2011-01-11T00:02:53 *** medrimonia1 has joined #aichallenge 2011-01-11T00:37:34 *** Mathnerd314 has quit IRC (Ping timeout: 264 seconds) 2011-01-11T00:41:12 *** closedbracket has quit IRC (Remote host closed the connection) 2011-01-11T01:06:04 *** bhasker has joined #aichallenge 2011-01-11T01:41:28 *** sigh has quit IRC (Ping timeout: 240 seconds) 2011-01-11T01:49:53 *** bhasker has quit IRC (Quit: bhasker) 2011-01-11T01:56:34 *** acieroid` is now known as acieroid 2011-01-11T02:37:01 *** sigh has joined #aichallenge 2011-01-11T03:28:38 *** perror has joined #aichallenge 2011-01-11T03:36:17 *** aerique has joined #aichallenge 2011-01-11T04:09:59 *** Naktibalda has joined #aichallenge 2011-01-11T04:27:05 *** AlliedEnvy has joined #aichallenge 2011-01-11T04:53:30 *** Palmik has joined #aichallenge 2011-01-11T04:56:26 *** mega1 has joined #aichallenge 2011-01-11T04:57:45 I've just spammed the forum about webinar that I'm going to do: http://ai-contest.com/forum/viewtopic.php?f=19&t=1241 2011-01-11T04:58:22 *** Utkarsh has quit IRC (Ping timeout: 265 seconds) 2011-01-11T04:58:45 I'll try to balance it between programming/AI/planetwars/Lisp 2011-01-11T04:59:37 the boss, who will also speak, is going to concentrate on "what it means for the semantice web" :-) 2011-01-11T05:00:04 anyway, if you are interested you can listen in, ask questions. 2011-01-11T05:04:36 *** Utkarsh has joined #aichallenge 2011-01-11T05:11:00 *** boegel has joined #aichallenge 2011-01-11T05:17:19 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2011-01-11T05:26:11 *** Utkarsh has joined #aichallenge 2011-01-11T05:33:09 mega1: it needs windows or mac 2011-01-11T05:33:17 I know :-( 2011-01-11T05:33:22 aww 2011-01-11T05:33:40 I'm going to have to boot windows 2011-01-11T05:33:48 I hope it works in virtualbox ... 2011-01-11T05:34:41 how is it any different from skype 2011-01-11T05:38:23 I've used skype but never used gotomeeting :-) 2011-01-11T05:38:43 it's a screencast I think 2011-01-11T05:38:50 with audio conferencing 2011-01-11T06:26:39 *** Utkarsh has quit IRC (Read error: Connection reset by peer) 2011-01-11T06:27:16 *** Utkarsh has joined #aichallenge 2011-01-11T07:12:34 *** delt0r_ has joined #aichallenge 2011-01-11T07:13:47 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2011-01-11T07:35:24 *** sigh has quit IRC (Remote host closed the connection) 2011-01-11T07:38:48 *** mega1 has quit IRC (Ping timeout: 246 seconds) 2011-01-11T08:32:10 *** amstan_ has quit IRC (Ping timeout: 240 seconds) 2011-01-11T08:48:54 *** closedbracket has joined #aichallenge 2011-01-11T08:49:38 *** closedbracket has quit IRC (Remote host closed the connection) 2011-01-11T08:49:53 *** closedbracket has joined #aichallenge 2011-01-11T09:04:16 well.. this seems up now: https://github.com/amstan/aichallenge/graphs/impact 2011-01-11T09:04:23 janzert and jeff are in the lead 2011-01-11T09:06:34 *** UndiFineD has quit IRC (Quit: Ex-Chat) 2011-01-11T09:07:08 *** UndiFineD has joined #aichallenge 2011-01-11T09:16:10 *** tapwater has joined #aichallenge 2011-01-11T09:21:26 *** UndiFineD has quit IRC (Quit: Ex-Chat) 2011-01-11T09:24:42 *** UndiFineD has joined #aichallenge 2011-01-11T09:38:02 *** Migi32 has joined #aichallenge 2011-01-11T11:03:51 *** aerique has quit IRC (Quit: ...) 