2011-07-24T00:01:22 bhasker: which browser and version are you using btw? 2011-07-24T00:06:33 *** mathis_ has quit IRC (Quit: Page closed) 2011-07-24T00:47:55 chrome 2011-07-24T00:47:56 on the mac 2011-07-24T00:48:13 frontier how are you calculating fog of war? 2011-07-24T00:55:28 *** jbrechtel_ has quit IRC (Ping timeout: 252 seconds) 2011-07-24T00:59:36 mmm 2011-07-24T01:01:09 bhasker: like this: http://pastebin.com/qK0FG7xr 2011-07-24T01:04:05 i suck at reading other peoples code :\ 2011-07-24T01:04:27 http://pastebin.com/gqCA8u5N that's how i do vision, but i think sigh had a faster way 2011-07-24T01:04:38 antimatroid: oh you will soon work for someone and read their code :) 2011-07-24T01:05:03 might not, contemplating phd next year :P 2011-07-24T01:05:26 ignore all the other people :) 2011-07-24T01:05:58 to be fair i have a call to interpolate in it and render the fog into a huge 32-bit bitmap (120x120 x 20x20 = 5,760,000 Pixels) 2011-07-24T01:06:59 interpolate calculates the properties of an ant at a given time (color, size, position, ...) 2011-07-24T01:07:07 i'm still not over the fact that i have to index matrices of bools as visited[loc.row][loc.col] instead of visited[loc], stupid c++ :( 2011-07-24T01:07:33 antimatroid: again, use byte as the datatype and you're done 2011-07-24T01:07:51 but doesn't that slow it/take more space whatever? :P 2011-07-24T01:07:53 add a typedef if you want a more fitting name, like bytebool 2011-07-24T01:08:23 antimatroid: it takes 8x the space, but it should be faster (i think) 2011-07-24T01:08:30 hmmm 2011-07-24T01:08:37 space is generally not really that much of a concern 2011-07-24T01:09:06 it is a simple check for 0, whereas the bit-array includes some bitshifting 2011-07-24T01:09:46 is it better to check is/n't 0 rather than is/n't 1? 2011-07-24T01:09:49 at least i don't know how you could check for a set bit without bit shifting 2011-07-24T01:11:01 let me have a look 2011-07-24T01:11:27 *** sigh has joined #aichallenge 2011-07-24T01:11:28 hmm. iirc after an arithmetic operation the result bits are updated (overflow, zero, parity, ...) so the compiler could use those. generally testing for zero or one... I don't know. That might not make a difference 2011-07-24T01:12:08 It has bin some time since I last wrote in assembly 2011-07-24T01:12:16 *been 2011-07-24T01:12:47 k that code will take some time to make sense of 2011-07-24T01:13:24 antimatroid: I always check for == or != 0, just to be sure ;). It can't hurt anyway. 2011-07-24T01:13:42 can you explain is simple words how you are building the fog of war 2011-07-24T01:14:25 one way would be to just do a bfs around each ant and paint each visible square 2011-07-24T01:14:49 that's what i do 2011-07-24T01:15:01 yea i looked at your code and modified it for other stuff that i do 2011-07-24T01:15:04 bhasker: I first set all squares to fog, then I ... seem to prepare some sort of lookup table 2011-07-24T01:15:15 another is to make a set of translate amounts that gives the vision squares and paint that around each ant 2011-07-24T01:15:22 another way would be to do some precalculation upfront 2011-07-24T01:15:31 just for each location build a set of visible locations up front 2011-07-24T01:15:41 then just sort + uniq the total 2011-07-24T01:16:02 then each iteration just extract a list of locations for each ant + sort + uniq it 2011-07-24T01:16:11 but doubt that will be faster than doing a bfs around each ant 2011-07-24T01:16:34 i think my second one would be faster than bfs 2011-07-24T01:16:37 no wait, i first set visible all squares on a + around the ant 2011-07-24T01:16:40 it's essentially the same thing but without the bfs 2011-07-24T01:17:17 look at antimatroids C++ starter bot 2011-07-24T01:17:26 it has a function to paint the vision of a bot 2011-07-24T01:17:35 that should work for building a fog of war 2011-07-24T01:17:45 and then there is a lookup for the wrapping, when I reach the top, bottom, left or right side of the map 2011-07-24T01:18:38 again look at antimatroid's c++ bot it does it without any lookups 2011-07-24T01:18:46 well not without any lookups 2011-07-24T01:18:50 there is a small lookup 2011-07-24T01:19:13 what's a lookup? 