2011-04-22T00:14:03 *** janzert1 has quit IRC (Read error: Connection reset by peer) 2011-04-22T00:14:34 *** yasith__ has joined #aichallenge 2011-04-22T00:15:15 *** janzert has joined #aichallenge 2011-04-22T00:16:01 *** yasith_ has quit IRC (Ping timeout: 248 seconds) 2011-04-22T00:16:55 *** Kingpin13 has quit IRC (Ping timeout: 260 seconds) 2011-04-22T00:40:36 *** yasith__ has quit IRC (Ping timeout: 246 seconds) 2011-04-22T00:47:56 *** siddharth has joined #aichallenge 2011-04-22T00:50:13 *** sids_aquarius has joined #aichallenge 2011-04-22T00:53:14 sigh: :P 2011-04-22T00:55:01 *** yasith has joined #aichallenge 2011-04-22T01:22:33 *** _flag <_flag!~flag@69.165.173.172> has quit IRC (Quit: Lost terminal) 2011-04-22T01:36:54 *** McLeopold has joined #aichallenge 2011-04-22T01:38:25 *** yasith has quit IRC (Remote host closed the connection) 2011-04-22T01:49:41 *** Ice_Harley has joined #aichallenge 2011-04-22T01:58:25 fuuuu amazon 2011-04-22T01:59:15 http://www.reddit.com/r/reddit.com/comments/gv5s8/were_back/ 2011-04-22T01:59:16 huzzah 2011-04-22T01:59:35 :P 2011-04-22T01:59:36 still can't login though 2011-04-22T01:59:55 must have been up for like 2 seconds 2011-04-22T02:00:28 somebodies voting on it 2011-04-22T02:01:35 people are adding commengts 2011-04-22T02:01:54 *** ahmedhelal has quit IRC (Read error: Connection reset by peer) 2011-04-22T02:01:59 must be bringing things up in stages 2011-04-22T02:02:24 *** ahmedhelal has joined #aichallenge 2011-04-22T02:02:55 it says there is one comment 2011-04-22T02:03:00 but then shows none for me :\ 2011-04-22T02:03:28 look at this one: http://www.reddit.com/r/reddit.com/comments/gv5s5/and_were_back/ 2011-04-22T02:05:04 interesting 2011-04-22T02:05:09 i can get those comments 2011-04-22T02:05:24 but now i see just 3 comments when loading the page, i think i'm being fed a cached version? 2011-04-22T02:07:45 *shrug* 2011-04-22T02:10:51 *** kaemo has joined #aichallenge 2011-04-22T02:14:21 *** McLeopold has left #aichallenge 2011-04-22T02:30:20 haha sorry, what did I manage to do 2011-04-22T02:31:41 you managed to type an obscure control character into the channel the other day, which caused the logs to not get served correctly 2011-04-22T02:34:04 might I inquire as to where? 2011-04-22T02:35:35 it *was* the 5th line of http://contestbot.hypertriangle.com/channellogs/%23aichallenge/%23aichallenge.2011-04-20.log 2011-04-22T02:35:59 you had typed "He don' seem to be.n" where the . is the character 0x10 2011-04-22T02:36:39 apparently it's called the "Data link escape"... no idea what it does 2011-04-22T02:37:10 *** berak has joined #aichallenge 2011-04-22T02:37:29 * janzert wonders if it was random data corruption somewhere along the line 2011-04-22T02:40:06 *** sigh has quit IRC (Quit: Leaving...) 2011-04-22T02:40:19 *** sigh has joined #aichallenge 2011-04-22T02:42:57 interesting, most places other than wikipedia seem to list 0x10 as backspace 2011-04-22T02:44:01 janzert: ascii on my system says it is DLE 2011-04-22T02:44:45 ascii 0x10 2011-04-22T02:44:45 ASCII 1/0 is decimal 016, hex 10, octal 020, bits 00010000: called ^P, DLE 2011-04-22T02:44:47 Official name: Data Link Escape 2011-04-22T02:45:01 the first line is the command I ran 2011-04-22T02:46:17 yeah, I see that a few other places now as well. two examples that list it as backspace are http://www.asciitable.com and http://bestofthisweb.com/blogs/tag/ascii-table/ 2011-04-22T02:46:50 although it looks like the second is just a rip off of the first 2011-04-22T02:47:07 backspace is 010 2011-04-22T02:47:12 those sites get it right 2011-04-22T02:47:23 are you sure you are looking at them correctly :P 2011-04-22T02:47:35 oops, yeah I was looking at the octal column 2011-04-22T02:49:59 I wonder why it forced a download... I think it's the browser's fault, not apache 2011-04-22T02:50:27 must trigger binary file handling? 