2010-12-18T00:07:26 *** amstan_ has joined #aichallenge 2010-12-18T00:07:26 *** ChanServ sets mode: +o amstan_ 2010-12-18T00:14:24 *** Frontier has left #aichallenge 2010-12-18T00:15:03 *** Frontier has joined #aichallenge 2010-12-18T00:44:38 *** amstan_ is now known as amstan 2010-12-18T02:05:33 *** amstan has quit IRC (Remote host closed the connection) 2010-12-18T02:21:14 *** Palmik has joined #aichallenge 2010-12-18T02:40:41 i think i'm almost done 2010-12-18T02:40:51 other than each program seems to use up more memory the longer it runs 2010-12-18T02:40:56 i can't work out why 2010-12-18T03:10:10 *** iNtERrUpT has joined #aichallenge 2010-12-18T03:14:07 *** choas has joined #aichallenge 2010-12-18T03:31:27 ah memory leaks... fun 2010-12-18T03:40:31 i swear i'm not using any pointers : 2010-12-18T03:40:48 i'm thinking somethings happening with my streams 2010-12-18T03:40:53 it's a huge jumble of mess for that 2010-12-18T03:42:11 sigh: interested in taking a look? :P 2010-12-18T03:42:21 you get to watch the game play too 2010-12-18T03:42:27 only an ascii version of it though 2010-12-18T03:42:29 let valgrind look at it first 2010-12-18T03:43:19 you say that like i've ever used it :P 2010-12-18T03:44:13 it's not that hard :P 2010-12-18T03:45:50 i'll have a quick look, if you want 2010-12-18T03:45:56 Frontier: I don't really support ants (am working on something else right now), and i only have 1 process per player. Also all the source is in the download 2010-12-18T03:52:02 just cleaning up something i can zip 2010-12-18T03:52:30 setup a public repo already :P 2010-12-18T03:57:13 meh, google docs will do for the moment 2010-12-18T03:58:32 i've included an arena folder, hopefully you should be able to cd there and then just enter 2010-12-18T03:58:32 engine.exe "amap.txt" "receiver.exe" "bot.exe" "bot.exe" "6" "50" "1.1" 2010-12-18T03:58:43 https://docs.google.com/leaf?id=0B5joEFt1EpwZZjZiN2M0NjUtZDY4Yy00M2YyLThlYjYtMGFiZThiZmEyMzY5&sort=name&layout=list&num=50 2010-12-18T04:00:44 the code for engine is horrible 2010-12-18T04:00:58 but i think it's at least spaced out to make sense for what spots are doing what 2010-12-18T04:01:18 is there a makefile or something? 2010-12-18T04:01:30 i have no idea how to write them, sorry :P 2010-12-18T04:02:05 i could rezip with codeblocks project files? 2010-12-18T04:03:45 hmmm 2010-12-18T04:05:29 mah... there's a bunch of windows stuff that isn't protected by ifdefs 2010-12-18T04:05:44 is it just #include "windows.h"? 2010-12-18T04:05:50 i couldn't work out how to not have that outside 2010-12-18T04:06:17 otherwise,ahhhh crap 2010-12-18T04:10:59 Process.cc:67: warning: missing sentinel in function call 2010-12-18T04:11:10 hmm 2010-12-18T04:11:21 the joys of C++ 2010-12-18T04:11:38 hmm, i didn't write that 2010-12-18T04:11:54 SmkMnstr: you about? 2010-12-18T04:13:19 nvm, I fixed it 2010-12-18T04:13:37 did it fully compile? because that should be the only problemish part 2010-12-18T04:13:44 well, and it running as expected 2010-12-18T04:14:55 ok, I got it compile 2010-12-18T04:15:33 i think i left the outputting of an ant square to debug.txt as broken, but that doesn't really matter 2010-12-18T04:15:57 game.txt contains almost full information of the board at each state (don't see harvest or hp etc.) 2010-12-18T04:16:29 uncomment lines 137-139 of Engine.cc if you want to see what is in game.txt to the console each turn 2010-12-18T04:17:27 antimatroid: If you have a replay file for me that would be great. Something that includes all the special cases as well. 2010-12-18T04:17:29 it seems to be running 2010-12-18T04:18:37 ok, it ran :) 2010-12-18T04:19:12 Frontier: just a minute 2010-12-18T04:19:18 sigh: it's slow as fuck aye :P 2010-12-18T04:19:24 yeah 2010-12-18T04:19:56 Frontier: are you taking the final line 2010-12-18T04:19:56 fin winner 2010-12-18T04:19:56 atm? 2010-12-18T04:20:48 