2011-07-24T01:19:24 a translation table 2011-07-24T01:19:27 https://github.com/aichallenge/aichallenge/blob/epsilon/ants/dist/starter_bots/cpp/State.cc that has the starter bot updatevision function 2011-07-24T01:20:29 "THE OBVIOUS WAY OF TRYING TO IMPROVE IT BREAKS USING THE EUCLIDEAN METRIC" Oo 2011-07-24T01:20:56 i still think it's worth setting it up as a graph and removing neighbour information with water, the visualiser even knows all the water information at the start, so it doesn't need to worry about having to remove edges from the graph 2011-07-24T01:21:09 mleise: yeah, i kinda wish we hadn't used euclidean for vision, but we did 2011-07-24T01:21:31 it's because of: 2011-07-24T01:21:31 a.* 2011-07-24T01:21:31 .a. 2011-07-24T01:21:47 both ants have the same "walk/search" distance to * but a different euclidean distance 2011-07-24T01:22:09 so you suggest we should have used manhattan? 2011-07-24T01:22:09 and i think there was another problem too, cause my trivial idea to fix that didn't work 2011-07-24T01:22:10 a graph in the visualizer? what are you talking about? 2011-07-24T01:22:17 manhatten or chebyshev 2011-07-24T01:22:25 i have no idea what chebyshev is 2011-07-24T01:22:29 squares? 2011-07-24T01:22:41 melise: i store my grid in my actual bot with neighbour locations at each square, so i don't actually need to call get location after the first turn 2011-07-24T01:23:17 so you are practically 'position independent'? 2011-07-24T01:23:19 cdist(x,y) := min (|x.row-y.row|, |x.col-y.col|) i think 2011-07-24T01:23:30 position independent? 2011-07-24T01:23:39 hmm you still use coordinates -.- 2011-07-24T01:23:48 I though you just have a graph :D 2011-07-24T01:23:54 change that to max 2011-07-24T01:24:06 i still have a matrix with my squares 2011-07-24T01:24:24 and index the locations with that, but each square has a vector of neighbour locations that aren't water which i use for pathfinding etc. 2011-07-24T01:24:54 although i've given up on having time to write a proper bot 2011-07-24T01:26:52 basically, euclidean "circle" is a circle, manhatten "circle" is a diamond, and chebyshev "circle" is a square 2011-07-24T01:27:10 wow this c++ code looks like it will be *slow* in JavaScript ;) 2011-07-24T01:27:27 what from? 2011-07-24T01:28:09 there are so many hash objects generated 2011-07-24T01:29:00 mleise: what i think would make it faster is the have a get_translate_loc function, then at the start, generate the translate numbers to get every square within vision of a particular location, then just call that for each ant location and mark it as visible 2011-07-24T01:29:09 to* have 2011-07-24T01:29:36 I use that in my bot but what would the maximum memory use be? 2011-07-24T01:29:55 what do you mean? 2011-07-24T01:30:08 how much RAM will this consume at max. 2011-07-24T01:30:08 i wouldn't store vision information for every location, just the translate amounts to generate it for a location 2011-07-24T01:30:14 ah ok 2011-07-24T01:30:54 beyond that, sigh might have a faster way 2011-07-24T01:31:25 how do you know about sigh's method? 2011-07-24T01:31:55 my method for what? 2011-07-24T01:32:03 because i was around when he found the error with searching from all ants simultaneously with euclidean and know he did waht the engine is using 2011-07-24T01:32:08 sigh: calculating vision information 2011-07-24T01:32:23 ah, i see 2011-07-24T01:32:28 what language? 