2011-04-22T02:50:37 yeah 2011-04-22T02:51:10 (HTTP response headers for the borked log file was the same as the other files back before it was fixed, which is why I don't think it is apache) 2011-04-22T02:51:29 *** berak has quit IRC (Read error: Connection reset by peer) 2011-04-22T02:55:24 bah... if I serves the logs to myself with "python -m SimpleHTTPServer" it thinks all the logs are binary files 2011-04-22T02:55:50 wrong mime type in the headers? 2011-04-22T02:56:44 yeah, application/octet-stream 2011-04-22T02:57:27 I believe a DLE needs to be terminated by another command 2011-04-22T02:57:41 what does it actually do? 2011-04-22T02:57:53 Interprets everything following as raw data 2011-04-22T02:58:33 http://www.lammertbies.nl/comm/info/ascii-characters.html#data 2011-04-22T02:59:25 yeah... I've been reading that 2011-04-22T02:59:42 never seen it before, I wonder what uses it 2011-04-22T02:59:59 it sounds to me like not much of anything ever used it 2011-04-22T03:00:09 does this mean that I somehow managed to type three plus signs in a row? 2011-04-22T03:00:17 no 2011-04-22T03:00:27 nah, three plus signs are safe 2011-04-22T03:00:28 ++ 2011-04-22T03:00:30 +++ 2011-04-22T03:00:40 hayes modems used the three plus signs instead of the DLE character 2011-04-22T03:01:12 (well, with silence before and after) 2011-04-22T03:01:39 Ah yes I see 2011-04-22T03:01:58 So then it is command p that's the worry? 2011-04-22T03:02:19 ^P is just the way it is rendered in plain text 2011-04-22T03:02:43 gotcha 2011-04-22T03:03:24 it's a byte with just the 4th bit set ;) 2011-04-22T03:03:45 I have a feeling that any wierd control code probably triggers the browser to treat it as binary data 2011-04-22T03:04:05 well this was a strange and informative incident. Sorry to anyone I caused problems for : ) 2011-04-22T03:04:38 :) 2011-04-22T03:04:48 jaycobcoleman: what IRC client do you use? 2011-04-22T03:05:14 I'm using freenode 2011-04-22T03:05:37 web interface? 2011-04-22T03:05:59 freenode is the server, what program are you using? 2011-04-22T03:07:19 or http://webchat.freenode.net/ possibly? 2011-04-22T03:07:47 oh yes, sorry 2011-04-22T03:11:03 *** Frontier has joined #aichallenge 2011-04-22T03:21:32 Why did you want to know the client? 2011-04-22T03:35:20 *** _0x47 <_0x47!~slash@p3EE0BFBC.dip.t-dialin.net> has joined #aichallenge 2011-04-22T03:36:11 *** berak has joined #aichallenge 2011-04-22T03:39:27 *** janzert has quit IRC (Ping timeout: 276 seconds) 2011-04-22T03:39:46 *** amstan_ has quit IRC (Ping timeout: 250 seconds) 2011-04-22T03:47:12 *** Kingpin13 has joined #aichallenge 2011-04-22T03:47:14 *** Kingpin13 has joined #aichallenge 2011-04-22T03:48:37 *** woudshoo` has joined #aichallenge 2011-04-22T03:53:36 jaycobcoleman: just wondering if it might have been able to add those types of characters accidentally 2011-04-22T03:57:15 Alright, I imagined so. I tried looking around myself and there's shockingly little reading marerial on the subject 2011-04-22T04:33:30 *** sids_aquarius has quit IRC (Ping timeout: 258 seconds) 2011-04-22T04:34:26 *** sids_aquarius has joined #aichallenge 2011-04-22T04:37:08 *** Stocha has joined #aichallenge 2011-04-22T04:37:54 is there any tutorial for how the server stuff works ? what has been implemented already ? what is still in work. Question for begining to thing about a tcp server 2011-04-22T04:49:20 There's no up-to-date info on server setup 2011-04-22T04:58:54 What aspects are you looking for help with? because I believe simply running server_setup.py on a clean ubuntu install works at this point 2011-04-22T04:59:08 for setup, obviously 2011-04-22T05:00:15 but if you're looking for tutorials on what to do once you've gotten that then I think you're out of luck 2011-04-22T05:03:41 What i would like is way for n bot to get a ranking and put it on a web page. 2011-04-22T05:03:54 incremental update would be best. 2011-04-22T05:04:19 then i could try to specify a way to furnish the bot set from distant bots through tcp means. 