ever(4047), george(4001), rsergio(3987), GreenTea(3958), Accoun(3918), dmj111(3907), Xrillo(3904), medrimonia(3898) 2010-11-08T16:16:00 monodev on windows, running my tcp bots on mono on linux 2010-11-08T16:16:04 what does const void even mean? 2010-11-08T16:16:42 but I haven't taken the time to build out an official server clone for that, I really should 2010-11-08T16:16:50 see if I can isolate some more of the crashing issues 2010-11-08T16:21:39 *** Apophis has quit IRC (Ping timeout: 276 seconds) 2010-11-08T16:21:52 *** yasith_ has quit IRC (Ping timeout: 265 seconds) 2010-11-08T16:22:33 *** Aaaarg has quit IRC (Ping timeout: 255 seconds) 2010-11-08T16:30:41 lavalamp_ const void just means you can't change it i guess haha 2010-11-08T16:33:18 alright one more time jeez 2010-11-08T16:33:40 * p4p4 tries to push the void a bit to the left 2010-11-08T16:34:09 const void << 1; 2010-11-08T16:34:59 argh now it has a problem with qsort 2010-11-08T16:35:00 wttwetwrklj; 2010-11-08T16:35:04 didn't have one before 2010-11-08T16:36:06 *** yasith has joined #aichallenge 2010-11-08T16:36:22 wttwetwrklj? 2010-11-08T16:37:23 that's "I'm frustrtated" in qwerty 2010-11-08T16:37:45 what in the fuck 2010-11-08T16:37:49 does it want 2010-11-08T16:37:55 and why are warnings errors 2010-11-08T16:38:16 what's the warning/error? 2010-11-08T16:38:17 oh that's fucking hilarious 2010-11-08T16:38:26 qsort pointer 4 is incompatible 2010-11-08T16:38:29 pointer 4 is a const void 2010-11-08T16:38:33 which is an error 2010-11-08T16:38:44 if i make it const void 2010-11-08T16:38:46 const void * I understand 2010-11-08T16:38:53 const void I would barf on, too 2010-11-08T16:39:26 ok i have no idea what to do here 2010-11-08T16:39:37 87 MB of logs... I should log less :P 2010-11-08T16:40:32 oh crap, forgot to clear my logs from the weekend 2010-11-08T16:41:14 why can't you just lose the "const"? 2010-11-08T16:41:22 i already did that 2010-11-08T16:41:28 i only get one error, qsort pointer 4 2010-11-08T16:41:36 trying to make the predicate function a const that's my last idea 2010-11-08T16:41:49 I guess I'd have to see the line of code 2010-11-08T16:41:51 nope 2010-11-08T16:42:01 won't take an int, a static int, or a const int as an argument 2010-11-08T16:42:08 for the predicate func 2010-11-08T16:42:12 ok i'll pastebin it 2010-11-08T16:42:15 i am out of ideas 2010-11-08T16:42:16 yeah 2010-11-08T16:42:49 i hate const woes... 2010-11-08T16:44:11 http://pastebin.com/a2b7bNjW 2010-11-08T16:44:21 i also tried making the ascendints const int 2010-11-08T16:44:23 didn't take that either 2010-11-08T16:44:38 *** janstaunton has quit IRC (Quit: janstaunton) 2010-11-08T16:45:42 *** spacebat has joined #aichallenge 2010-11-08T16:45:46 *** irchs has joined #aichallenge 2010-11-08T16:45:56 can you paste the definition of qsort? 2010-11-08T16:46:08 phreeza beat me to it 2010-11-08T16:46:28 http://www.cplusplus.com/reference/clibrary/cstdlib/qsort/ 2010-11-08T16:46:47 oh never new that had a quicksort ^^ 2010-11-08T16:46:59 const void * should do it 2010-11-08T16:47:18 try qsort(tempsort, kMaxPlanets, sizeof(int), &ascendints) 2010-11-08T16:47:19 int ascendints (const void * a, const void * b) 2010-11-08T16:47:25 yea i am trying that now 2010-11-08T16:47:35 lavalamp 2010-11-08T16:47:40 i'll try janzert if that doesn't work lol 2010-11-08T16:47:47 you want a pointer to ascendints basically 2010-11-08T16:48:00 const void*a worked 2010-11-08T16:48:07 a func is a pointer tho really 2010-11-08T16:48:11 janzert I don't think that should make a difference, there's an implied & when you mention a function, at least it C++ 2010-11-08T16:48:12 well i'm not gonna get into that lol 2010-11-08T16:48:19 ok, wasn't sure 2010-11-08T16:48:22 lavalamp_ and i had the same idea it worked :) 2010-11-08T16:48:31 now the bot is proably broken from me playing with modifiers lol 2010-11-08T16:48:45 yup 2010-11-08T16:48:48 sweet lol 2010-11-08T16:48:56 http://ai-contest.com/profile.php?user_id=9166 2010-11-08T16:48:58 O.o 2010-11-08T16:49:24 hurgh 2010-11-08T16:49:45 heh it segfaults? 2010-11-08T16:49:49 i dunno 2010-11-08T16:49:53 it works fine on my machine 2010-11-08T16:50:01 lol, language "Unknown" in one of your submissions. Fucked up beyond all recognition :P 2010-11-08T16:50:10 Unknown = mybot.c 2010-11-08T16:50:16 ah. ok :P 2010-11-08T16:50:18 now you get to the tricky part of debugging the contest entry :/ 2010-11-08T16:50:32 demerzel, maybe it is time to set up that ubuntu virtualbox 2010-11-08T16:50:39 pass 2010-11-08T16:50:44 if i get to that point i'm just done with this contest 2010-11-08T16:51:05 *shrug* it didn't take me long, although I've done it before 2010-11-08T16:51:06 it might just be too ambitious 2010-11-08T16:51:21 i barely know linux and don't have any idea how to set up any virtualboxes 2010-11-08T16:51:34 ah ok 2010-11-08T16:51:40 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-11-08T16:52:15 toning down the lookahead hopefully that's it if not i'll grab one of the standard bot players and see if it's busted 2010-11-08T16:52:38 does it work in release mode on your own system? 2010-11-08T16:52:49 oh i can't :( 2010-11-08T16:52:54 stupid 10 minute rule 2010-11-08T16:53:03 yes 2010-11-08T16:53:04 and/or turn on all the same optimizations 2010-11-08T16:53:46 *** sigh has joined #aichallenge 2010-11-08T16:55:27 *** sigh_ has joined #aichallenge 2010-11-08T16:55:27 *** sigh has quit IRC (Read error: Connection reset by peer) 2010-11-08T16:57:08 damn galcon; now every time I close my eyes all I see is fleets converging on planets 2010-11-08T16:58:05 like, tiny red arrows from everywhere ? 