2011-07-24T01:32:33 JavaScript :D 2011-07-24T01:33:24 * mleise should really not render full 2400x2400 pixel rgba bitmaps when zoomed in 2011-07-24T01:33:48 you might want to look at how the engine does it, it might be appropriate 2011-07-24T01:33:57 for the visualiser, where you have full state info 2011-07-24T01:36:24 sigh: Deltas... I have a deja vu 2011-07-24T01:37:13 I thought, if I talk about my vision code, people will propose deltas and then I'll have to say that my code is ... too flexible for deltas :p 2011-07-24T01:37:55 *** bhasker has quit IRC (Quit: bhasker) 2011-07-24T01:38:02 I don't imply that ants move exaclty one square per turn in one of 4 directions 2011-07-24T01:41:07 i'm interested in people writing code for luring enemy ants to places that you want 2011-07-24T01:41:41 either leading them with a lone ant away from your "hive/base/whatever" or luring them in to one ant in the middle of a trap where your ants come in from outside the enemies vision on all sides 2011-07-24T01:43:29 btw, I improved the attack lines a bit. They are now based on the final positions of the ants in the current turn and are visible during the whole turn. Often when one turn ends the lines are immediatly replaced by new ones which makes the ants look really busy. 2011-07-24T01:45:24 :) 2011-07-24T01:45:59 sigh: we won the tour de france :) 2011-07-24T01:56:12 antimatroid: cool, what do we get? 2011-07-24T01:56:25 our anthem played, i guess? :P 2011-07-24T01:56:42 :P 2011-07-24T01:59:42 a doping scandal? 2011-07-24T01:59:56 sorry 2011-07-24T03:43:50 *** mcstar has joined #aichallenge 2011-07-24T03:46:28 i dreamt with gene hackman 2011-07-24T04:06:28 *** berak has joined #aichallenge 2011-07-24T04:20:54 *** amstan has quit IRC (Ping timeout: 255 seconds) 2011-07-24T04:39:22 *** Palmik has joined #aichallenge 2011-07-24T05:18:52 *** choas has joined #aichallenge 2011-07-24T05:28:46 *** vdp|afk is now known as vdp 2011-07-24T05:36:30 yeah fair enough 2011-07-24T05:36:37 erm fwrong window :P 2011-07-24T05:58:52 *** ArchMonkey has joined #aichallenge 2011-07-24T06:01:23 *** humanfromearth has joined #aichallenge 2011-07-24T06:02:09 hi, I read the rules of the contest and it says: Use of multiple processes or threads is prohibited. 2011-07-24T06:02:38 do coroutines(microthreads, fibers..) count as threads? 2011-07-24T06:03:03 specifically goroutines 2011-07-24T06:21:41 i don't know much about that kind of stuff, green threads whatever they are are allowed 2011-07-24T06:21:51 if that helps, great, otherwise no idea sorry :) 2011-07-24T06:28:58 antimatroid: I don't know about the design of goroutines. Here is a short definition: a function executing in parallel with other goroutines in the same address space. 2011-07-24T06:29:08 Goroutines are multiplexed onto multiple OS threads so if one should block, such as while waiting for I/O, others continue to run. Their design hides many of the complexities of thread creation and management. 2011-07-24T06:29:29 i don't really know anything to help you sorry, hopefully someone else will come along at some point who knows more 2011-07-24T06:29:43 thanks 2011-07-24T06:31:37 *** vdp is now known as vdp|afk 2011-07-24T06:36:51 Goroutines are multiplexed onto multiple OS threads 2011-07-24T06:37:09 it wont work if it uses multiple os level threads 2011-07-24T06:38:34 i guess you can somehow tune your system to not spawn other threads than a main one 2011-07-24T06:38:50 but you will use any parallel speedups of course 2011-07-24T06:39:02 will lose* 2011-07-24T06:48:34 mcstar: by default it uses just one. It can be adjusted by setting GOMAXPROCS to something bigger than 1. So if I use the default setting I should be fine? 2011-07-24T06:49:09 i guess, yes 2011-07-24T06:49:24 *** Kingpin13 has quit IRC (Ping timeout: 258 seconds) 2011-07-24T06:49:27 mcstar: thanks 2011-07-24T06:52:09 i dont see anything in the Go starter bot that resembles setting such an option 2011-07-24T06:52:34 my guess is that the go compile script automatically takes care of this so you neednt do anything 2011-07-24T06:53:16 when you see mcleopold you should talk to him if you need more info 2011-07-24T07:22:00 *** foRei has quit IRC (Quit: Bye) 2011-07-24T07:24:52 http://www.reddit.com/r/australia/comments/iy9li/streaming_in_australia/ haha 2011-07-24T07:24:59 