2011-01-11T11:19:13 *** JamesMG has joined #aichallenge 2011-01-11T11:37:39 *** boegel has quit IRC (Ping timeout: 264 seconds) 2011-01-11T11:57:07 *** Revelations has joined #aichallenge 2011-01-11T11:57:51 *** JamesMG has quit IRC (Remote host closed the connection) 2011-01-11T11:57:56 *** Revelations is now known as JamesMG 2011-01-11T12:10:40 *** steve____ has joined #aichallenge 2011-01-11T12:13:05 *** steve____ has quit IRC (Client Quit) 2011-01-11T12:13:42 *** Palmik has quit IRC (Read error: Connection reset by peer) 2011-01-11T12:14:05 *** Palmik has joined #aichallenge 2011-01-11T12:15:01 *** JamesMG has quit IRC (Quit: leaving) 2011-01-11T12:15:56 *** JamesMG has joined #aichallenge 2011-01-11T12:16:07 *** mega1 has joined #aichallenge 2011-01-11T12:32:15 *** Utkarsh has quit IRC (Ping timeout: 240 seconds) 2011-01-11T12:38:58 *** Utkarsh has joined #aichallenge 2011-01-11T12:50:22 *** Utkarsh has quit IRC () 2011-01-11T12:51:22 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.8/20100722155716]) 2011-01-11T13:02:07 *** Kingpin13 has joined #aichallenge 2011-01-11T13:23:09 *** Kingpin13 is now known as Kingpin13|food 2011-01-11T13:48:45 *** perror has quit IRC (Quit: Bye all !) 2011-01-11T14:03:39 *** Naktibalda has joined #aichallenge 2011-01-11T14:35:38 *** boegel has joined #aichallenge 2011-01-11T14:55:02 *** boegel has quit IRC (Ping timeout: 240 seconds) 2011-01-11T15:13:07 *** delt0r___ has joined #aichallenge 2011-01-11T15:14:14 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-01-11T15:23:22 *** Palmik_ has joined #aichallenge 2011-01-11T15:24:04 *** Palmik has quit IRC (Read error: Connection reset by peer) 2011-01-11T15:40:32 *** sigh has joined #aichallenge 2011-01-11T15:46:23 *** BtbN has quit IRC (*.net *.split) 2011-01-11T15:46:23 *** davidd has quit IRC (*.net *.split) 2011-01-11T15:46:23 *** jbroman has quit IRC (*.net *.split) 2011-01-11T15:49:07 *** BtbN has joined #aichallenge 2011-01-11T15:49:08 *** jbroman has joined #aichallenge 2011-01-11T15:49:08 *** davidd has joined #aichallenge 2011-01-11T15:55:37 *** Palmik_ has quit IRC (Remote host closed the connection) 2011-01-11T16:03:02 *** sigh has quit IRC (Remote host closed the connection) 2011-01-11T16:09:38 *** boegel has joined #aichallenge 2011-01-11T16:13:34 *** bhasker has joined #aichallenge 2011-01-11T16:18:20 *** bhasker has quit IRC (Ping timeout: 260 seconds) 2011-01-11T16:24:34 *** bhasker has joined #aichallenge 2011-01-11T16:29:33 *** Cold-Phoenix has joined #aichallenge 2011-01-11T16:29:40 *** mega1 has quit IRC (Ping timeout: 276 seconds) 2011-01-11T16:32:28 *** fgump has joined #aichallenge 2011-01-11T16:36:26 oh man.. 2011-01-11T16:36:31 people have to come here 2011-01-11T16:36:39 we're starting deving at 5:45 2011-01-11T16:37:20 "here"? 2011-01-11T16:37:27 i am still working at 17:45 est 2011-01-11T16:42:07 *** Mathnerd314 has joined #aichallenge 2011-01-11T16:43:10 where is here? 2011-01-11T16:45:47 i was just thinking... what if we had a pay-what-you-want (allowing $0) entry fee? money could go toward costs and/or a prize pool. no promised benefits for paying at all, just essentially an anonymous donation 2011-01-11T16:46:24 kind of inspired by things like the humble indie bundle and stuff 2011-01-11T16:47:41 problem: people don't like choices. 2011-01-11T16:47:50 prizes are probably subject to regulatory laws, too 2011-01-11T16:48:25 only prizes with monetary value or export restrictions 2011-01-11T16:51:30 well, there goes that idea then 2011-01-11T16:52:57 i would recommend against offering prizes 2011-01-11T16:53:07 yeah it might take the fun out of it anyway 2011-01-11T16:55:19 *** boegel has quit IRC (Quit: *poof!