2010-11-08T16:58:19 all too often 2010-11-08T16:58:24 your nightmares will never be the same ;) 2010-11-08T16:58:37 ;( 2010-11-08T16:59:41 in space, no-one can hear you scream 2010-11-08T16:59:50 over the roar of the armada's engines 2010-11-08T17:00:51 galcon is like Reach For The Stars but for cocaine addled monkeys 2010-11-08T17:01:09 * spacebat wipes nose 2010-11-08T17:03:34 *** davidjliu has joined #aichallenge 2010-11-08T17:04:10 grr how the hell do you give jbotmanager a map 2010-11-08T17:04:19 i am escalating in anger 2010-11-08T17:04:39 type in the right directory in the box? 2010-11-08T17:04:48 i set lookahead to 1 and it's still busted 2010-11-08T17:04:56 i put map1.txt in the damn dir with it 2010-11-08T17:05:03 and also put maps in a dir next to it 2010-11-08T17:05:30 *** iFire has joined #aichallenge 2010-11-08T17:06:35 welp i have no idea what the hell that means but i guess that means i broke it dicking with gcc's complaints 2010-11-08T17:06:42 it made a green line down then stopped 2010-11-08T17:06:45 so i guess that means bad idk 2010-11-08T17:06:55 *** choas has quit IRC (Quit: leaving) 2010-11-08T17:07:48 I have my bot make a replay file, then I run it from the command line to see where it crashes 2010-11-08T17:08:08 i don't htink it's getting to that point 2010-11-08T17:08:12 *** iFire has quit IRC (Remote host closed the connection) 2010-11-08T17:08:13 but it sounds like you could just make a file with the map followed by the word go 2010-11-08T17:08:24 pipe it into your bot 2010-11-08T17:08:28 and see what happens 2010-11-08T17:08:32 if i just type go to it it says go back so idk i'll play with it 2010-11-08T17:08:35 *** iFire has joined #aichallenge 2010-11-08T17:08:58 MyBot.exe < test.txt 2010-11-08T17:13:28 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T17:13:28 MyBot.exe > test.txt or MyBot.txt | test.txt will be better I think 2010-11-08T17:13:51 *** _W_ <_W_!~w@irc.w-wins.com> has quit IRC (Ping timeout: 255 seconds) 2010-11-08T17:15:57 Top 10 players: bocsimacko(4211), shangas(4049), bix0r4ever(4042), george(3991), rsergio(3986), GreenTea(3960), Accoun(3919), Xrillo(3904), medrimonia(3898), dmj111(3898) 2010-11-08T17:16:30 Migi32, I wasn't sure if that would work on windoze or not 2010-11-08T17:16:34 *** Aaaarg has joined #aichallenge 2010-11-08T17:17:12 ah, neither am I 2010-11-08T17:17:20 on ubuntu I use cat test.txt | MyBot 2010-11-08T17:17:30 but I believe MyBot.exe < test.txt overwrites MyBot.exe 2010-11-08T17:17:43 with test.txt 2010-11-08T17:18:00 that's not what happened when I used it in the debugger 2010-11-08T17:18:16 *** Bobng has quit IRC (Read error: Connection reset by peer) 2010-11-08T17:18:17 *** irchs has quit IRC (Quit: irchs) 2010-11-08T17:18:26 It's supposed to send test.txt on stdin, afaik 2010-11-08T17:18:29 *** bhasker_ has joined #aichallenge 2010-11-08T17:18:38 ok demerzel_ listen to lavalamp_ he probably knows this a lot better than me :) 2010-11-08T17:18:48 lol 2010-11-08T17:19:03 *** Bobng has joined #aichallenge 2010-11-08T17:19:19 type x.txt | MyBot.exe 2010-11-08T17:19:29 *** Utkarsh has joined #aichallenge 2010-11-08T17:19:46 *** _W_ <_W_!~w@irc.w-wins.com> has joined #aichallenge 2010-11-08T17:21:49 *** bhasker has quit IRC (Ping timeout: 260 seconds) 2010-11-08T17:29:45 this is seriously frustrating 2010-11-08T17:29:49 it appears to be working on my machine 2010-11-08T17:30:25 demerzel_ does it handle both windows and unix line endings? 2010-11-08T17:30:33 *** perestrelka has quit IRC (Ping timeout: 276 seconds) 2010-11-08T17:30:43 it's not that tho 2010-11-08T17:30:47 it's saying it doesn't start/crashes 2010-11-08T17:31:04 oh hm 2010-11-08T17:31:04 works fine on mine, i'm gonna see if removing the reset gamestate function is it if not i have no idea 2010-11-08T17:31:20 i only really used it to reset between tcp games anyway 2010-11-08T17:32:20 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-11-08T17:32:22 *** Azrathud has joined #aichallenge 2010-11-08T17:32:38 *** iFire has joined #aichallenge 2010-11-08T17:32:40 how would you keep state between different games ? 2010-11-08T17:32:51 *** perestrelka has joined #aichallenge 2010-11-08T17:33:06 bbl 2010-11-08T17:33:11 *** lavalamp__ has quit IRC (Quit: Page closed) 2010-11-08T17:34:14 ok 2 GB logfiles this is getting out of hand... 2010-11-08T17:35:35 *** demerzel_ has quit IRC (Quit: Page closed) 2010-11-08T17:37:24 *** amstan has quit IRC (Ping timeout: 240 seconds) 2010-11-08T17:38:35 *** demerzel_ has joined #aichallenge 2010-11-08T17:38:48 amstan janzert or whoever can you please tell me if the bot is timing out or crashing 2010-11-08T17:40:04 *** Apophis has joined #aichallenge 2010-11-08T17:42:59 I really love this challenge but I don't like that much playing the real Galcon, same for you? 2010-11-08T17:44:36 *** Bobng has quit IRC (Ping timeout: 255 seconds) 2010-11-08T17:45:14 I've never played Galcon 2010-11-08T17:45:20 *** sigh_ is now known as sigh 2010-11-08T17:45:41 waiting 10 minutes between trying things is really damn annoying 2010-11-08T17:45:48 all i want to know is if it's crashing or timing out 2010-11-08T17:46:26 i went from cool