poor bastard 2011-07-24T07:26:18 *** humanfromearth has quit IRC (Ping timeout: 255 seconds) 2011-07-24T07:40:06 *** humanfromearth has joined #aichallenge 2011-07-24T07:47:53 *** choas has quit IRC (Ping timeout: 250 seconds) 2011-07-24T07:51:38 *** mathis_ has joined #aichallenge 2011-07-24T08:07:51 *** Lithosphere has joined #aichallenge 2011-07-24T08:08:59 *** mcstar has left #aichallenge ("WeeChat 0.3.5") 2011-07-24T08:15:31 *** chturne has joined #aichallenge 2011-07-24T08:27:21 *** Accoun has quit IRC (Read error: Connection reset by peer) 2011-07-24T08:28:39 *** Accoun has joined #aichallenge 2011-07-24T08:51:17 *** mceier has joined #aichallenge 2011-07-24T09:02:36 *** humanfromearth has quit IRC (Ping timeout: 255 seconds) 2011-07-24T09:07:57 *** FireFly has joined #aichallenge 2011-07-24T09:18:57 *** humanfromearth has joined #aichallenge 2011-07-24T09:32:23 *** jbrechtel_ has joined #aichallenge 2011-07-24T09:39:32 we could use something like cpulimit to limit the percent of cpu that bots can use. that would mean we don't have to have that rule about threads anymore 2011-07-24T09:40:47 this would mean that a process can only use up to one cpu's worth of resources: cpulimit -l 100 2011-07-24T09:41:00 hmm, i guess multithreading could still interfere with other bots though 2011-07-24T09:41:21 i guess that shoots down the idea then 2011-07-24T09:42:25 oh, a better alternative: 2011-07-24T09:42:30 taskset 2011-07-24T09:42:39 just assign a process to a particular core 2011-07-24T09:57:50 jmcarthur: I think that running bots on the server should be optional. Someone should be able to use it's own machine to run the bot. Like tcpserver for planet wars last year. 2011-07-24T10:00:14 this reduces a number of security concerns and makes the bot creators more free to use different techniques such as running a bot on a cluster, however that may be unfair for others that don't have a 'cluster'. 2011-07-24T10:06:42 humanfromearth: This probably gives an unfair advantage to people with more expensive computing hardware. 2011-07-24T10:07:01 jbroman: Yes, I pointed that out 2011-07-24T10:07:08 can someone take a look at my pull request? It solves a number of issues. 2011-07-24T10:08:42 it would also give an unfair advantage to people with really bad internet connections 2011-07-24T10:08:48 err 2011-07-24T10:08:51 disadvantage 2011-07-24T10:09:28 well i guess they could opt for the server if they wanted, but it probably wouldn't be worth it to support *two* ways to compete 2011-07-24T10:10:33 you're right, but then again.. 2mb zip files + 1 process/thread is very limiting in these days 2011-07-24T10:11:35 <_flag> It's less limiting then ever before. What do you mean? 2011-07-24T10:12:02 <_flag> It's millions and millions of instructions 2011-07-24T10:12:40 Maybe not all of them are instructions, maybe some are statistical data 2011-07-24T10:13:01 <_flag> I'm talking about the process/thread thing 2011-07-24T10:13:46 <_flag> I'm pretty sure you can create files within your directory and collect data on the server anyway (but I'm not 100% sure) 2011-07-24T10:14:03 _flag: you cannot write files on the server 2011-07-24T10:14:08 it's in the rulles 2011-07-24T10:14:33 <_flag> Then how were people able to create learning algorithms last contest? 2011-07-24T10:14:47 they didn't 2011-07-24T10:14:49 >) 2011-07-24T10:14:56 <_flag> They did 2011-07-24T10:15:27 maybe locally and then uploaded the results on the server, but you can't write files on the server.. you can just read them. 2011-07-24T10:15:53 <_flag> Hmmm..., I guess they must've used the tcp server to configure it 2011-07-24T10:16:17 probably, but during the actual contest tcpserver was not used 2011-07-24T10:16:38 see how people won last 2 times 2011-07-24T10:16:56 <_flag> All the games are played on the server, yes 2011-07-24T10:17:05 no GA, neural nets.. 2011-07-24T10:17:07 <_flag> But everyone used the tcp server to test changes 2011-07-24T10:17:49 maybe people just don't know how to code them, or they didn't apply.. I'm not an expert myself.. 