*) 2011-01-11T16:55:31 *** Kingpin13|food has quit IRC (Ping timeout: 276 seconds) 2011-01-11T16:56:52 *** mceier has joined #aichallenge 2011-01-11T17:20:22 *** sigh has joined #aichallenge 2011-01-11T17:26:55 *** Naktibalda has quit IRC (Remote host closed the connection) 2011-01-11T17:56:15 *** amstan_ has joined #aichallenge 2011-01-11T17:56:15 *** ChanServ sets mode: +o amstan_ 2011-01-11T18:04:45 *** sigh has quit IRC (Read error: Connection reset by peer) 2011-01-11T18:05:10 *** sigh has joined #aichallenge 2011-01-11T18:08:22 *** steve____ has joined #aichallenge 2011-01-11T18:09:16 *** steve____ has quit IRC (Client Quit) 2011-01-11T18:09:36 k... all ready? 2011-01-11T18:12:12 ready for what? 2011-01-11T18:12:15 coding.. 2011-01-11T18:12:17 what else? lol 2011-01-11T18:12:36 jeff's hooked up to his comp, finishing his ant sim 2011-01-11T18:13:12 i can't at the moment 2011-01-11T18:13:39 but figure out which branch we should be using and i'll add when i can 2011-01-11T18:14:49 master should be fine 2011-01-11T18:15:05 though.. 2011-01-11T18:15:07 hmm 2011-01-11T18:15:16 hold on, this won't work with jeff's plans 2011-01-11T18:15:22 we should probably continue the pw branch 2011-01-11T18:15:31 and merge pw-worker into it 2011-01-11T18:15:42 janzert: you there? 2011-01-11T18:15:56 pw is a funny name now that it isn't planet_wars anymore 2011-01-11T18:16:24 and i think a clean branch would be better if we're coding from scratch 2011-01-11T18:16:33 (pw has a lot of stuff we don't need, too) 2011-01-11T18:17:36 hmm? 2011-01-11T18:23:57 *** Migi32 has quit IRC (Quit: BOINC) 2011-01-11T18:25:54 yeah.. 2011-01-11T18:26:01 we're not coding from scratch yet though 2011-01-11T18:26:11 we're going by those plans in the topic 2011-01-11T18:26:22 so.. we should maintain that code.. 2011-01-11T18:26:25 * Zannick should read /topic more often 2011-01-11T18:26:38 aichallenge: Alexandru Stan master * r018ee28 / django/contest/models.py : added Map model - http://bit.ly/fie2wR 2011-01-11T18:27:23 janzert: so.. do i just merge the 2 branches 2011-01-11T18:27:30 is there anything i have to worry about? 2011-01-11T18:28:06 a ton of conflicts 2011-01-11T18:28:27 yeah, I wouldn't merge necessarily 2011-01-11T18:28:55 but? 2011-01-11T18:29:03 we should have some clean code though.. 2011-01-11T18:29:16 you said the compile pipeline from the trunk was not active 2011-01-11T18:29:22 I don't remember offhand any files that are used by both at this point but it'll probably be easier to just take the worker backend directory and put it in a worker directory 2011-01-11T18:29:34 then strip out the stuff that isn't actually used on a worker 2011-01-11T18:29:53 also strip the stuff out of the current backend that is no longer used on the master server 2011-01-11T18:30:14 it's a better/more natural separation anyway 2011-01-11T18:30:43 hm, so, reading the roadmap 2011-01-11T18:30:58 it sounds like we want two branches 2011-01-11T18:31:19 one to reuse the old code to get off the ground for this contest 2011-01-11T18:31:42 and another to develop new code for the one after 2011-01-11T18:31:48 Zannick: i have 3 branches now.. 2011-01-11T18:31:58 pw* is from the old contest, master is new 