lemem put this up on official to holy shit i hate this contest's shitty server 2010-11-08T17:47:05 :( 2010-11-08T17:48:00 and lack of communication i should say 2010-11-08T17:48:09 everything about this contest is cool except the official server haha 2010-11-08T17:48:48 yeah, they've had a lot of issues with it 2010-11-08T17:49:17 it was worse earlier, emails were only working intermittently or not at all 2010-11-08T17:49:31 and bocsimacko's brain :) Top1 in lisp, must be hot like hell :) 2010-11-08T17:49:38 *** demerzel_ has quit IRC (Quit: Page closed) 2010-11-08T17:56:06 *** RainCT has quit IRC (Remote host closed the connection) 2010-11-08T17:56:53 *** flag has quit IRC (Quit: Lost terminal) 2010-11-08T17:59:48 *** irchs has joined #aichallenge 2010-11-08T18:08:15 *** p4p4 has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2010-11-08T18:13:32 *** Migi32 has quit IRC (Quit: http://boinc.berkeley.edu) 2010-11-08T18:14:35 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T18:15:57 Top 10 players: bocsimacko(4233), bix0r4ever(4053), shangas(4039), rsergio(4010), george(4007), GreenTea(3977), medrimonia(3949), Accoun(3947), Xrillo(3924), dmj111(3899) 2010-11-08T18:16:07 *** JamesMG has quit IRC (Quit: Leaving.) 2010-11-08T18:21:15 *** Utkarsh has joined #aichallenge 2010-11-08T18:22:00 *** irchs has quit IRC (Quit: irchs) 2010-11-08T18:25:06 *** amstan has joined #aichallenge 2010-11-08T18:27:28 *** iFire has quit IRC (Read error: Connection reset by peer) 2010-11-08T18:30:10 *** iFire has joined #aichallenge 2010-11-08T18:37:04 *** amstan has quit IRC (Read error: Connection reset by peer) 2010-11-08T18:37:21 *** amstan has joined #aichallenge 2010-11-08T18:39:58 *** Utkarsh has quit IRC (Ping timeout: 260 seconds) 2010-11-08T18:44:39 *** Utkarsh has joined #aichallenge 2010-11-08T18:49:51 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-11-08T18:54:01 *** superflit has quit IRC (Quit: superflit) 2010-11-08T18:55:00 *** bhasker has joined #aichallenge 2010-11-08T18:56:02 *** Utkarsh has joined #aichallenge 2010-11-08T18:58:34 *** amstan has quit IRC (Ping timeout: 240 seconds) 2010-11-08T18:58:57 *** bhasker_ has quit IRC (Ping timeout: 276 seconds) 2010-11-08T18:59:46 *** kjdf has left #aichallenge 2010-11-08T19:00:05 *** mceier has quit IRC (Quit: leaving) 2010-11-08T19:00:47 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T19:06:15 *** demerzel_ has joined #aichallenge 2010-11-08T19:06:25 i am stumped 2010-11-08T19:06:34 i can run it and send go to it and it flags back 2010-11-08T19:06:45 but if i run it in the sandboxer it does nothing 2010-11-08T19:07:11 i have no idea where to look lol 2010-11-08T19:07:47 i can even pass it some bogus planets like P 1 1 1 1 50 5 and it makes orders 2010-11-08T19:08:02 but on teh site and in planetwarrior.exe (iouri's thing) it doesn't do jack 2010-11-08T19:11:37 *** _Utkarsh <_Utkarsh!~admin@117.201.81.174> has joined #aichallenge 2010-11-08T19:12:34 *** Azrathud has quit IRC (Read error: Connection reset by peer) 2010-11-08T19:15:58 Top 10 players: bocsimacko(4224), bix0r4ever(4044), shangas(4029), rsergio(4002), george(3998), GreenTea(3967), Accoun(3939), medrimonia(3935), Xrillo(3915), dmj111(3897) 2010-11-08T19:16:07 *** Appleman1234 has quit IRC (Ping timeout: 265 seconds) 2010-11-08T19:16:36 *** _Utkarsh <_Utkarsh!~admin@117.201.81.174> has quit IRC (Ping timeout: 265 seconds) 2010-11-08T19:18:54 *** kjdf has joined #aichallenge 2010-11-08T19:20:48 *** Utkarsh has joined #aichallenge 2010-11-08T19:20:57 *** nullkuhl has quit IRC (Remote host closed the connection) 2010-11-08T19:25:16 *** McLeopold has joined #aichallenge 2010-11-08T19:25:40 *** Utkarsh has quit IRC (Ping timeout: 272 seconds) 2010-11-08T19:25:49 I was clenching my teeth watching this one: http://72.44.46.68/canvas?game_id=916956 2010-11-08T19:30:11 *** Azrathud has joined #aichallenge 2010-11-08T19:31:10 *** Utkarsh has joined #aichallenge 2010-11-08T19:36:11 *** Utkarsh has quit IRC (Ping timeout: 265 seconds) 2010-11-08T19:38:37 why the name ant bot? 2010-11-08T19:41:56 *** Utkarsh has joined #aichallenge 2010-11-08T19:44:05 kjdf: originally, it was to describe my funneling logic 2010-11-08T19:44:31 but the logic sucked, and I removed it, and kept developing 2010-11-08T19:44:36 i have no idea what to do i can't fix this i dont' even know what's broke -- i have it spitting out every character it gets and then it just dies at p18 2010-11-08T19:44:56 after getting the full line? idk :( 2010-11-08T19:45:04 what's a p18 2010-11-08T19:45:06 changing maps doesn't matter 2010-11-08T19:45:11 planet #18 2010-11-08T19:45:46 btw: http://72.44.46.68/getplayer?player=demerzel_dev03 2010-11-08T19:46:10 :) 2010-11-08T19:46:20 i've been staring at this for the last hour: it's a spit out of everything it receives 2010-11-08T19:46:20 http://pastebin.com/rQq5Kt5g 2010-11-08T19:46:28 *** Utkarsh has quit IRC (Ping timeout: 260 seconds) 2010-11-08T19:46:41 oh i'm on 4 now if you didn't beat <4 i'd be pretty surprised lol 2010-11-08T19:46:49 in fact if you didn't beat 4 i'd be surprised 2010-11-08T19:47:11 but i can't figure this out - it runs fine on my machine but just dies there in sandbox (and on the main server) 2010-11-08T19:47:22 i should say 'in IDE' 2010-11-08T19:47:32 what game engine are you using? 