2011-07-24T10:19:16 <_flag> I'm not sure how those would apply to either PlanetWars or Tron 2011-07-24T10:19:16 it still seems rather strange that standard AI stuff didn't help much 2011-07-24T10:21:38 but maybe, just maybe it's related to the fact that you can't really run a massively parallel bot on the contest server 2011-07-24T10:22:38 so that being the limitation people resorted to heuristic stuff 2011-07-24T10:22:46 I'm just speculating here... 2011-07-24T10:24:46 <_flag> I don't think people are limiting themselves to heuristic stuff 2011-07-24T10:25:26 <_flag> You couldn't really describe the strategies currently being employed on the server as "heuristic stuff" 2011-07-24T10:26:00 <_flag> Second, I don't think the idea is to suck as much processing power as possible, it's to be efficiant 2011-07-24T10:26:34 <_flag> It adds an extra constraint on the competition 2011-07-24T10:27:31 humanfromearth: for the record, zip, tar.gz, tar.bz2, and tar.xz are all supported now 2011-07-24T10:27:45 humanfromearth: so you can squeeze in quite a bit of stuff, especially with xz 2011-07-24T10:28:35 _flag: the learning bots just learned form replays available on the web site 2011-07-24T10:28:39 *from replays 2011-07-24T10:30:45 i'd be fine with allowing multithreading if we divided the time limit by the number of cores and gave each bot relatively exclusive access to machine resources, one at a time 2011-07-24T10:30:45 *** sigh has quit IRC (Remote host closed the connection) 2011-07-24T10:30:55 but then you're disadvantaged if you *don't* use parallelism 2011-07-24T10:31:18 and it turns the contest into a bit more of an efficiency contest, really 2011-07-24T10:32:20 arguably, though, since multicore is the modern thing, there isn't much harm in favoring modern optimizations like parallelism 2011-07-24T10:32:37 it also disadvantages some languages yet further though 2011-07-24T10:32:48 due to global locks and such that some runtimes have 2011-07-24T10:33:00 jmcarthur: you mean python? 2011-07-24T10:33:35 that would be an example, yeah 2011-07-24T10:33:48 i mean, you could always spawn multiple processes, but that kind of sucks 2011-07-24T10:34:23 to be more exact: cpython.. pypy would be a good alternative for that problem 2011-07-24T10:34:28 sure 2011-07-24T10:34:46 ocaml is another example. you can use multiple OS threads in ocaml, but it comes with risks due to the way the GC works 2011-07-24T10:36:09 jmcarthur: I think you're missing the point here.. it's not a language contest.. it's about making clever bots 2011-07-24T10:36:30 exactly 2011-07-24T10:36:43 so if a language is disadvantaged then quite simply.. don't use it 2011-07-24T10:36:57 use something else that does the job 2011-07-24T10:37:21 i agree, but the majority may not 2011-07-24T10:38:19 but to further drive home our mutual agreement, languages *already* impose significant advantages/disadvantages on your, depending on the kind of AI you are writing 2011-07-24T10:38:24 *on you 2011-07-24T10:39:07 jmcarthur: :) true 2011-07-24T10:39:24 also, we want to attract programmers with all kinds of backgrounds 2011-07-24T10:40:23 many might not have much experience with C++. since the *main* point is to write good AI instead of competing in a number crunching contest, it's nice to eliminate language inadequacies when we can, at least 2011-07-24T10:40:39 otherwise we would just make everybody use fortran ;) 2011-07-24T10:41:03 jmcarthur: you wouldn't have a contest then.. 2011-07-24T10:41:24 well, you would, but it wouldn't be an AI contest 2011-07-24T10:41:48 or maybe i'm misinterpreting what you are saying 2011-07-24T10:42:26 not many young people know fortran these days... 2011-07-24T10:42:40 my point is that if we design the contest to balance the different languages as much as possible then we will have a much greater turnout, that's all 2011-07-24T10:42:57 i think you are agreeing