Frontier: http://pastebin.com/T4xZYmRd 2010-12-18T04:20:53 you may want to edit/remove that last line 2010-12-18T04:21:17 oh, and did you want direction for walk? 2010-12-18T04:21:23 i gave it to you 2010-12-18T04:22:33 http://pastebin.com/2UNY4Nrf 2010-12-18T04:22:44 that's the format i currently pass to the visualiser with 2010-12-18T04:22:48 antimatroid: At the moment I'm not taking any of that ^^. I thought it would be wise to wait till you have something that works for you. 2010-12-18T04:23:20 (i think) take a look at that sample file and see if it matches 2010-12-18T04:24:03 sigh: i have a feeling i made more of a mess with string streams than just taking a char at a time, i may try changing all that after and see if it changes anything 2010-12-18T04:24:13 That's difficult to see. In the specs the direction is given twice for "walk" 2010-12-18T04:24:35 ?? 2010-12-18T04:25:06 Round 1 and 3 both have a direction field, but it should always be the same, correct? 2010-12-18T04:25:26 i have a feeling i'm not sending you walk rotation moves till later 2010-12-18T04:25:35 i can fix that 2010-12-18T04:26:20 yeah, i'm currently only sending you walk information at round 3, including rotation 2010-12-18T04:28:01 i found a neat little trick for the starter bot home path, assing each square a direction home, start them all at 'X', and every time an ant move into a square with 'X', mark the direction that to which you just came from, then any ant of yours has a (very inefficient) path back to base mapped out 2010-12-18T04:28:46 I see. I get the initial rotation from the "A" lines so I don't need any for turn 1, but on turn 2 following it would be good to have the direction for attacks/walk so that the visualization makes sense 2010-12-18T04:29:48 wait, is for each start ant meant to be just at turn 0, or every turn? 2010-12-18T04:29:49 antimatroid: I think the starter bot is coming along fine :p. 2010-12-18T04:30:12 only turn 0 2010-12-18T04:30:16 okay good 2010-12-18T04:30:47 it's just so I have something to initialize the graphics with. 2010-12-18T04:30:47 then round 1, i send you info about deposit and attack rotations atm, then round 3 walk rotation information (and walk information) 2010-12-18T04:31:18 i don't give you any information about spawning ants either, can you just do it if there is grains on a home base after round 2? 2010-12-18T04:31:20 ok that is probably better than a separate rotation line anyway 2010-12-18T04:32:10 You don't give me a changed square for a spawning ant? Ok, I think I can write a quick hack for that. 2010-12-18T04:32:47 i do give you a changed square for it, but i'm not sure if you wanted to animate the spawn before resolving conflicts? 2010-12-18T04:32:55 you get information about that square after the resolution 2010-12-18T04:34:00 sigh: how does the example game look? :P 2010-12-18T04:34:34 Frontier: http://pastebin.com/TszZfnhe 2010-12-18T04:34:42 that's what an ascii example game looks like 2010-12-18T04:40:58 i'll have a play around with popen in python later and see if i can't get that working 2010-12-18T04:41:40 did any of you ever play future cop lapd? 2010-12-18T04:42:19 http://www.youtube.com/watch?v=W2rPuRDtL_s 2010-12-18T04:42:29 that should jog your memory, that is what i think my ant game feels like 2010-12-18T04:43:19 antimatroid: I can still animate the spawn of ants by adding the animation to the previous turn ;) And no, I never played future cop lapd. I liked to play the demo of this game, though: http://www.youtube.com/watch?v=7KWsywyJtjA 2010-12-18T04:43:27 what are you getting working in python? 