2011-04-22T05:04:30 but i have to have the local server working properly first. 2011-04-22T05:04:40 so i have to know if it already exists i guess :p 2011-04-22T05:04:45 and they learn to set it up. 2011-04-22T05:04:48 then 2011-04-22T05:05:58 i need to know what exists so far. maybe i 'll have a look at server_setup.py when i have time. But if you tell me how it works ? 2011-04-22T05:10:24 The idea is to have a local server than can rank all bot from a local bot set. Then i will make it so that that a local bot can in fact call through tcp a distant bot. 2011-04-22T05:10:38 and also log the games 2011-04-22T05:17:20 well if you want real answers then Janzert is the man you wanna talk to 2011-04-22T05:19:41 I would tell you what I do know, but actually janzert just updated the packages a few days ago and it's well documented enough that I think it'd be relatively straight-forward for you to understand what's going on 2011-04-22T05:22:10 i'll try to have a look if / when i have time then. 2011-04-22T05:23:23 sigh: I think I have a very simple and elegant solution for the problem of dead ants overwriting food info from the same turn 2011-04-22T05:23:45 don't display dead ants on the map? 2011-04-22T05:24:24 well there's an option. really why was it setup to do that in the first place? 2011-04-22T05:24:40 dunno, ask mcleopold 2011-04-22T05:25:01 the slightly more effort way is to place the food after the dead ants 2011-04-22T05:25:11 but really, there is no reason to mark dead ants on the map 2011-04-22T05:26:56 yes indeed, nonetheless I think right before the end of the render_changes function in the ants.py that runs the game we should just insert visible_updates.sort() 2011-04-22T05:27:45 It would get rid of the need for a parsing system, which should make updates run slightly faster 2011-04-22T05:31:23 sure, but it's not so good to make the bots rely on the particular order that the engine sends out updates 2011-04-22T05:31:36 unless that is made part of the specs 2011-04-22T05:31:54 especially when there is no real need to mark dead ants on the map 2011-04-22T05:33:28 they wouldn't rely on it. The current setup would still work, only it would allow for a more efficient system to be used in its place 2011-04-22T05:33:54 and wether or not you care about marking dead ants, it will speed up updates, which is very important 2011-04-22T05:34:55 why would it speed up updates? 2011-04-22T05:35:57 it would require less testing of each input to see which method should be followed 2011-04-22T05:36:59 the updates are already grouped by type 2011-04-22T05:37:39 but the bots should not be relying on that 2011-04-22T05:38:18 currently, the engine DOES sort, but the order is explictly ants, food, dead ants... dead ants is last because it is the least important 2011-04-22T05:38:45 this can be changed BUT, making the order part of the specs is overly restrictive 2011-04-22T05:39:20 *** needsch has joined #aichallenge 2011-04-22T05:39:22 if it isn't part of the specs the bots should not be relying on it 2011-04-22T05:39:40 I'm happy to change the engine if there is some consensus on this though 2011-04-22T05:39:48 what specs are you referring to? 2011-04-22T05:41:10 well, currently it is on the wiki... but the point is that if your bot is relying on it then it should have some guarantee that it will always work that way 2011-04-22T05:41:17 concensus about droping dead ants ? 