with me without realizing it 2011-07-24T10:44:09 all that said, i think parallelism is fun and i wish we were allowed to use it 2011-07-24T10:44:27 it's just that i see that there are some good reasons not to allow it 2011-07-24T10:46:21 jmcarthur: and I think that's a big limitation, I guess what I wanted to ask was how decides this? 2011-07-24T10:48:15 *** onensora has joined #aichallenge 2011-07-24T10:53:01 humanfromearth: at this point i'm not sure. normally i'd say jeff, but he seems to never be around 2011-07-24T10:53:56 the leadership around here seems to mostly be de facto nowadays 2011-07-24T10:54:38 jmcarthur: I see 2011-07-24T10:54:48 *** jbrechtel_ has quit IRC (Ping timeout: 276 seconds) 2011-07-24T10:54:48 overall i see that as a positive, but it does have the downside that you never know who to talk to about various issues, and everybody is afraid to make certain kinds of major decisions 2011-07-24T10:56:16 also my pull request hangs there for 1 week.. 2011-07-24T10:56:48 i think that's just because everybody is slowing down and gradually becoming less excited about this next contest :( 2011-07-24T10:58:09 i'll check out your pull request 2011-07-24T10:58:22 thanks 2011-07-24T11:00:32 *** Kingpin13 has joined #aichallenge 2011-07-24T11:01:25 i'm not a Go programmer, but it looks like there's been a decent amount of review on it, and of course for the most part there aren't any major changes, so i guess i can go ahead merge it 2011-07-24T11:01:57 that's right 2011-07-24T11:02:22 it solves a few python issues as well 2011-07-24T11:02:31 aichallenge: Jake McArthur epsilon * rc2da8d8 / (6 files in 2 dirs): 2011-07-24T11:02:31 aichallenge: Merge pull request #222 from humanfromearth/epsilon 2011-07-24T11:02:31 aichallenge: A few bugs for python package. Formated Go code with gofmt and removed unicode stuff. - https://github.com/aichallenge/aichallenge/commit/c2da8d884ac6006d8aa1c2c0d3a265eb0469a611 2011-07-24T11:03:16 there you go :) 2011-07-24T11:03:32 sorry that took so long, but i think the delay was valuable anyway since it got some review 2011-07-24T11:03:46 thanks.. I think you can close #220 and #221 2011-07-24T11:05:51 done and done 2011-07-24T11:07:05 thanks.. one more thing.. do you happen to know who is maintaining the go starter package? 2011-07-24T11:07:49 I made a few changes to it and maybe I can work out a merge 2011-07-24T11:10:45 i have no idea at all 2011-07-24T11:10:55 nor do i know where to look 2011-07-24T11:13:48 ok.. I'll just make a pull request and see what happens 2011-07-24T11:20:08 *** chturne has quit IRC (Remote host closed the connection) 2011-07-24T11:26:53 *** mceier has quit IRC (Quit: leaving) 2011-07-24T11:29:45 *** Parsley has joined #aichallenge 2011-07-24T11:32:29 from what i understand, ga and neural networks don't really work very well? 2011-07-24T11:32:37 and only really get used when there is no other option 2011-07-24T11:32:55 but yeah, in general you need to train any variables locally 2011-07-24T11:33:06 there was a ga bot for planet wars, spaceinvaders maybe? 2011-07-24T11:33:14 it didn't do all that well 2011-07-24T11:34:31 it did already considering that it didn't have very long to learn 2011-07-24T11:34:35 *did alright 2011-07-24T11:34:59 *** mathis_ has quit IRC (Quit: Page closed) 2011-07-24T11:37:56 *** ArchMonkey has quit IRC (Quit: ArchMonkey) 2011-07-24T11:38:06 *** mathis_ has joined #aichallenge 2011-07-24T11:53:10 *** mathis_ has quit IRC (Quit: Page closed) 2011-07-24T11:57:52 *** savanus has joined #aichallenge 2011-07-24T11:58:38 *** savanus has quit IRC (Client Quit) 2011-07-24T12:02:24 *** mathis_ has joined #aichallenge 2011-07-24T12:25:40 *** mcstar has joined #aichallenge 2011-07-24T12:56:28 *** amstan has joined #aichallenge 2011-07-24T12:56:28 *** ChanServ sets mode: +o amstan 2011-07-24T13:03:50 *** nullkuhl has quit IRC (Read error: Connection reset by peer) 2011-07-24T13:05:56 *** nullkuhl has joined #aichallenge 2011-07-24T13:05:56 *** nullkuhl has joined #aichallenge 2011-07-24T13:13:53 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-07-24T13:14:03 *** antimatroid has joined #aichallenge 2011-07-24T13:34:41 *** Bhaskerh has joined #aichallenge 2011-07-24T13:36:13 *** [airvey] <[airvey]!