2011-01-11T18:32:12 pw is the svn trunk, pw-worker is the worker branch from svn 2011-01-11T18:32:26 i would like to merge the pws or clean them up a little 2011-01-11T18:32:30 oh, okay, so master is the new clean one? and pw is the old one? 2011-01-11T18:32:32 actually, compile_anything.py is used by both the master and workers 2011-01-11T18:33:02 and should be identical in both branches already 2011-01-11T18:34:02 Zannick: yes, except we won't be using the master one yet 2011-01-11T18:34:09 Zannick: except for the new backend rewrite 2011-01-11T18:34:16 which is not essential at this point 2011-01-11T18:34:29 amstan_: okay. 2011-01-11T18:35:12 i will try to split my time between them, though 2011-01-11T18:35:21 that's what all of us will do 2011-01-11T18:35:26 :) 2011-01-11T18:35:48 *** amstan is now known as amstanirrsi 2011-01-11T18:35:51 *** amstan_ is now known as amstan 2011-01-11T18:36:14 *** Cold-Phoenix has quit IRC () 2011-01-11T18:36:42 Jeff: "so.. i won't do anything, won't go on irc won't do anything until this ants thing is done" 2011-01-11T18:36:48 next thing... he's watching a video 2011-01-11T18:36:56 "it's rendering now..." 2011-01-11T18:40:03 i'm gonna do some coding later tonight, not now 2011-01-11T18:40:53 oh man.. they're cool, those ants 2011-01-11T18:41:39 i'm kinda busy tonight 2011-01-11T18:42:14 i was going to work first on rewriting compile_anything, but that isn't quite as necessary as improving the things that need improvement 2011-01-11T18:43:22 animated gif coming soon! 2011-01-11T18:48:11 janzert: so.. what is used on the worker? 2011-01-11T18:57:26 aichallenge: Alexandru Stan pw * ra6d5fce / (64 files in 4 dirs): renamed worker backend - http://bit.ly/fGB8Ur 2011-01-11T18:57:27 aichallenge: Alexandru Stan pw * ra825b82 / (32 files in 2 dirs): copied worker folder from the other branch, hence merged them - http://bit.ly/gQhLcn 2011-01-11T18:57:34 aichallenge: Alexandru Stan pw-worker * ra6d5fce / (64 files in 4 dirs): renamed worker backend - http://bit.ly/fGB8Ur 2011-01-11T19:09:39 oh man.. it's so exciting 2011-01-11T19:11:07 it should be possible to tell cia that Alexandru Stan => amstan and such 2011-01-11T19:11:42 probably not 2011-01-11T19:11:52 it's how it appears in the commit 2011-01-11T19:12:19 one of my other channels uses cia in that manner 2011-01-11T19:12:30 i'm not sure if it's cia or the repo host 2011-01-11T19:12:31 yeah, but they probably have a post-commit hook 2011-01-11T19:12:39 and you can configure that part 2011-01-11T19:12:40 could be 2011-01-11T19:12:56 but.. the github stuff is not flexible in terms of cia settings 2011-01-11T19:13:16 i dunno man 2011-01-11T19:13:32 there's 2 buttons: enable and test hook 2011-01-11T19:13:34 it's cool 2011-01-11T19:13:47 it submits it to the cia project with the same name 2011-01-11T19:13:51 if it's already taken 2011-01-11T19:13:53 too bad, lol 2011-01-11T19:14:42 amstan: the stuff in the worker branch backend directory, easiest spot to start to get an idea of how it works is probably by looking first at game_server_build.sh 2011-01-11T19:14:56 that loads everything onto the worker and sets it up 2011-01-11T19:15:13 then start_worker.sh is used to start everything 2011-01-11T19:16:40 that will lead you to tournament_manager.py after a couple of steps and there is where the actual work is coordinated 2011-01-11T19:18:22 and the basic idea is that it keeps