2011-04-22T05:41:51 what about dead ants only marked when there is no food/ no ant that have been there ? 2011-04-22T05:41:52 currently the only reason the engine sorts is to ensure deterministic output 2011-04-22T05:42:05 Stocha: you can do that, but why? 2011-04-22T05:42:23 i though it would simplify. maybe i didn't get the problem 2011-04-22T05:42:37 incomplete infomation like that is next to useless 2011-04-22T05:42:47 yeah but dead ant seems useless 2011-04-22T05:42:51 In it's current state, the updating method, at least in the python bot, is already restrictive to the extent that basically any changes to it will result in it no longer working 2011-04-22T05:43:04 except maybe if an ant dies at the limit of the vision radius 2011-04-22T05:43:08 dead ants is useful, because there is no other way to track ants properly 2011-04-22T05:43:10 That's because the input recieved is so restrictive 2011-04-22T05:43:10 wich means it wasn't killed by you. 2011-04-22T05:43:36 usefull for bots ? 2011-04-22T05:44:11 the thing with dead ant for bot, is that it can be deduced, except in the limit cases above stated. 2011-04-22T05:44:12 If somebody is planning on making a bot they're pretty much going to have to rely on our pre-made updating functions anyway, so I think we might as well make them as fast as possible 2011-04-22T05:44:18 jaycobcoleman: the only restrictions are those detailed in the specifications 2011-04-22T05:44:34 how do they take jaycobcoleman ? 2011-04-22T05:44:46 if it's less than 10ms, then it's probably no big deal. 2011-04-22T05:44:57 if it's higher then i agree. 2011-04-22T05:45:08 I ignored dead ants, but now I noticed I like the info. When I look at my bot and you don't have dead ant information, it is sometimes hard to see what happened 2011-04-22T05:45:10 how long do they take 2011-04-22T05:45:47 you mean you think building the dead status from scratch is too demanding ? 2011-04-22T05:45:49 I haven't timed it, but I'd like to be able to use the same sytem for a large variety of maps 2011-04-22T05:46:13 jaycobcoleman: fast as possible is not the goal, if someone wants fast as possible (1) they shouldn't be using python, and (2) they will need to rewrite the starter pack anyway 2011-04-22T05:46:14 i guess i just don't get what the discussion is really about :p 2011-04-22T05:46:37 and you are looking at an area of the code which will never ever be a bottleneck 2011-04-22T05:46:57 if you want to improve the speed then work on the vision algorithm 2011-04-22T05:46:58 haha, I think using python probably is a reason to try to make what you're doing as fast as possible 2011-04-22T05:47:17 What's wrong with it? 2011-04-22T05:47:32 just slow? 2011-04-22T05:47:36 well, the current one is broken :) 2011-04-22T05:47:43 in the python starter bot 2011-04-22T05:48:09 the one in the engine is fast and works, but will NOT be put into the starter bot because it is complicated 2011-04-22T05:48:09 It works perfectly, the problem is with parsing at setup 2011-04-22T05:48:34 are you taking about vision? 2011-04-22T05:48:43 yes 2011-04-22T05:48:56 it is broken 2011-04-22T05:49:06 does it return all false or do you get an error of some sort? 2011-04-22T05:49:35 no, it returns results which are slightly wrong 2011-04-22T05:49:44 hang on, I'll pull up an example 2011-04-22T05:49:58 that'd be lovely 2011-04-22T05:49:59 Stocha: No, it is just that you can't see which move an enemy made just before it died. When watching replays (with my own visualizing thingy) this is really puzzling sometimes. 