2010-12-18T04:44:22 sigh: possibly something cleaner 2010-12-18T04:45:00 might want to look at the ai contest repo then... the SadBox class there does all the process talky stuff 2010-12-18T04:52:07 *** choas has quit IRC (Ping timeout: 250 seconds) 2010-12-18T04:55:01 *** boegel has joined #aichallenge 2010-12-18T04:57:46 *** fgump has joined #aichallenge 2010-12-18T05:01:07 *** boegel has quit IRC (Quit: Leaving) 2010-12-18T05:03:12 *** tapwater has quit IRC (Quit: tapwater) 2010-12-18T05:10:42 oh, i also think i might be letting the bots send multiple moves for their (or their enemies!) ants at once 2010-12-18T05:10:45 i will fix that later 2010-12-18T05:20:25 * Frontier looks at the PlanetWars visualizer and realizes how one really writes JavaScript applications. 2010-12-18T05:38:31 hehe, it can't be worse than my engine code 2010-12-18T05:39:56 *** choas has joined #aichallenge 2010-12-18T05:52:09 *** jaspervdj has joined #aichallenge 2010-12-18T06:00:48 Frontier: nice visualizer I saw 2010-12-18T06:01:40 Error323: hopefully in a day or so it will actually be showing a real game :) 2010-12-18T06:02:29 antimatroid: :D 2010-12-18T06:02:37 antimatroid: ur working on that? 2010-12-18T06:03:15 yeah, i have a semi working engine 2010-12-18T06:03:23 cool 2010-12-18T06:03:27 http://pastebin.com/TszZfnhe 2010-12-18T06:03:39 that's very bad view of a sample game that just goes for 30 turns 2010-12-18T06:05:36 yeah i'm not gonna parse that just now, but i'm happy u got it almost working :) 2010-12-18T06:06:42 Error323: hehe. thx for the compliment. i am on the right track with the visuals then :) 2010-12-18T06:07:40 *** medrimonia1 has joined #aichallenge 2010-12-18T06:07:46 yeep, looks real nice 2010-12-18T06:08:12 I love it how this contest manages to keep using javascript for their visualizations. 2010-12-18T06:08:22 Instead of turning to flash 2010-12-18T06:09:09 *** medrimonia has quit IRC (Ping timeout: 260 seconds) 2010-12-18T06:14:56 *** delt0r___ has quit IRC (Read error: Connection reset by peer) 2010-12-18T06:16:12 *** delt0r___ has joined #aichallenge 2010-12-18T06:24:21 *** choas has quit IRC (Ping timeout: 272 seconds) 2010-12-18T07:11:03 sigh: what did you end up having to change before? 2010-12-18T07:11:40 and any idea how i can get the one file to compile across systems? ie. how do you put in #include "windows.h" when you need that for the #ifdef 2010-12-18T07:13:09 just put the #include "windows.h" inside a #ifdef _WIN32 2010-12-18T07:13:54 also had to change the last arg to execl in Process.cc:67 from 0 to NULL 2010-12-18T07:14:29 then it doesn't compile :P 2010-12-18T07:14:46 i think it needs that to have _WIN32 be defined 2010-12-18T07:14:52 hence why i had it outside there 2010-12-18T07:15:38 then how does Process.cc work? it uses #ifdef _WIN32 2010-12-18T07:16:07 yes, but if i move the include windows.h it falls through to the #else 2010-12-18T07:16:57 wait, neverming, got it 2010-12-18T07:17:01 nevermind* 2010-12-18T07:19:07 *** nullkuhl has joined #aichallenge 2010-12-18T07:26:09 should the bots receive who won at the end too? 2010-12-18T07:26:43 could be helpful for people training 2010-12-18T07:36:12 antimatroid: whats this ? a new contest ? 2010-12-18T07:36:26 nullkuhl: just a possible game for the next contest 2010-12-18T07:36:35 antimatroid: and what is the game ? 2010-12-18T07:36:41 antwars 2010-12-18T07:36:55 http://pastebin.com/TszZfnhe 2010-12-18T07:37:01 that's a very basic 30 turn sample game 2010-12-18T07:37:30 antimatroid: ...that is not a very good sample to get the idea what the game is like ^^ 2010-12-18T07:37:38 players bases are indicated start with 9 grains and an ant on them, then you need to go harvest grains and deposit them at home for more ants to be spawned 2010-12-18T07:37:52 Frontier: it's sort of the best i have atm :P 2010-12-18T07:38:07 if one scrolls down they can see the ants spreading out 2010-12-18T07:38:52 it works better when the console is printing it in the same position every second 2010-12-18T07:39:00 if one scrolls in sync to the monitor to get the shutter effect... maybe :p 2010-12-18T07:43:23 *** choas has joined #aichallenge 2010-12-18T07:47:25 it will be a million times better once i can watch it in a your visualiser though :) 2010-12-18T07:47:34 i started calling it gui with all my stuff, it's shorter 2010-12-18T07:56:39 I'll put a text field on the web site where you can paste your replays for the JavaScript to pick them up. 2010-12-18T07:57:53 there's a website? :P 2010-12-18T07:57:57 and awesome 2010-12-18T07:58:35 got it 2010-12-18T07:59:08 :p 2010-12-18T08:00:48 *** Naktibalda has joined #aichallenge 2010-12-18T08:02:38 when i generate a sample set of 100 maps i've make sure there's a decent bit of variation with wall density 2010-12-18T08:02:46 i'll* 2010-12-18T08:03:30 that sounds good, sizes will also be from 16 to 32, right? 2010-12-18T08:03:41 i was doing 40x40 atm 2010-12-18T08:04:25 people like maxing at 32 don't they? 2010-12-18T08:04:47 I can't fit that on a laptop monitor with 1280x800 as long as I make a square 20x20 px. 2010-12-18T08:05:07 what's your max? 2010-12-18T08:05:27 i have a big ass 24 inch, so i don't think about those things :P 2010-12-18T08:05:31 32x32 works good 2010-12-18T08:05:39 okay, i'll make them all 32x32 2010-12-18T08:06:07 i don't think it really changes much if the overall dimensions change up 2010-12-18T08:06:35 *** mega1 has joined #aichallenge 2010-12-18T08:06:57 it is about the size of the planetwars visualizer then 2010-12-18T08:08:12 I thought about adding a border to the map, so you can see what is happening behind the wrap 2010-12-18T08:08:27 so repeating data? 2010-12-18T08:08:47 might just be more confusing i think 2010-12-18T08:09:17 Right, but shaded darker so you can distinguish it from the 'real' map 2010-12-18T08:09:42 maybe worth a shot, although if the map could be bigger, i'd probably prefer to do it with new squares :P 2010-12-18T08:10:31 Well I guess I could always zoom out the canvas until it fits the screen :p 2010-12-18T08:10:51 that's possibly not such a bad idea if easy enough :P 2010-12-18T08:11:02 we can still use smaller dimensions if it's considered better 2010-12-18T08:11:40 I don't know if I get bad ass drawing artifacts from the fractional coordinates though 2010-12-18T08:12:09 i think it looks pretty good atm size wise 2010-12-18T08:12:17 i wouldn't want to make them much smaller for readability 2010-12-18T08:13:11 how long do you gess a game of 40x40 could be? 2010-12-18T08:13:18 *guess 2010-12-18T08:13:23 probably too long 2010-12-18T08:14:17 my intentions are for the games to be along the lines of 2010-12-18T08:14:18 go out and discover the map and build up grains on the home base for army 2010-12-18T08:14:18 big epic battle as players try to invade each others bases and defend 2010-12-18T08:14:54 epic... well, how can I help in making it epic? 2010-12-18T08:15:03 explosions? 2010-12-18T08:15:23 metal shreds flying through the air? 2010-12-18T08:15:38 ant sounds? 2010-12-18T08:16:14 hmmm, not sure 2010-12-18T08:16:26 my thoughts are mostly on the strategy being interesting 2010-12-18T08:17:09 i think the optimal thing to do when multiple opponent ants are near each other switches all over the place here 2010-12-18T08:17:24 which is what i think gives it the same kind of strategic complexity that tron and planet wars had 2010-12-18T08:17:55 you could maybe have ants that land on the same square do some kind of fight animation? 2010-12-18T08:18:16 and if an ant moves onto the enemies base, explode the base and all ants? 2010-12-18T08:18:33 That is the most difficult situation because they actually just disapear as soon as they get there :p 2010-12-18T08:19:20 to the enemies base? ah yeah, you would need to do it once being sent fin i think 2010-12-18T08:19:36 i think it's best if you move there their ant spawns before collision fights/fusion 2010-12-18T08:20:23 ...no I mean when ants move onto the same square I cannot do play an animation as they just get merged 2010-12-18T08:20:43 i send you the move information at round 3, then updated square information in round 4 2010-12-18T08:20:48 couldn't you do it between those two rounds? 