asking the master for a game and gets back a pair of bots and a map, then it downloads/compiles the bots if necessary, runs the game and reports the outcome back to the master 2011-01-11T19:25:22 *** mceier has quit IRC (Quit: leaving) 2011-01-11T19:27:32 there we go: http://www.ai-contest.com/forums/viewtopic.php?f=21&t=1243 2011-01-11T19:30:13 janzert, jmcarthur, Zannick: ants! 2011-01-11T19:30:19 i saw 2011-01-11T19:41:54 *** Mathnerd314 has quit IRC (Remote host closed the connection) 2011-01-11T19:56:37 *** closedbracket has quit IRC (Remote host closed the connection) 2011-01-11T19:58:47 aichallenge: Alexandru Stan simpleants * r3a3a863 / (.gitignore four_map.png Makefile playback/foo.avi): added makefile for the playbacks, changed map so there's no deadends, removed random avi file (+5 more commits...) - http://bit.ly/eAQtuO 2011-01-11T19:59:01 and there's the repo.. 2011-01-11T19:59:12 branch i mean 2011-01-11T20:08:18 *** closedbracket has joined #aichallenge 2011-01-11T20:20:20 aichallenge: Jeff Cameron simpleants * rd9731fe / map.py : g - http://bit.ly/dHED0Q 2011-01-11T20:40:47 oh man.. he fed the ants.. 2011-01-11T20:40:50 now there's a problem.. 2011-01-11T20:41:13 Zannick: you there? 2011-01-11T20:41:23 janzert, jmcarthur: ready? 2011-01-11T20:42:34 I'm ready to watch it all happen ;) 2011-01-11T20:46:38 yeah.. 2011-01-11T20:46:45 the forum didn't like the huge ass gif 2011-01-11T20:46:47 so.. 2011-01-11T20:46:49 getting csclub 2011-01-11T20:46:52 k.. 2011-01-11T20:46:53 drumrolls! 2011-01-11T20:47:18 DO NOT FEED THE ANTS: http://csclub.uwaterloo.ca/~amstan/aichallenge/simpleantswithfood.gif 2011-01-11T20:49:00 aichallenge: Alexandru Stan simpleants * r61b46da / map.py : Merge branch 'simpleants' of github.com:amstan/aichallenge into simpleants - http://bit.ly/eUERIQ 2011-01-11T20:49:01 aichallenge: Alexandru Stan simpleants * reae2567 / four_map.png : jeff fed the ants in the previous commit, I gave them more space now - http://bit.ly/hYRMk4 2011-01-11T20:59:16 i'm trying to decide if i think this simple ants game has any depth or not 2011-01-11T20:59:32 * janzert concurs 2011-01-11T21:01:24 amstan: i'm about to be offline for a while, sorry 2011-01-11T21:02:15 so you send your ants out to turn food into more ants. is there any way that you can do that more efficiently than everybody else? 2011-01-11T21:02:19 *** fgump has quit IRC (Quit: Leaving) 2011-01-11T21:02:26 maybe a little traveling saleman in there 2011-01-11T21:02:36 *salesman 2011-01-11T21:02:41 is there limited sight? 2011-01-11T21:02:46 it doesn't look like it 2011-01-11T21:03:50 the only reason i think fighting has any meaning at all is that you don't want to fight in the early game, since it's a free for all 2011-01-11T21:04:01 *** amstan has quit IRC (Ping timeout: 250 seconds) 2011-01-11T21:04:04 you want to delay that as long as possible while still collecting as much food as possible 2011-01-11T21:04:25 so... i guess i can see a little challenge in there 2011-01-11T21:04:34 i still can't seem to get excited about it though 2011-01-11T21:12:01 I guess my concern is there doesn't seem to be a lot of room for different strategies, just optimization of food gathering and if you have the global state while that optimization is probably still difficult to do it doesn't seem as interesting to me 2011-01-11T21:12:58 same here 2011-01-11T21:15:52 but I suppose without