2011-04-22T05:50:36 jaycobcoleman: http://pastebin.com/gscdvw3u 2011-04-22T05:52:56 and parsing is never going to be a bottleneck, there is absolutely no reason to worry about optimising it for the starter pack 2011-04-22T05:53:18 main issue atm is correctness, so let's get that right :P 2011-04-22T05:53:27 Alright, well then moving onto this issue : ) I can't see why that would happen. 2011-04-22T05:53:41 however, I would like to use numpy arrays for a couple of reasons 2011-04-22T05:53:53 sure, you can use whatever you like 2011-04-22T05:53:59 they're simpler to see created 2011-04-22T05:54:02 but it won't be part of he starter pack :) 2011-04-22T05:54:07 Why 2011-04-22T05:55:03 they're much simpler to understand for someone who isn't great with programming 2011-04-22T05:55:21 maybe, but then they have to install numpy 2011-04-22T05:55:37 the fewer dependencies the better for a starter pack 2011-04-22T05:55:49 numpy has been standard for a while I thought 2011-04-22T05:56:08 numpy doesn't come with python 2011-04-22T05:56:40 hmm. It's weird that I already had it then, on a 2 week old install of Ubuntu 2011-04-22T05:57:30 *** ginDeVine has joined #aichallenge 2011-04-22T05:57:51 maybe you installed something else which has a numpy dependancy, or your ubuntu dist might have had it installed already 2011-04-22T05:58:07 *** Eruonen has joined #aichallenge 2011-04-22T06:00:13 I also have it on my mac evidently, and I don't do anything programming related with that 2011-04-22T06:00:26 that's unfortunate though 2011-04-22T06:01:02 the ability to call indexes in tuple form would make the code a lot simpler to understand 2011-04-22T06:03:20 *shrug* dunno... numpy is not part of python standard library 2011-04-22T06:03:23 anything further about starter packs, tell McLeopold :P 2011-04-22T06:03:48 woudshoo: good point 2011-04-22T06:04:06 alright. I haven't run into any odd vision problems like that though and I can't see any reason for it to happen 2011-04-22T06:04:48 jaycobcoleman: the problems are small enough that you generally won't notice unless you explictly test for it 2011-04-22T06:05:25 alright I'll run some test then 2011-04-22T06:06:04 the canonical failure case is when two ants are diagonal each other 2011-04-22T06:06:40 there's already a script ants/util/vision_compare.py to test still stuff 2011-04-22T06:07:32 oh wait, I understand exactly what's happening. Alright, I'm gonna mess with it for a little bit 2011-04-22T06:08:38 jaycobcoleman: if you are looking at vision algorithms then have a look at ants/util/vision_compare.py 2011-04-22T06:09:12 vision_by_ants_faster is the fastest for a python bot 2011-04-22T06:10:09 engine uses an algorithm which is even faster, but it can't be implemented by bots in the general case 2011-04-22T06:11:14 Oh I see. This is lovely 2011-04-22T06:12:45 if the not portion of "if (not vision[n_row][n_col] and self.vision_distance(a_loc, n_loc) <= self.viewradius2):" is removed from vision_by_square then it should work perfectly 2011-04-22T06:13:13 without getting weird diagonals problems 2011-04-22T06:13:24 *** ginDeVine has quit IRC (Quit: ginDeVine) 2011-04-22T06:14:09 jaycobcoleman: well, it will go into an infinite loop 2011-04-22T06:14:53 haha, silly me, what was I thinking :) 2011-04-22T06:15:02 if you want to fix vision_by_squares you end up with vision_by_distance_2 2011-04-22T06:16:05 as you can see, I've been over all this :P 2011-04-22T06:16:17 I can tell : ) 2011-04-22T06:16:53 *** Eruonen has quit IRC (Ping timeout: 260 seconds) 2011-04-22T06:17:18 alright, then what are you looking for in terms of vision code? 2011-04-22T06:18:50 sigh: is your reddit working yet? :P 2011-04-22T06:19:13 *** Frontier has quit IRC (Quit: Leaving.) 