2010-11-08T19:47:39 iouri's 2010-11-08T19:47:47 but it dies on main server too 2010-11-08T19:48:01 i never put in stdin/stdout until now so i'm guessing it has something to do with that 2010-11-08T19:48:13 but what i have no idea - i changed fgets to an individual char read 2010-11-08T19:48:31 is this c code? 2010-11-08T19:48:35 yea 2010-11-08T19:48:44 i even manually added a '\0' to the end to make sure 2010-11-08T19:49:28 C code: http://pastebin.com/Y5ys6SC0 -- io_output is a global char[128] 2010-11-08T19:50:19 tried ch = io_output, &io_output[0] is my latest attempt until i realized it was actually reading it 2010-11-08T19:50:55 who is io_output declared? 2010-11-08T19:50:58 *how 2010-11-08T19:51:06 it's global, declared above it 2010-11-08T19:51:11 *** Utkarsh has joined #aichallenge 2010-11-08T19:51:12 why not make it bigger ;) 2010-11-08T19:51:16 I mean, what is the declaration? 2010-11-08T19:51:39 char io_output[128]; - have it in the .h too 2010-11-08T19:52:04 buffer overrun code... 2010-11-08T19:52:16 it's not going over 128 2010-11-08T19:52:25 either way, this code is not very safe 2010-11-08T19:52:40 yea i know it was a quick attempt to look at one char at a time 2010-11-08T19:52:59 i'd solidify it but there's the whole 'crashign at 18' provblem 2010-11-08T19:53:59 it's successfully creating the planet too - MADE PLANET indicates it went through fine and the index is correct 2010-11-08T19:54:02 i am totally stumped 2010-11-08T19:56:45 *** cfaftw has joined #aichallenge 2010-11-08T19:58:03 demerzel_: idk 2010-11-08T19:58:43 do you know which line of code is the last to execute? 2010-11-08T19:59:13 anyone have a quick link to a C starter pack? 2010-11-08T19:59:29 from you output, it looks like the fprintf function? 2010-11-08T19:59:31 i'm searching the forum... but C always sucks as a search term 2010-11-08T19:59:45 how about starter? 2010-11-08T20:00:14 cfaftw: on tcp right now? 2010-11-08T20:00:34 no 2010-11-08T20:01:01 no but the fact that the string is smaller on the last line tells me something broke badly 2010-11-08T20:01:13 mcleopold, what's new with your bot? 2010-11-08T20:01:38 cfaftw: if you find a c starter pack hook me up cuz i had the cpp one but it's a bunch of << >> whatever 2010-11-08T20:01:38 cfaftw: the framework 2010-11-08T20:01:45 i even took the state out completely from that and it still broke :( 2010-11-08T20:01:47 And the targeting system. 2010-11-08T20:03:01 demerzel_, i thought you were already working on a C bot 2010-11-08T20:03:28 i have one but it is broken horribly with the stdin 2010-11-08T20:03:34 haha, you were the one that got me hyped up to do my bot in C 2010-11-08T20:03:38 and i have no idea why haha 2010-11-08T20:03:40 *** Appleman1234 has joined #aichallenge 2010-11-08T20:03:47 gotcha 2010-11-08T20:03:53 * brx stretches. 2010-11-08T20:03:54 i don't know of a C starter pack honestly 2010-11-08T20:04:01 i built my bot with tcp builtin it never had stdin till now haha 2010-11-08T20:04:03 somebody has to have one... there are more than just a few people doing theirs in C 2010-11-08T20:04:43 the C++ one would work but you'd have to move all the funcs out etc 2010-11-08T20:04:48 mcleopold, i'm starting from scratch... going to try minimax 2010-11-08T20:05:01 yeah, I may have to do that. 2010-11-08T20:05:45 demerzel_, i might just have to write my own... in reality it's not that much work, i guess 2010-11-08T20:06:12 the only thing i want from a starter pack is the parsing routines... which i suppose i might be able to steal from the C++ pack 2010-11-08T20:06:42 cfaftw: parsing is pretty easy. Just roll your own. should take only an hour. 2010-11-08T20:07:13 mcleopold, yeah... i was just hoping for something premade to save me that hour :) 2010-11-08T20:08:27 parsing is super easy with scanf 2010-11-08T20:09:18 here, adapt to your own structs: http://pastebin.com/ng711edr 2010-11-08T20:11:15 thanks 2010-11-08T20:11:57 never mind that it is crashing... 2010-11-08T20:12:05 haha 2010-11-08T20:12:16 it works perfectly it's the stdin part that's crashing :( 2010-11-08T20:12:28 :P 2010-11-08T20:12:30 i been using that code since day 1 on tcp - and while my strategy needs work the state was fine 2010-11-08T20:15:31 *** rangzen has quit IRC (Quit: Quitte) 2010-11-08T20:15:59 Top 10 players: bocsimacko(4238), bix0r4ever(4047), shangas(4034), george(4034), rsergio(4009), GreenTea(3990), Accoun(3948), Xrillo(3938), dmj111(3923), medrimonia(3920) 2010-11-08T20:17:35 *** McLeopold has quit IRC (Quit: Leaving) 2010-11-08T20:20:10 i wonder if there's some flush or something i need to do 2010-11-08T20:20:14 fgets just hangs completely 2010-11-08T20:20:32 that's the 'safe' way to do it since it allows a maxbuffer size 2010-11-08T20:20:37 demerzel_, when does it hang? 2010-11-08T20:20:49 upon any invocation of fgets? 