~herve@obs92-5-88-163-187-42.fbx.proxad.net> has joined #aichallenge 2011-07-24T14:05:36 *** choas has joined #aichallenge 2011-07-24T14:06:47 *** amstan_ has joined #aichallenge 2011-07-24T14:06:47 *** ChanServ sets mode: +o amstan_ 2011-07-24T14:20:51 *** amstan has quit IRC (Remote host closed the connection) 2011-07-24T14:22:06 *** amstan_ has quit IRC (Ping timeout: 255 seconds) 2011-07-24T14:23:20 *** amstan has joined #aichallenge 2011-07-24T14:23:20 *** ChanServ sets mode: +o amstan 2011-07-24T14:24:15 mleise: http://www.unixmen.com/software/1834-nvidia-27521-has-been-released-with-installation-instructions 2011-07-24T14:24:32 mleise: will fix that crash thing with maximizing konsole 2011-07-24T14:29:47 *** McLeopold has joined #aichallenge 2011-07-24T14:31:11 @later tell amstan I'll have my head back in the contest on Thursday, on vacation 2011-07-24T14:31:11 McLeopold: As you wish. 2011-07-24T14:31:32 *** jbrechtel_ has joined #aichallenge 2011-07-24T14:32:49 *** Bhaskerh has quit IRC (Ping timeout: 240 seconds) 2011-07-24T14:34:24 hey guys, would it be possible to add ccl to the available languages? 2011-07-24T14:34:33 (sorry, implementation) 2011-07-24T14:36:47 McLeopold: k 2011-07-24T14:37:13 *** humanfromearth has left #aichallenge 2011-07-24T14:43:26 mcstar: put together an installation guide and patch the source tree and you should be gold 2011-07-24T14:43:44 o.O 2011-07-24T14:44:24 im sure theres an apt command that lets you install it easier 2011-07-24T14:44:28 or wth do you mean? 2011-07-24T14:45:47 is clozure cl not among natty's packages? jmcarthur is this what you meant? 2011-07-24T14:46:46 jmcarthur: can i ask you a big favor? are you near youre arch system? 2011-07-24T14:47:04 *** jbrechtel__ has joined #aichallenge 2011-07-24T14:48:42 mcstar: the instructions would be whatever packages need to be installed, etc. 2011-07-24T14:48:51 this would also need to be added to the worker setup script 2011-07-24T14:48:57 *** jbrechtel_ has quit IRC (Ping timeout: 258 seconds) 2011-07-24T14:49:00 and you also need to tell the worker how to build and run it 2011-07-24T14:49:09 i am on my arch system, yes 2011-07-24T14:49:29 it seems i have a problem with my sbcl 2011-07-24T14:49:49 i was asking around on #lisp, and it seems i dont get the same debugging output as others 2011-07-24T14:50:04 do you have by any chance sbcl installed? 2011-07-24T14:50:20 no. does it require any configuration? 2011-07-24T14:50:27 no 2011-07-24T14:50:34 its 10MB 2011-07-24T14:50:41 i'll install it 2011-07-24T14:50:50 ah, thx, youre sweet 2011-07-24T14:51:09 so once it's installed what would you like me to do? 2011-07-24T14:51:18 it's installed 2011-07-24T14:51:25 http://pastebin.com/xrFd4WTF 2011-07-24T14:51:38 the contents of this paste should go into a file.lisp 2011-07-24T14:52:14 and then? 2011-07-24T14:52:19 then, in the same directory, type sbcl 2011-07-24T14:52:26 (load "file.lisp") 2011-07-24T14:52:31 (fun1 34234234) 2011-07-24T14:52:44 and i am interested in the full output that it shows 2011-07-24T14:52:53 so pls pastebin it 2011-07-24T14:53:57 http://pastebin.com/YAZ0SqhC 2011-07-24T14:54:25 thx god 2011-07-24T14:54:30 its the same as mine 2011-07-24T14:54:40 and isnt what it supposed to be 2011-07-24T14:54:50 lovely 2011-07-24T14:54:55 jmcarthur: thank you very much 2011-07-24T14:55:00 you're welcome 2011-07-24T14:56:19 *** jbrechtel_ has joined #aichallenge 2011-07-24T14:59:29 *** jbrechtel__ has quit IRC (Ping timeout: 250 seconds) 2011-07-24T15:14:44 amstan: what crash with maximizing konsole? I'm using gnome-terminal anyway :p 2011-07-24T15:15:07 mleise: then it wasn't you 2011-07-24T15:15:21 idk, i was talking to someone in this channel about the nvidia crashes in kde 2011-07-24T15:15:29 amstan: 2011-07-24T15:15:31 hello 2011-07-24T15:15:35 :D 2011-07-24T15:15:40 mcstar: yes, probably you 2011-07-24T15:15:45 probably 2011-07-24T15:15:51 mcstar: http://www.unixmen.com/software/1834-nvidia-27521-has-been-released-with-installation-instructions 2011-07-24T15:16:09 yeah, several weeks ago it solved itself 2011-07-24T15:16:22 but since i dont use kde anymore.... 