getting some people writing bots it's probably impossible to tell if that's really true 2011-01-11T21:16:26 maybe write a "Greedy" bot then see how hard it is to write something else that beats it 2011-01-11T21:17:36 where I define greedy as something like a bot that sends the closest free ant it has toward a food particle and never sends more than one to any particular food 2011-01-11T21:34:08 so, jeff an alex _really_ do want multi-player ratings :/ 2011-01-11T21:34:15 * dmj111 goes back to reading the trueskill papers. 2011-01-11T21:34:34 well i don't 2011-01-11T21:34:41 :) 2011-01-11T21:34:42 :\ 2011-01-11T21:36:01 i'm trying to decide if i want to bother with some linux-only code in my engine. it would mean we would finally have a way to time bots based on cpu usage instead of wall clock time 2011-01-11T21:36:16 but that functionality would be linux only 2011-01-11T21:38:26 as long as it still works on clock time on windows, if it gets distributed to everyone. i am assuming the percentage of windows users was pretty high last time. 2011-01-11T21:38:29 if you do, remember that you can't use cpu time alone otherwise bot can sleep forever and hang the game from completing 2011-01-11T21:38:57 also I would contend that IO wait time should be charged to bots as well 2011-01-11T21:39:41 i.e. if you're doing a bunch of disk operations it's your fault that it's slow :P 2011-01-11T21:43:36 yeah it wouldn't be CPU time alone 2011-01-11T21:43:56 could also include time in the kernel on behalf of the bot 2011-01-11T21:44:25 if you're doing a bunch of disk IO you're doing something evil :P 2011-01-11T21:44:53 I'm making use of pre-calculated stuff if reading it off disk is free :) 2011-01-11T21:45:53 i'll just put this on the "maybe do later" list that i'm keeping in my head 2011-01-11T21:46:55 probably a good idea given the criticality of it :) 2011-01-11T21:47:11 it would be a lot of work and, well, as much as it sucks i've gotten kind of used to wall clock time 2011-01-11T21:47:33 *** medrimonia1 has quit IRC (Ping timeout: 240 seconds) 2011-01-11T21:48:01 well, it wouldn't be all that much work to *poll* for the bot's cpu time, but blocking on it for a specific amount of cpu time would be more work 2011-01-11T21:55:53 *** amstan has joined #aichallenge 2011-01-11T21:55:53 *** ChanServ sets mode: +o amstan 2011-01-11T21:57:30 janzert, jmcarthur, dmj111: there will be limited sight 2011-01-11T22:01:36 *** ssalbiz has joined #aichallenge 2011-01-11T22:01:52 ssalbiz: so.. why not? jeff tought it was a good idea 2011-01-11T22:02:28 amstan: slowness is the major concern 2011-01-11T22:02:49 jeff was claiming pretty low overhead, what do you know? 2011-01-11T22:03:07 one obstacle that I never really bothered to fix was knowing when the vm was spun up properly (otherwise ssh hangs on response) 2011-01-11T22:03:33 *** medrimonia has joined #aichallenge 2011-01-11T22:03:35 so, from a terminal, doing a single testcase, there is relatively little _perceivable_ overhead 2011-01-11T22:04:18 so that's the big overhead then? spinning up the vm? 2011-01-11T22:04:29 if you try to open an ssh connection to a vm without a sleep of ~5s in between it turns out that the sshd inside the vm isn't spun up properly yet and bots 2011-01-11T22:04:46 after that it's not too bad 2011-01-11T22:05:06 problem is getting the vm to signal when sshd is ready to receive incoming connectinos 2011-01-11T22:05:10 connections, even 2011-01-11T22:05:32 not too hard to solve, but you still end up with ~10s overhead to spin up two contest vms 2011-01-11T22:06:20 yeah, that's a lot, we were doing 10 games per minute/machine last contest 2011-01-11T22:06:22 i think 2011-01-11T22:06:29 janzert could confirm 2011-01-11T22:07:07 ssalbiz: what about the stuff we're using now? are you aware of any weakness? 