2010-11-08T20:20:52 instantly 2010-11-08T20:21:00 weird 2010-11-08T20:21:22 yea - when i switched to getc or fgetc or fgets("%c") it got to char 807 then died 2010-11-08T20:21:35 or getchar - any of those 2010-11-08T20:22:05 i've been usingi t on tcp forever i never had stdin problems lol 2010-11-08T20:26:32 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T20:27:36 *** smellyhippy has quit IRC (Ping timeout: 264 seconds) 2010-11-08T20:27:53 wow... some people that I recognize from the top10 are now no longer in the top 100 2010-11-08T20:28:01 it's insane how much the competition has improved 2010-11-08T20:28:42 sigh, yep 2010-11-08T20:28:48 where are you now? 2010-11-08T20:29:15 uploaded a submission recently... at 89 so far with one loss 2010-11-08T20:29:36 i'm attempting to redo my bot in C and use minimax 2010-11-08T20:29:42 literally just started 2010-11-08T20:29:48 we'll see how that goes :P 2010-11-08T20:30:05 it should go ok if i've figured uot how to use stdin by then lol 2010-11-08T20:30:07 I tried minimax... but prelimarary results were not good, so I gave up 2010-11-08T20:30:11 nodes are easy as hell in C lol 2010-11-08T20:30:33 and fast - i can go 4 deep no sweat 2010-11-08T20:30:46 well i already have an idea how to implement it... i just haven't used C in a while 2010-11-08T20:30:46 that's 2 moves apiece tho not 4 apiece haha 2010-11-08T20:30:51 actually, scratch that... the most amazing thing is that McLeopold is 74 with the bot he uploaded at the start of the competition 2010-11-08T20:30:59 so my time upfront is shaking the rust off 2010-11-08T20:31:05 cfaftw: why c? just for speed? 2010-11-08T20:31:25 sigh, that, and I feel like i need a language change to force me to think differently 2010-11-08T20:31:36 yeah, makes sense 2010-11-08T20:31:47 every time i tried "redoing" it in Python, i ended up using the same basic methods 2010-11-08T20:31:54 :) 2010-11-08T20:31:56 it was hard to break the mold i had created 2010-11-08T20:32:44 yeah... the core of my bot rests on a planet scoring function that hasn't changed significantly since I started 2010-11-08T20:32:44 *** smellyhippy has joined #aichallenge 2010-11-08T20:32:44 *** smellyhippy has joined #aichallenge 2010-11-08T20:32:58 *** Utkarsh has joined #aichallenge 2010-11-08T20:33:23 ditto 2010-11-08T20:34:29 *** demerzel_ has quit IRC (Quit: Page closed) 2010-11-08T20:39:30 *** Mathnerd314 has quit IRC (Disconnected by services) 2010-11-08T20:39:32 *** amstan has joined #aichallenge 2010-11-08T20:39:32 *** ChanServ sets mode: +o amstan 2010-11-08T20:40:05 *** Mathnerd314_ has joined #aichallenge 2010-11-08T20:40:06 *** Mathnerd314_ is now known as Mathnerd314 2010-11-08T20:40:21 *** Utkarsh has quit IRC (Ping timeout: 276 seconds) 2010-11-08T20:44:03 *** demerzel_ has joined #aichallenge 2010-11-08T20:45:34 *** Utkarsh has joined #aichallenge 2010-11-08T20:51:44 *** sigh has quit IRC (Read error: Connection reset by peer) 2010-11-08T20:52:02 *** sigh has joined #aichallenge 2010-11-08T20:56:23 *** lavalamp_ has quit IRC (Quit: Page closed) 2010-11-08T20:57:05 does anybody know anything about piped input in C i have narrowed it down to that lol 2010-11-08T20:57:43 *** lavalamp_ has joined #aichallenge 2010-11-08T20:59:32 *** amstan has quit IRC (Remote host closed the connection) 2010-11-08T20:59:54 *** bhasker has quit IRC (Quit: bhasker) 2010-11-08T21:04:08 well, it took an all nighter, but i'm done 2010-11-08T21:07:50 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T21:11:36 *** phreeza_ has joined #aichallenge 2010-11-08T21:13:35 *** Utkarsh has joined #aichallenge 2010-11-08T21:14:14 *** phreeza has quit IRC (Ping timeout: 245 seconds) 2010-11-08T21:15:59 Top 10 players: bocsimacko(4220), bix0r4ever(4053), shangas(4028), george(4027), rsergio(4002), GreenTea(3986), Accoun(3944), Xrillo(3936), dmj111(3931), Hazard(3892) 2010-11-08T21:19:02 antimatroid1, done with what? 2010-11-08T21:19:08 undergrad 2010-11-08T21:19:20 totally? 2010-11-08T21:19:27 I started two versions of my old bot on the tcp server as a public service so people have stuff to play against :) 2010-11-08T21:19:31 yep, 3 majors done and dusted 2010-11-08T21:19:37 now for honours and then post grad 2010-11-08T21:19:42 wow, congratulations! 2010-11-08T21:19:50 now start working on your bot 2010-11-08T21:19:52 :P 2010-11-08T21:19:53 i'll be starting a complete rewrite later 2010-11-08T21:20:14 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2010-11-08T21:20:21 *** bhasker has joined #aichallenge 2010-11-08T21:20:24 i have nothing but that scheduled till the end of the contest now :) 2010-11-08T21:27:04 *** Utkarsh has joined #aichallenge 2010-11-08T21:28:20 *** cfaftw has quit IRC (Quit: Leaving) 2010-11-08T21:36:16 *** davidjliu has quit IRC (Ping timeout: 265 seconds) 2010-11-08T21:44:23 *** Utkarsh has quit IRC (Ping timeout: 250 seconds) 2010-11-08T21:49:43 *** amstan has joined #aichallenge 2010-11-08T21:49:43 *** ChanServ sets mode: +o amstan 2010-11-08T21:59:44 *** tapwater_ has joined #aichallenge 2010-11-08T22:01:04 *** Utkarsh has joined #aichallenge 2010-11-08T22:02:17 *** amstan has quit IRC (Ping timeout: 252 seconds) 2010-11-08T22:02:44 *** tapwater has quit IRC (Ping timeout: 272 seconds) 2010-11-08T22:02:46 *** tapwater_ is now known as tapwater 2010-11-08T22:06:42 *** amstan_ has joined #aichallenge 2010-11-08T22:06:42 *** ChanServ sets mode: +o amstan_ 2010-11-08T22:10:42 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-11-08T22:11:07 *** kjdf has left #aichallenge 2010-11-08T22:15:59 Top 10 players: bocsimacko(4244), bix0r4ever(4086), george(4054), shangas(4029), rsergio(4019), GreenTea(4005), Accoun(3979), Xrillo(3969), luksian(3965), dmj111(3954) 2010-11-08T22:18:51 *** Utkarsh has joined #aichallenge 2010-11-08T22:21:55 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-11-08T22:29:20 *** Utkarsh has joined #aichallenge 2010-11-08T22:31:18 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2010-11-08T22:31:37 *** tapwater has quit IRC (Ping timeout: 250 seconds) 2010-11-08T22:32:37 *** tapwater has joined #aichallenge 2010-11-08T22:33:20 *** Prillicy has joined #aichallenge 2010-11-08T22:33:30 I can see a clear downside to jumping almost 2000 ranks.... 