2011-04-22T06:19:15 something that works :P 2011-04-22T06:19:16 I changed the vision code in the starter pack once, but McLeopold reverted it :P 2011-04-22T06:19:21 mine sure isn't 2011-04-22T06:19:22 so I'm just leaving it up to him 2011-04-22T06:19:38 antimatroid: no 2011-04-22T06:19:50 what's the go with mcleopold? 2011-04-22T06:20:00 he doesn't seem to be around much atm 2011-04-22T06:20:02 dunno 2011-04-22T06:20:13 I haven't seen him at all today 2011-04-22T06:21:05 sigh: vision_by_ant_faster is the fastest you have so far? 2011-04-22T06:21:36 for a bot yeah... engine is faster though 2011-04-22T06:21:57 So I've been told :P 2011-04-22T06:22:09 technically the bot can be made even fast by doing a hybrid with what the engine is doing 2011-04-22T06:23:25 either vision_by_ant or a non-cached version of it should go in the starter pack 2011-04-22T06:23:45 *** skynetz has joined #aichallenge 2011-04-22T06:29:16 *** amstan_ has joined #aichallenge 2011-04-22T06:29:16 *** ChanServ sets mode: +o amstan_ 2011-04-22T06:29:45 it's just an abstract sort of idea at the moment, but I think it would be very fast to create a small array just large enough to fit the entire view radius of one ant into it, then specify a row/col offset for each element to be placed(of course based on the position of the ant) 2011-04-22T06:30:30 each turn you create a False array and for each ant compare the two using bitwise & and restoring it as the main map 2011-04-22T06:31:11 it's like drawing a picture and then pasting it everywhere you want it 2011-04-22T06:31:44 stocha was doing/trying something like that 2011-04-22T06:32:09 was he successful? 2011-04-22T06:33:20 Stocha: sigh says you had an idea similar to mine for vision. How has that turned out? 2011-04-22T06:33:32 dunno, but either way that approach is much more suited to a language like c 2011-04-22T06:33:40 Why is that? 2011-04-22T06:34:14 are you packing each cell as a bit? 2011-04-22T06:34:44 *** berak has quit IRC (Read error: Connection timed out) 2011-04-22T06:34:49 possibly? 2011-04-22T06:35:02 Why wouldn't you? 2011-04-22T06:35:30 *** berak has joined #aichallenge 2011-04-22T06:35:47 by the way, I meant bitwise or 2011-04-22T06:35:50 no reason, just hat the overhead in python doesn't seem worth it... but try it out and see 2011-04-22T06:36:26 psyco should speed an operation like that up almost to c standards 2011-04-22T06:36:43 try it and see how it goes 2011-04-22T06:36:53 there's nothing wrong with the actual approach 2011-04-22T06:36:58 Alright I'll do so :P 2011-04-22T06:38:19 I don't exactly understand the dos and don'ts of different programming languages yet 2011-04-22T06:39:13 know the dos or don'ts or why they happen? 2011-04-22T06:40:32 as in, what is good for one language yet terrible in another 2011-04-22T06:41:26 and of course I also don't know the reasons why the things happen that I don't know are happening 2011-04-22T06:42:29 i meant generally not specifically, but it's alright, i was just making sure you weren't saying you don't understand why some things work better in some langauges than others :P 2011-04-22T06:44:25 Well I don't entirely know what works in what yet. 2011-04-22T06:49:06 *** Eruonen has joined #aichallenge 2011-04-22T06:58:59 well my vision algorithm is okay. i didn't time it or anything. 2011-04-22T06:59:17 nor did i try to optimize it. 2011-04-22T06:59:49 but it uses the parallel capacity of 64 int bit logic. 2011-04-22T07:00:03 int 64 bit logic. 2011-04-22T07:00:40 I draw the view radius from every ant at the same time. 2011-04-22T07:01:00 Hmm. I don't think we're doing the same thing 2011-04-22T07:01:39 but you should compare it to the other methods in util/vision_compare.py 2011-04-22T07:02:01 Stocha: be careful doing that for vision 2011-04-22T07:02:08 it gets hairy with the euclidean metric 2011-04-22T07:02:17 *** berak has quit IRC (Read error: Connection timed out) 2011-04-22T07:02:22 ...i don't have much time for it. The thing is, i have working bitmap framework. so i will use that of non critical task. like vision. 