2010-12-18T08:21:23 that's why i'm doing those rounds, so we can see the ants move on top of each other before the morph 2010-12-18T08:21:43 antimatroid: Actually I'm still trying to get the movement smooth so I recombine all rounds into one. If you really insist I can split it up again. 2010-12-18T08:22:00 it's your choice :) 2010-12-18T08:22:09 but i think it's better if we can see what each ant did 2010-12-18T08:22:34 actually, no do it smoothly 2010-12-18T08:22:40 your way is better :) 2010-12-18T08:22:51 cause we will still see them walk etc. yeah? 2010-12-18T08:22:52 Well you see that. The two ants move onto one square.... and then *pop* they are one. 2010-12-18T08:22:59 yeah, that's cool 2010-12-18T08:23:16 the smooth flow of the game looks really good, same with planet wars 2010-12-18T08:24:16 It just needs the little bit of extra information, like the direction an ant is facing, btw. do you add the direction to the "S" lines of round ...what was it 1? now? 2010-12-18T08:25:41 round 1: directions ants face for attacks and deposits 2010-12-18T08:25:41 round 2: new information for squares after attacks, harvests and deposits 2010-12-18T08:25:41 round 3: directions for ants walking (implement these walk moves) 2010-12-18T08:25:41 round 4: information for squares changed after walks/spawning/conflicts 2010-12-18T08:26:03 round 1 is "R", rounds 2 and 4 are "S" and round 3 is "W" 2010-12-18T08:26:37 R row column direction 2010-12-18T08:26:37 W row column direction 2010-12-18T08:26:38 S row column antPlayer antHP antHarvest squareGrains 2010-12-18T08:26:40 hmm, ok. just thought that round 1 and 3 share the same information 2010-12-18T08:27:09 no, i don't process walk moves until after round 2, because an ant that dies from an attack doesn't move 2010-12-18T08:27:10 wait... you wont send a round 3 for an ant that has a round 1, right? 2010-12-18T08:27:17 no i wont 2010-12-18T08:27:41 ok. this is gonna be a mess of a source code :p 2010-12-18T08:27:54 should i be doing that differently? 2010-12-18T08:28:10 otherwise you will see an ant walk that dies before walking 2010-12-18T08:28:15 no im happy we have something we can agree on. 2010-12-18T08:29:06 i'm flexible, i just need to be convinced another way is better :) 2010-12-18T08:29:08 i think it would be easier for me if i could identify the objects as ants, not as "square x changed". 2010-12-18T08:29:21 if i am convinced a particular way works best, i get really stubborn 2010-12-18T08:29:37 what id do fused ants get? 2010-12-18T08:30:03 I had a stubborn co-worker. Since then I don't agrue any more. 2010-12-18T08:30:18 i enjoy a good argument :( 2010-12-18T08:30:53 fused ants... hmm... good question. Is it maybe a new ant that just spawned there? 2010-12-18T08:32:10 that could possibly be done 2010-12-18T08:32:12 the ant id would be lastId+1 2010-12-18T08:32:15 is it really getting easier if i do that? :P 2010-12-18T08:33:15 if it makes your life easier i can do it 2010-12-18T08:33:19 I don't want to check 1000 squares to find 20 ants that I want to draw, so I use a list of ants anyway internally 2010-12-18T08:33:22 but i think it makes both the engine and bot's life harder 2010-12-18T08:33:54 Right now when I manage the animations I have to track back origin squares and stuff like that ^^ 2010-12-18T08:34:21 hmmm 2010-12-18T08:34:37 How do you handle that internally? Do you iterate over every square to find own ants in the bot? 2010-12-18T08:35:06 i think i do in the engine atm, some places i store a vector of locations to squares with ants 2010-12-18T08:35:12 i really need to clean everything up 2010-12-18T08:35:15 i can do ant ids 2010-12-18T08:35:17 i'll do that 2010-12-18T08:35:24 hehe, ok 2010-12-18T08:35:31 you respecify gui input and i'll do it to that 2010-12-18T08:39:38 Now I have to make sure I don't end up with a bunch of helper functions again, in other words get it right on the first try. hmm a special "spawn" and "die" command probably. even attack, deposit and harvest could be used for some basic animation. 