2011-01-11T22:08:26 amstan: haven't looked at it in a while, let me refresh my memory 2011-01-11T22:08:29 is it the systrace stuff? 2011-01-11T22:09:14 hmm 2011-01-11T22:09:30 i'm not exactly sure either, let me check the repo 2011-01-11T22:09:47 this is it, if you wanna peek: https://github.com/amstan/aichallenge/tree/pw/planet_wars/backend 2011-01-11T22:10:14 ssalbiz: it's just the jail 2011-01-11T22:11:17 i was just thinking about just using ptrace, but it would be kind of slow 2011-01-11T22:12:13 on the other hand, it would make timing based on cpu very easy, so we wouldn't have to worry about unfairness regarding the wall clock timer 2011-01-11T22:12:16 jmcarthur: I haven't used ptrace before, is it difficult to configure custom permissions for what invoked shits can/cannot do? 2011-01-11T22:12:35 the jvm, for example, needs to open up a bunch of sockets for some reason 2011-01-11T22:12:51 ssalbiz: essentially it allows the parent process to decide what to do with all syscalls 2011-01-11T22:13:03 really? that makes things difficult then... 2011-01-11T22:13:20 because we certainly don't want to just allow arbitrary network access... although a firewall would be fine for that 2011-01-11T22:13:24 yeah, one of the many reasons using systrace was sub-optimal 2011-01-11T22:13:42 with ptrace you should be able to even decide what to do based on the parameters to the syscall though 2011-01-11T22:15:04 anyway, my idea for the timing was that with ptrace you could register a SIGALRM on behalf of the child process that sends the child a signal once it has exhausted its cpu quota for the current turn, but the parent process would be able to get the signal first (and the child would be paused automatically) 2011-01-11T22:15:35 downside is that the child wouldn't be able to use SIGALRM 2011-01-11T22:16:15 eh, that seems like a fair tradeoff 2011-01-11T22:16:46 i also just don't know if i feel like messing with all this right now, although it would make sense since i'm writing an engine anyway 2011-01-11T22:17:12 I can look into it over the weekend 2011-01-11T22:17:15 my main problem with it is that ptrace is kind of slow 2011-01-11T22:17:20 s/kind of/very/ 2011-01-11T22:17:54 jmcarthur: yeah, running some examples to test concept considered useful 2011-01-11T22:19:12 http://www.xs4all.nl/~weegen/eelis/geordi/ <-- this uses ptrace, for reference 2011-01-11T22:19:22 and in turn, codepad.org uses a modified version of geordi 2011-01-11T22:21:02 based on these comments, it looks like kind of a sucky process https://github.com/Eelis/geordi/raw/master/src/EvalCxx.hsc 2011-01-11T22:21:47 but that code is public domain and i'm writing my engine in Haskell anyway, so maybe i could just rip it out and use it 2011-01-11T22:41:39 *** abc_ has joined #aichallenge 2011-01-11T23:13:16 *** delt0r_ has joined #aichallenge 2011-01-11T23:14:18 *** delt0r___ has quit IRC (Ping timeout: 240 seconds) 2011-01-11T23:27:02 *** Mathnerd314 has joined #aichallenge