2011-04-22T07:02:38 (unless vision somehow becomes time crititical, but there are few chances at this point) 2011-04-22T07:03:08 antimatroid: ? 2011-04-22T07:03:28 i use the straight forward double for loop. 2011-04-22T07:03:33 *** berak has joined #aichallenge 2011-04-22T07:03:40 with inequality x2 + y2 < rad2 2011-04-22T07:03:55 sigh: okay, so what you're doing in vision_by_ant _faster is almost exactly what I was talking about with the exception that I'm caching the entire overlay as an array so I can do it more quickly per ant. 2011-04-22T07:04:00 sure okay 2011-04-22T07:04:00 i didn't dare to optimized. Although you can calculate x2 y2 incrementally. 2011-04-22T07:04:11 why are you using x and y? 2011-04-22T07:04:16 You probably already figured that out, but I hadn't looked through your methods yet 2011-04-22T07:04:20 use row and col, or you will just confuse the hell out of yourself 2011-04-22T07:04:25 i say x, y doesn't really matter how you call them. 2011-04-22T07:04:31 hehe 2011-04-22T07:04:39 you about to get yelled at 2011-04-22T07:04:44 x implied indexing of the horizontal dimension 2011-04-22T07:04:45 i agree row and col are better for specifications though. 2011-04-22T07:05:05 i don't agree that it is better for quick prototyping :p 2011-04-22T07:05:19 antimatroid: just turn your monitor on its side :P 2011-04-22T07:05:19 transposing stuff is painful and unecessary 2011-04-22T07:05:29 i don't transpose 2011-04-22T07:05:37 sigh: for making moves you need to transpose directions 2011-04-22T07:05:42 i use x for what you call col. and y for what you call row 2011-04-22T07:05:49 and then you need to worry about where you do and don't need to convert 2011-04-22T07:06:02 Stocha: that's just even less clear to someone reading your code :P 2011-04-22T07:06:06 i never need to convert anything. I always use the same. 2011-04-22T07:06:15 actually, that's what one would expect 2011-04-22T07:06:20 so fare nobody is supposed to have read the code. 2011-04-22T07:06:21 but my above statements still stand 2011-04-22T07:06:48 it's still to be proven that row and cols are unambiguous 2011-04-22T07:06:57 unambiguous? 2011-04-22T07:07:03 what about arabian developper ? 2011-04-22T07:07:24 the contest is done in english 2011-04-22T07:08:24 x, y are clear to some. i always use top to bottome and left to right directions. 2011-04-22T07:09:03 because it was the coor system that was used on the early MacOs systems. 2011-04-22T07:09:17 when i learned about programming. in 1980's 2011-04-22T07:10:10 stil i agree that rows and cols are better for talking about coordinate. as long as it go left to right and top to bottom :p 2011-04-22T07:10:38 *** contestbot has joined #aichallenge 2011-04-22T07:10:39 *** ChanServ sets mode: +o contestbot 2011-04-22T07:10:49 antimatroid:are you sure ? 2011-04-22T07:10:51 yeah.. i'm playing with it a little 2011-04-22T07:11:02 :P 2011-04-22T07:11:26 is there a shelter for no-log time periods ? 