2010-12-18T08:45:28 i can send you all made moves, but you will get more information 2010-12-18T08:45:56 or do you want to also process moves yourself? 2010-12-18T08:48:42 no, i would like to let the server handle the numbers. 2010-12-18T08:50:20 actually, i don't send you any more information 2010-12-18T08:50:44 and it's almost no difference, i just split R lines up into D and A 2010-12-18T08:50:49 i'll take your word on that 2010-12-18T08:50:50 actually, i also need to send you H lines 2010-12-18T08:51:07 wait a sec 2010-12-18T08:51:36 no, you should wait now, im in the middle of refining the input format :p 2010-12-18T08:51:56 okay :) 2010-12-18T08:52:21 but split round one up into lines starting with 'A', 'D' and 'H' ;) 2010-12-18T08:54:40 Stupid question: Do I need to think in rounds when I merge them all to have a smooth movement? 2010-12-18T08:56:09 ? 2010-12-18T08:57:18 an ant should be able to deposit before dying from attack 2010-12-18T08:57:21 i have that wrong atm 2010-12-18T08:57:21 I might be doing it wrong, but I don't see a need for this abstraction in my visualizer right now 2010-12-18T08:57:42 deposit before attack... hmm 2010-12-18T09:02:19 *** mceier has joined #aichallenge 2010-12-18T09:13:26 antimatroid: http://pastebin.com/Zv9K4b0T - That would solve some problems for me :) 2010-12-18T09:14:18 sorry forgot an id in the move command: http://pastebin.com/CmLYWPhL 2010-12-18T09:15:17 can move be walk? 2010-12-18T09:15:25 i like to call all options a move 2010-12-18T09:15:32 oh yes, sure 2010-12-18T09:15:43 i think it makes it clearer 2010-12-18T09:15:50 alright, give me a little while :) 2010-12-18T09:16:07 should bots receive ant ids? 2010-12-18T09:16:42 what's the point of providing a spawn direction on an ant? 2010-12-18T09:17:16 can you just set it as north and south for each bot and not receive that info? 2010-12-18T09:23:17 Frontier: do you want ant id's to be unique for each player or for all ants? 2010-12-18T09:23:17 I could do that, but it adds control to the way the ants start out 2010-12-18T09:23:53 antimatroid: I had the same question a few days back. It doesn't really matter, but make them global. 2010-12-18T09:24:10 i'm wondering how large that may get 2010-12-18T09:24:16 This way we can later add the brain bug, that assimilates enemy ants. 2010-12-18T09:24:54 I hope it gets smaller this way :p 2010-12-18T09:24:54 you can't have more ants than total starting grains + 2 2010-12-18T09:24:59 we can cap that if necessary 2010-12-18T09:26:22 antimatroid: As a means to make the game replays smaller you can cap the number. Let's see how much data the starter bot produces on a 32x32 map :) 2010-12-18T09:26:52 okay, i still need to work out what is causing memory leak issues 2010-12-18T09:27:14 and i'm hoping that is what causes it to get slower and not the more ants 2010-12-18T09:27:17 use valgrind it helped me a lot 2010-12-18T09:27:24 the bot still seems to return it's moves fairly fast 2010-12-18T09:27:58 as long as it doesn't run O^3 loops its going to stay this way :) 2010-12-18T09:29:00 there are a lot of cool algorithms for painting a discrete grid 2010-12-18T09:29:08 cool == simple 2010-12-18T09:29:36 what do you think starting hp of an ant should be? 2010-12-18T09:29:44 it's currently 5 and i think that's about right 2010-12-18T09:30:41 *** nullkuhl has left #aichallenge ("Konversation terminated!") 