2010-11-08T22:33:40 now I look like I suck compared to the bots I'm fighting 2010-11-08T22:36:10 i have paper and pacer out, i'm working out how to do my reimplementation properly :) 2010-11-08T22:36:21 i'ma get this resolution business right 2010-11-08T22:37:22 *** tapwater_ has joined #aichallenge 2010-11-08T22:39:22 demerzel_: If you use linux, and native code or python or perl, i can test your bot in my server, in case you don't have a debugging server. Or does your bot work now? 2010-11-08T22:39:31 *** Utkarsh has joined #aichallenge 2010-11-08T22:40:36 no i've narrowed it down to a piping issue 2010-11-08T22:40:42 ie if i type to it it works fine 2010-11-08T22:40:46 but piping somehow breaks the buffer 2010-11-08T22:40:51 i'm not sure if i'm hitting EOF or what 2010-11-08T22:40:53 *** tapwater has quit IRC (Ping timeout: 260 seconds) 2010-11-08T22:40:54 *** tapwater_ is now known as tapwater 2010-11-08T22:41:30 i have written some pipe code for my server code (child process communication) 2010-11-08T22:41:30 it's not limited to linux either apparently - just using iouri's visualizer breaks it too 2010-11-08T22:41:44 around 800char - and the standard is 512 - somehow i think it's overrunning the buffer 2010-11-08T22:41:55 have you? is it C or cpp 2010-11-08T22:42:15 i'm thinking to try setting a linebuf proxy but if you know a direct way that would be awesome 2010-11-08T22:42:21 this is a weird weird thing to get stuck on lol 2010-11-08T22:42:23 a pipe is usually the size of a page, and that is 4096 bytes in my kernel, my code is c/c++ 2010-11-08T22:43:03 well i'm mostly interested in making sure the buffer doesn't overrun, i'm gonna try a linebuf thing here in a sec and if that doesn't work i'll try to adapt yours if you're feeling nice ;) 2010-11-08T22:43:49 *** Utkarsh has quit IRC (Ping timeout: 245 seconds) 2010-11-08T22:43:49 http://pastebin.com/1xHGetne 2010-11-08T22:44:57 I was a new to this stuff too and it cost me hours to get it right. I often ended up in dead-locks, but it seems to work well now as long as nothing unexpected happens ^^ 2010-11-08T22:46:13 It is missing the class-declaration and the logger class, but I think you'll find through. If not, just ask. 2010-11-08T22:49:44 *** Utkarsh has joined #aichallenge 2010-11-08T22:50:05 *** virdo has joined #aichallenge 2010-11-08T22:50:31 the wait_for_data() method is where it tries to read from the child process' stdout & stderr. You can change that to read from stdin only. Basically it is checking for a HUP (indicating the pipe was closed by the other side) and IN (indicating the arrival of data). On internet sockets caution must be taken. POLIN can fire even if the arrived data was rejected due to checksum errors or something, so a subsequent read may han 2010-11-08T22:50:31 was blocking. 2010-11-08T22:51:03 what's funny is my socket code was flawless haha 2010-11-08T22:51:12 ok i'll look at that since setvbuf didn't work for me 2010-11-08T22:55:39 *** Prillicy has quit IRC (Quit: I be bangin') 2010-11-08T22:56:46 *** wh1teside_ has quit IRC (Ping timeout: 240 seconds) 2010-11-08T22:57:29 *** wh1teside has quit IRC (Ping timeout: 265 seconds) 2010-11-08T22:58:51 *** McLeopold has joined #aichallenge 2010-11-08T22:59:29 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T23:03:46 *** Utkarsh has joined #aichallenge 2010-11-08T23:05:42 i honestly have no idea what's going on in your code 2010-11-08T23:05:48 i am so frustrated this is nuts 2010-11-08T23:06:00 *** phreeza_ has quit IRC (Quit: Computer has gone to sleep) 2010-11-08T23:08:12 *** Prillicy has joined #aichallenge 2010-11-08T23:08:17 is the site down for anyone else or is it just me? 2010-11-08T23:08:24 *** Utkarsh has quit IRC (Ping timeout: 264 seconds) 2010-11-08T23:08:35 its down 2010-11-08T23:08:41 down for a bit 2010-11-08T23:08:44 k 2010-11-08T23:08:47 going to bed 2010-11-08T23:08:56 oh god, i never did that last night 2010-11-08T23:08:58 database backup is about %20 done 2010-11-08T23:09:10 it's 3pm, i'm going to drop dead/asleep at some point 2010-11-08T23:09:33 will finish all of the fixes to my bot tomorrow hopefully and then we'll see how this strategy actually should would have placed me if I got the code right the first 4 times 2010-11-08T23:09:55 *** Prillicy has quit IRC (Client Quit) 2010-11-08T23:11:55 ahhh, i remember those thoughts 2010-11-08T23:14:49 *** Utkarsh has joined #aichallenge 2010-11-08T23:15:47 My bugs are slowly getting easier to find and fix... 2010-11-08T23:16:09 amstan: Error: There was a problem accessing the interface to ai-contest.com 2010-11-08T23:21:53 amstan: ignore that 2010-11-08T23:21:56 nothing fucking works - not fflush not fseek not bufsize not reassigning to to a char* i am totally out of ideas 2010-11-08T23:22:01 raedwulf help damnit 2010-11-08T23:22:47 demerzel_, can you pastebin your stdin reading code? 2010-11-08T23:22:59 there's nothing wrong with it 2010-11-08T23:23:02 *** ratatata has joined #aichallenge 2010-11-08T23:23:12 the problem is it's overrunning whatever stdin buffer there is when it's piped 2010-11-08T23:23:13 or whatever it is that's failing ^^ 2010-11-08T23:23:22 i can type to it just fine 2010-11-08T23:23:27 I have an ubuntu setup, I can try to run it 2010-11-08T23:23:39 but if it's piped or fed or whatever you want to call it (ie sandboxed) then it overruns the buffer 2010-11-08T23:23:52 it happens in windows and linux 2010-11-08T23:24:04 hmm 2010-11-08T23:24:09 you'd think it'd be something i could google easily but i have read article after article 2010-11-08T23:24:14 how often do you read from the pipe? 2010-11-08T23:24:25 i've tried every method of reading from freadf to fgetc to getchar to blahblah 2010-11-08T23:24:27 er, stdin as it appears to you 2010-11-08T23:24:36 lavalamp which is your latest version? 