2011-07-24T15:16:35 i was switching to awesome at that time when we talked about this 2011-07-24T15:16:46 but thx 2011-07-24T15:26:05 *** McLeopold has left #aichallenge 2011-07-24T15:28:31 *** jbrechtel__ has joined #aichallenge 2011-07-24T15:31:11 *** jbrechtel_ has quit IRC (Ping timeout: 240 seconds) 2011-07-24T15:42:35 i don't think, i'll ever get that tcp-server running, but have some tron again in the meantime .. http://tron-game.appspot.com/ 2011-07-24T15:48:22 Bots that caused more than 50 crashes or timeouts will see their source code exposed. 2011-07-24T15:48:24 huh? 2011-07-24T15:48:30 why is that good? 2011-07-24T15:48:32 *** ibdknox has joined #aichallenge 2011-07-24T15:48:38 i dont want to see those sources 2011-07-24T15:48:40 :D 2011-07-24T15:55:02 dont break my heart 2011-07-24T15:55:23 *** Accoun has quit IRC () 2011-07-24T15:59:18 *** [airvey] <[airvey]!~herve@obs92-5-88-163-187-42.fbx.proxad.net> has quit IRC (Read error: No route to host) 2011-07-24T16:00:29 *** [airvey] <[airvey]!~herve@obs92-5-88-163-187-42.fbx.proxad.net> has joined #aichallenge 2011-07-24T16:07:52 *** Accoun has joined #aichallenge 2011-07-24T16:15:17 mcstar, if u update it, it will start from scratch.. 2011-07-24T16:33:53 *** foRei has joined #aichallenge 2011-07-24T16:47:12 gn 2011-07-24T16:47:14 *** mcstar has left #aichallenge ("WeeChat 0.3.5") 2011-07-24T17:49:12 mleise amstan the output from the scheduler test started a while back http://pastebin.com/w8hc7yHy 2011-07-24T17:49:34 no time stamps though so can't tell if they all happened in one burst or not 2011-07-24T17:49:47 Shame on me. 2011-07-24T17:50:18 I do know when I checked it the first couple times there had been no output though 2011-07-24T17:50:55 At least there are no extreme values. That is good to know. 2011-07-24T17:51:12 janzert: didn't i run it with time? 2011-07-24T17:51:13 also I didn't stop it so it is still running 2011-07-24T17:51:27 janzert: i think i forgot 2011-07-24T17:51:33 doesn't appear to be 2011-07-24T17:51:41 but that wouldn't help anything here anyway 2011-07-24T17:51:46 janzert: can you restart it with time? 2011-07-24T17:51:58 janzert: would be nice to know how many time errors we have in a time period 2011-07-24T17:52:23 I can add time and date to the lines if you want 2011-07-24T17:52:23 a timestamp per line would be much better 2011-07-24T17:52:28 * janzert nods 2011-07-24T17:52:29 mleise: yes 2011-07-24T17:52:31 :p 2011-07-24T17:52:39 any ideas on the timestamp format? 2011-07-24T17:52:50 mleise: just get the default one 2011-07-24T17:53:17 date and time down to the millisecond would be great 2011-07-24T17:53:29 exact format doesn't matter 2011-07-24T17:53:48 the millisecond after the delay, ok 2011-07-24T17:53:52 just down to the second is really going to be enough of course 2011-07-24T17:55:08 I just have to look up date formatting... 2011-07-24T18:08:07 *** ltriant has joined #aichallenge 2011-07-24T18:17:16 *** martinording has joined #aichallenge 2011-07-24T18:18:25 32-bit executable is here: http://marco-leise.homedns.org/scheduler_test 2011-07-24T18:19:12 hey, anyone knows when c# bots will stop failing the tests? 2011-07-24T18:20:59 *** berak has quit IRC (Quit: ChatZilla 0.9.84 [SeaMonkey 2.0a3/20090223135443]) 2011-07-24T18:21:28 mleise: hmm, could you commit the updated source. the worker doesn't like that binary for some reason 2011-07-24T18:21:50 oh, was it maybe not statically linked? 2011-07-24T18:21:50 kk 2011-07-24T18:22:14 I'll just upload the source 2011-07-24T18:22:26 ok 2011-07-24T18:23:48