2010-12-18T09:31:17 *** yasith has joined #aichallenge 2010-12-18T09:32:06 Difficult to say. I'm more worried about the relation to the 'instagibs' when two ants walk onto the same square :D 2010-12-18T09:34:14 Look how the new format is smaller! Old: "W 28 11 N; S 27 11 1 5 0 8", new: "W 0 N" 2010-12-18T09:36:53 it's harder to process 2010-12-18T09:37:15 when an ant walks onto a square, i now need to track how many walked on and the id if it was just a single ant 2010-12-18T09:37:15 update M -> W, S -> G (was used twice): http://pastebin.com/0d0VmH7u 2010-12-18T09:37:42 i should be able to do it thought 2010-12-18T09:39:15 didn't you have to merge any number of ants on a square before? 2010-12-18T09:44:29 the ids are new, thats true. but it really helps me a lot to know where the life of an ant starts and where it ends. I can now combine two walk commands to a curve so the ants don't always turn in 90°. 2010-12-18T09:45:49 before i just took in new ant info for hp and grains from both players, then combined that with spawned info and saw which player still had an ant there if any 2010-12-18T09:46:02 i just need to tweak it to count the number of ants added 2010-12-18T09:46:13 and create a new ant id if it's more than 1 2010-12-18T09:46:47 any direction i pass you for spawned ants is going to be arbitrary though 2010-12-18T09:48:15 then drop that field for now. I'll see how I can use the new ant ID (!!) to find the direction the ant will use next 2010-12-18T09:49:04 *** yasith has quit IRC (Ping timeout: 276 seconds) 2010-12-18T09:49:14 Can you tell the bots if they won or lost and let them return a victory dance command? 2010-12-18T09:49:42 jeesh did I really propose that? 2010-12-18T09:50:58 yeah, i don't think that's necessary :P 2010-12-18T09:51:10 you could actually do a victory dance with this though 2010-12-18T09:53:52 use F for ants spawned by fusions instead of S. I might want to visualize that case differently 2010-12-18T09:54:06 hmmm i was just about to say 2010-12-18T09:54:17 when i spawn an ant, i do that before collisions 2010-12-18T09:54:31 so if i assign that an id, then it collides, i assign another id 2010-12-18T09:55:29 meaning ids may get used and never even appear 2010-12-18T09:55:31 so... an enemy ant can stand on my base, it spawns a new ant, they collide and my ant 'survives' with 1hp 2010-12-18T09:56:10 i'm not doing it like that atm at all 2010-12-18T09:56:25 so what case did you think of? 2010-12-18T09:56:34 *** sigh has quit IRC (Remote host closed the connection) 2010-12-18T09:56:54 i have newAnts[2][2] 2010-12-18T09:56:54 then spawn and move ant info goes to that, and square info for actual ants that move gets cleared, then i resolve it all in one sweep 2010-12-18T09:57:41 i can maybe do a little hack 2010-12-18T09:57:58 assign newID -1 to a spawned ant, and if that is the carry over ID for 1 new ant, assign a new ID 2010-12-18T10:00:03 can you print the command sequence I would receive for this? 2010-12-18T10:00:14 *** yasith has joined #aichallenge 2010-12-18T10:00:16 what do you mean? 2010-12-18T10:00:22 you don't see any of this 2010-12-18T10:00:31 this is just me thinking out loud how i can resolve it 2010-12-18T10:01:23 so when an ant spawns and it survives the collision it will show as spawned in the usual way using "S"? 2010-12-18T10:01:36 instead of "S" "K" "F" 2010-12-18T10:01:41 hmmm 2010-12-18T10:01:45 i haven't got that far yet 2010-12-18T10:01:53 i'm trying to assign ants an id 2010-12-18T10:01:53 that's what I was asking :p 2010-12-18T10:02:28 i don't see why an ant needs an id? 2010-12-18T10:02:34 We don't need to create any new id for fusions as long as max one of them always survives and all others die 2010-12-18T10:03:16 Frontier: which one gets to keep id? and then i need to store information about all ants that get added 2010-12-18T10:03:22 it really does make it quite a lot harder my end 2010-12-18T10:04:33 yeah forget that idea. The id is usefull to refer to in the replay and for me to keep track