2010-11-08T23:24:45 pwwax.12 ? 2010-11-08T23:24:47 anywhere from one char at a time to a line at a time to trying to read the whole thing 2010-11-08T23:24:48 or is it the gromble one 2010-11-08T23:24:51 hehe I'm running a few huh! 2010-11-08T23:24:57 pwwax.12 is my first bot 2010-11-08T23:25:05 oh k 2010-11-08T23:25:07 lavalamp_: basically it's a piped stdin issue and i can't figure it out :( 2010-11-08T23:25:09 gromble.sub.0 is my current submission 2010-11-08T23:25:19 my latest bot doesn't run >.< 2010-11-08T23:25:23 hehe 2010-11-08T23:25:29 which one is grumble-gen.12 2010-11-08T23:25:30 been stuck on this for literally 4 hours 2010-11-08T23:25:33 gromble* 2010-11-08T23:25:40 whats the issue demerzel 2010-11-08T23:25:47 That's the almost-latest one 2010-11-08T23:26:11 demerzel_, if you can get it into a small file I will see if I can figure it out 2010-11-08T23:26:15 c stdin piped overruns whatever buffer in linux or windows 2010-11-08T23:26:27 oh i've got it up one sec 2010-11-08T23:26:39 somebody was trying to help me earlier to no avail haha 2010-11-08T23:27:08 lol 2010-11-08T23:27:47 this was a quick and dirty one 2010-11-08T23:27:48 http://pastebin.com/Y5ys6SC0 2010-11-08T23:28:00 but i've used getc fgets fscanf blahblah 2010-11-08T23:28:12 bhasker, yeah gromble-gen.12 is my move generator with all my current bug fixes and it sucks, apparantly, lol 2010-11-08T23:28:15 removed state and parsing entirely and it's still broken 2010-11-08T23:28:22 hehe yea :-\ 2010-11-08T23:28:23 it's always around 7-800 chars and the standard is 512 2010-11-08T23:28:44 and like i said i can type to it just fine - for some reason piping doesn't clear th buffer 2010-11-08T23:29:22 you can't poll too quickly 2010-11-08T23:29:37 can you show where you create/allocate the buffer? 2010-11-08T23:29:41 it will lock the stream too fast 2010-11-08T23:29:54 it's just a char[128] extern / declared above it 2010-11-08T23:29:59 *** superflit has joined #aichallenge 2010-11-08T23:30:11 raedwulf so what is the solution 2010-11-08T23:30:29 i've learned more about filestreams in c than i thought was possible to know in teh last 4 hours and still got nothing 2010-11-08T23:30:54 *** McLeopold has quit IRC (Ping timeout: 240 seconds) 2010-11-08T23:30:56 that was my first thought that i was hitting an eof or something 2010-11-08T23:31:39 *** McLeopold has joined #aichallenge 2010-11-08T23:31:57 but i have no way to know :( i got it spitting out this from that code: http://pastebin.com/wVXrkGy7 2010-11-08T23:32:56 fflush although it's not compliant does nothign except make it not work at all, same with fseek, same with assigning linebuf to a char* 2010-11-08T23:33:12 wait wtf your bot is in c lemme see your stdin stuff lol 2010-11-08T23:33:48 to this point i'd been playing exclusively on tcp via socket code so never had to make stdin/out till today 2010-11-08T23:34:07 had NO idea it would be such a nightmare 2010-11-08T23:34:11 demerzel_: i had to delay it a bit 2010-11-08T23:34:20 what are you having so much trouble with? 2010-11-08T23:34:24 i read/parsed per line 2010-11-08T23:34:25 it shouldn't be that complicated 2010-11-08T23:34:31 delay it how 2010-11-08T23:34:36 seemed to work 2010-11-08T23:34:48 antimatroid1: i promise you if there was an easy solution available to me i would have taken it hours ago 2010-11-08T23:34:52 but i couldn't read an entire buffer until go 2010-11-08T23:34:58 raedwulf delay it how 2010-11-08T23:35:07 i just did some processing per line 2010-11-08T23:35:16 and read per line 2010-11-08T23:35:26 i do process - it creates the planets 2010-11-08T23:35:33 but then again this is on linux - might be different issue 2010-11-08T23:35:34 it's a per-line read :( 2010-11-08T23:35:44 nah it has same problem on main server 2010-11-08T23:35:56 so i need to slow my parser down somehow awesom 2010-11-08T23:36:06 i only create my planet objects once 2010-11-08T23:36:14 hell that's turn 1 - it's creating all the planets and getting doubles etc 2010-11-08T23:36:20 it's even faster turn 2+ 2010-11-08T23:36:38 any ideas raedwulf? 2010-11-08T23:36:52 demerzel_: lemme check what i used in the end 2010-11-08T23:36:56 shouldn't getchar block wtf is 'locking' the steam 2010-11-08T23:37:05 stream* 2010-11-08T23:37:13 demerzel_: i used fgets 2010-11-08T23:37:21 never had an issue where my program was too fast lol 2010-11-08T23:37:30 yea that's what i started with raed 2010-11-08T23:37:48 hmmm 2010-11-08T23:37:49 moved to per char so i could see that it was getting anythign at all 2010-11-08T23:38:43 fflush just completely kills it 2010-11-08T23:39:02 what is 'locking' the stream maybe i can go from there cuz i'm getting nowhere with the buffer angle 2010-11-08T23:39:11 tried every damn solution out there 2010-11-08T23:39:53 try fgets again? 