of an ant. 2010-12-18T10:06:32 actually, i do the spawning after doing all moves, i know if it's the only ant i add to the base 2010-12-18T10:06:39 therefore i can tell whether i need to assign a ned id 2010-12-18T10:06:41 problem averted :) 2010-12-18T10:06:53 you get a brand new id if any ants land together though 2010-12-18T10:07:53 no problem. the only effect would have been that you see the stronger of two ants 'rush through' 2010-12-18T10:08:33 you can have up to 6 ants end up on a square at once 2010-12-18T10:08:39 5 for normal, 6 on a base 2010-12-18T10:14:08 *** smellyhippy has quit IRC (Quit: BUT WAIT! What are those penguins up to?!) 2010-12-18T10:15:08 *** smellyhippy has joined #aichallenge 2010-12-18T10:15:11 *** bhasker has joined #aichallenge 2010-12-18T10:20:18 http://pastebin.com/j1qFsEdE - added 'F' fusion spawn, removed spawn directions 2010-12-18T10:21:23 hmmm, a lot of these changes require me to store and track a lot more information than before 2010-12-18T10:21:35 and here is me struggeling with ants without id: http://localhost/ant.html ^^ 2010-12-18T10:21:46 doesn't load 2010-12-18T10:21:53 sorry... 2010-12-18T10:22:11 http://marco-leise.homedns.org/ant.html 2010-12-18T10:22:42 they walked onto a wall :| 2010-12-18T10:23:08 i know... ignore the randomness... i wanted to point to the shadow ants appearing where an ant moves 2010-12-18T10:24:07 can i just send you one kind of spawn ant? 2010-12-18T10:24:43 ah crap, i need to track id's to kill as well 2010-12-18T10:24:53 *** Azrathud has quit IRC (Ping timeout: 240 seconds) 2010-12-18T10:25:14 you can, but if you find that you know what kind it is at the particular code line i can visualize them differently. that said it works with only "S" of course 2010-12-18T10:25:48 hmmm, i'll see what i end up with 2010-12-18T10:25:58 it just got even messier :P 2010-12-18T10:26:46 I try to swallow the guilt feelings :D 2010-12-18T10:27:28 You haven't worked with databases a lot have you? 2010-12-18T10:28:09 i have to add ants to a square, then if ants added == 1 and home base has more than 1 grains and no ant is current there after moves, spawn a new and with a new id otherwise if there are grains spawn an ant without a new id, then i need to do collision detections on all that, and send you ids for any ants that die and spawning of any new ones whether through collisions or spawning a new one 2010-12-18T10:28:17 no, i haven't :P 2010-12-18T10:28:28 more than 0 grains* 2010-12-18T10:28:39 i'm not sure this is going to help with my efficiency problem either :P 2010-12-18T10:29:05 *** holmar has quit IRC (Remote host closed the connection) 2010-12-18T10:30:50 that sounds complicated... i hope it is a lot simpler in reality. if it helps just spawn the ant, let it die in the collision and spawn a new ant if some hp remain for one player 2010-12-18T10:32:37 so you can end each processing state with a command (spawn, kill+spawn) instead of keeping all the information till the end of the pipeline, just to save an extra id 2010-12-18T10:34:08 the point is i'm currenty 2010-12-18T10:34:19 oops that was meant to be backspace not enter :) 2010-12-18T10:36:43 *** Anderson_ has joined #aichallenge 2010-12-18T10:36:49 heheh 2010-12-18T10:36:57 .. 2010-12-18T10:36:57 *** Anderson_ has quit IRC (Client Quit) 2010-12-18T10:37:04 *** Anderson_ has joined #aichallenge 2010-12-18T10:37:14 *** abm has joined #aichallenge 2010-12-18T10:37:15 hehee 2010-12-18T10:37:48 *** abm is now known as Anderson_BM 2010-12-18T10:39:27 Frontier: U and S, do I have to? :P 2010-12-18T10:41:36 hehe 2010-12-18T10:41:44 actually, i have my own counter proposal :P 2010-12-18T10:41:46 give me a min 2010-12-18T10:43:05 *** amstan has joined #aichallenge 2010-12-18T10:43:05 *** ChanServ sets mode: +o amstan 2010-12-18T10:46:01 Frontier: do you want to be told deposit amounts? 2010-12-18T10:46:31