2010-11-08T23:40:00 *** hellman has joined #aichallenge 2010-11-08T23:40:27 haha i already have :P 2010-11-08T23:40:36 what's the ascii code for EOF? 2010-11-08T23:40:50 lavalamp_: from what function? 2010-11-08T23:40:55 what OS 2010-11-08T23:41:01 demerzel_ your code seems to work fine for me :/ 2010-11-08T23:41:09 well ubuntu atm 2010-11-08T23:41:48 ctrl+d but if you're programming, all fgetc will give you EOF, fgets will give you a NULL string and feof will be 1 2010-11-08T23:42:09 what does getchar() return? 2010-11-08T23:42:10 10 2010-11-08T23:42:28 lavalamp_: no idea :) two sec 2010-11-08T23:42:29 i turned off buffer completely via setvbuf(stdin, NULL, _IONBF, 0); 2010-11-08T23:42:35 lavalamp_ 10 2010-11-08T23:42:42 it still doesn't work lol 2010-11-08T23:42:44 lavalamp_: EOF 2010-11-08T23:42:52 oh eof is -1 2010-11-08T23:43:02 yes but its better to use the constant 2010-11-08T23:43:05 ascii 04 wikipedia says 2010-11-08T23:43:26 well maybe it's just this bot player who knows i'll try uploading it in the am 2010-11-08T23:43:28 lavalamp_: but thats probably not what C returns - its defined in the C standard for the constant EOF 2010-11-08T23:43:29 *** Utkarsh has quit IRC (Ping timeout: 252 seconds) 2010-11-08T23:43:40 i am totally out of energy on this - 4+ hours of googling +c +stdin + piped +help 2010-11-08T23:44:02 my problems only lasted 2 hours :S 2010-11-08T23:44:09 ok 2010-11-08T23:44:38 lucky 2010-11-08T23:44:43 lavalamp_: you never get EOF anyway 2010-11-08T23:44:44 i never thought i'd get a hangup on stdin ffs 2010-11-08T23:44:56 typed to it etc to make sure it was working 2010-11-08T23:44:58 its useful for self-testing but the engine will never give you EOF - it just kills it 2010-11-08T23:45:00 http://pastebin.com/sCzuNrJp 2010-11-08T23:45:04 and hey it is ok then 2010-11-08T23:45:08 see if that program fails for you 2010-11-08T23:45:38 I piped a big replay file through it and it seemed to work fine 2010-11-08T23:46:04 well it breaks in windows and on the main server :( 2010-11-08T23:46:09 lavalamp_: usually piping works because its faster than what the engine sends through 2010-11-08T23:46:16 at the end it started spitting out eofs which is why I asked 2010-11-08T23:46:19 but maybe doing it one char at a time + turning buffer off will make ti work on the server at least 2010-11-08T23:46:37 raedwulf ok I'll test that 2010-11-08T23:47:54 demerzel_: http://pastebin.ca/1986003 2010-11-08T23:48:16 http://pastebin.ca/1986004 2010-11-08T23:49:14 what is that i thought you used fgets 2010-11-08T23:49:42 demerzel_: check the first pastebin 2010-11-08T23:49:50 first pastebin is part of the "driver" 2010-11-08T23:49:59 and then protocol_parse is just what i use for parsing 2010-11-08T23:50:05 oh right 2010-11-08T23:50:09 yea that looks a lot like mine 2010-11-08T23:50:14 why not scanf? lol 2010-11-08T23:50:49 *** Utkarsh has joined #aichallenge 2010-11-08T23:51:03 i suppose i could have done that 2010-11-08T23:51:28 but this solution worked ^^ 2010-11-08T23:51:32 http://pastebin.com/ng711edr this what parseline feeds 2010-11-08T23:51:36 so it's doing parsing per line like yours 2010-11-08T23:53:00 my original solution ws exactly yours in fact 2010-11-08T23:53:09 one line fgets ok that was easy 2010-11-08T23:53:21 then i uploaded it and it crashed and i was like wtf and ran it through bot tester and it crashed 2010-11-08T23:53:49 so i took state and parsing out of the equation completely and it still crashe 2010-11-08T23:54:05 I accidently modified the wrong version of my code. 2010-11-08T23:54:10 Now I'm merging... 2010-11-08T23:54:26 ...and I don't remember writing this stuff. O_o 2010-11-08T23:54:47 yea i did that a few times 2010-11-08T23:54:49 But, it's brilliant! 2010-11-08T23:54:57 demerzel_: :S 2010-11-08T23:55:02 god this is so stupid and yet completely blocking me from doing anything 2010-11-08T23:55:11 i know my code works because it works fine with rage bot thats still running on the main server 2010-11-08T23:55:25 well wtf then 2010-11-08T23:55:30 demerzel_: it's just your parsing code, and that's it? 2010-11-08T23:55:37 it's crashing on the server i have the exact same code except with an fgets 2010-11-08T23:55:52 demerzel_: checked the buffer size? 2010-11-08T23:55:57 no buffer overflow? 2010-11-08T23:56:05 i have mine at 128 per line 2010-11-08T23:56:09 should be enough 2010-11-08T23:56:15 hmm 2010-11-08T23:56:22 what is 'line' in your implementation 2010-11-08T23:56:33 remember windows doesn't prompt for memory bugs as often as linux 2010-11-08T23:56:57 char line[PROTOCOL_LINE_LENGTH]; 2010-11-08T23:56:59 my code on the server is one line - fgets(stdin, 128, io_output) 2010-11-08T23:57:16 ah ok so just like mine except with a define 2010-11-08T23:57:19 well then i really have no damn idea 2010-11-08T23:57:29 demerzel_: stdin is defined in the C headers 2010-11-08T23:57:38 not wise to override that name 2010-11-08T23:57:50 i even ran splint on it 2010-11-08T23:57:52 lol 2010-11-08T23:58:00 you sure thats the correct line? 2010-11-08T23:58:14 char *fgets(char *s, int size, FILE *stream); 2010-11-08T23:58:24 stream should be at the end of the function 2010-11-08T23:58:29 no i just typed that up 2010-11-08T23:58:32 ok 2010-11-08T23:58:34 ^^ 2010-11-08T23:58:45 ah yes then fgets(io_output, 128, stdin) 2010-11-08T23:58:56 lemme make sure that isn't the only problem cuz then i'll feel stupid but at least i'm done lol