2011-04-22T07:11:36 aww missed my chance to spam control characters into the channel without breaking the logs :P 2011-04-22T07:12:12 oh man.. why was that even installed? sudo apt-get purge cvs 2011-04-22T07:12:22 :P 2011-04-22T07:12:22 *** amstan_ is now known as amstan 2011-04-22T07:12:39 where do ppl usually run programs like these from? 2011-04-22T07:12:48 i'm thinking i should use /srv for this or something 2011-04-22T07:13:08 programs like what? 2011-04-22T07:13:15 supybot 2011-04-22T07:13:22 or game servers 2011-04-22T07:13:42 or annoying stuff that you don't want to use special users for.. but you don't want to run as root either 2011-04-22T07:14:55 /opt? 2011-04-22T07:15:17 i use opt, yes 2011-04-22T07:15:25 but i noticed that srv is more semantic than opt 2011-04-22T07:15:45 also: root@server:/srv# ls => cvs 2011-04-22T07:15:46 root@server:/srv# rm -Rf csv 2011-04-22T07:15:48 root@server:/srv# ls => cvs 2011-04-22T07:15:53 why is it still there? lol 2011-04-22T07:16:09 o_O 2011-04-22T07:16:16 haha 2011-04-22T07:16:22 try rm -Rf cvs 2011-04-22T07:16:40 oh 2011-04-22T07:16:56 you serious? i kept mistyping it 2011-04-22T07:17:19 I'm lazy, I would just have typed c 2011-04-22T07:17:36 or just since it's the only file in there 2011-04-22T07:20:03 some tab complete things refuse to do anything unless you type a few letters 2011-04-22T07:20:43 depends how you have things set up 2011-04-22T07:20:55 one of them is the kde start menu 2011-04-22T07:21:05 it doesn't do a search unless you type the first 3 letters 2011-04-22T07:21:10 :( 2011-04-22T07:21:42 I suppose that can give too many results sometimes 2011-04-22T07:23:21 i might as well update too 2011-04-22T07:27:07 *** contestbot has joined #aichallenge 2011-04-22T07:27:15 *** ChanServ sets mode: +o contestbot 2011-04-22T07:27:24 just in time! 2011-04-22T07:27:24 decent, now it has its own screen session 2011-04-22T07:27:31 :) 2011-04-22T07:29:30 I've been messing with screen all day... it is unbelievably annoying when it comes to changing sessionnames once a session has started 2011-04-22T07:30:48 what if you type them, and then contest bot is up, and get hit by them. 2011-04-22T07:31:04 o_O 2011-04-22T07:31:34 control characters. it's risky to send, even if contestbot is not up. 2011-04-22T07:31:40 *** aeter has joined #aichallenge 2011-04-22T07:32:21 haha... well at worst no one will be able to view the logs in their browser until amstan deleted the lines :P 2011-04-22T07:33:16 no core fusion reaction then ? 2011-04-22T07:33:40 is contestbot run near a shore ? 2011-04-22T07:34:44 again... 2011-04-22T07:34:45 i never undestood one thing, can i do cds in cron lines? 2011-04-22T07:34:57 like, how would i go adding a git commit to a cronjob 2011-04-22T07:35:14 you can use at 2011-04-22T07:35:43 instead of cron, unless you want it to execute every single time and not only once 2011-04-22T07:35:58 otherwise just make a script that does the git commit,cds etc. 2011-04-22T07:35:59 amstan: just specify the full path 2011-04-22T07:36:03 and point cron to it 2011-04-22T07:36:03 i want it to do every single time 2011-04-22T07:36:04 or that 2011-04-22T07:36:04 why do you need to cd? 2011-04-22T07:36:22 sigh: i don't think i can tell git add where the repo path is 2011-04-22T07:38:02 amstan: hmm... you are right 2011-04-22T07:38:57 what about git --git-dir=... ?? 2011-04-22T07:43:15 *** contestbot has joined #aichallenge 2011-04-22T07:43:15 *** ChanServ sets mode: +o contestbot 2011-04-22T07:43:25 woudshoo`: i guess that could work, too late, the damage has been done with a script 2011-04-22T07:50:47 *** berak has quit IRC (Read error: Connection timed out) 2011-04-22T07:50:51 sigh: 0 4 * * * /etc/alex/contestbot_cron good for you? 2011-04-22T07:51:00 4 meaninig 4am for me 2011-04-22T07:51:01 contestbot: time 2011-04-22T07:51:01 amstan: 07:51 AM, April 22, 2011 2011-04-22T07:51:17 sure, anytime is fine :) 2011-04-22T07:51:33 *** berak has joined #aichallenge 2011-04-22T07:57:34 sigh: git clone http://contestbot.hypertriangle.com/channellogs/%23aichallenge/git/ 2011-04-22T07:57:37 wiat 2011-04-22T07:57:39 noo! 2011-04-22T07:58:34 there we go 2011-04-22T07:58:37 now it has a dot too 2011-04-22T08:00:25 contestbot: time 2011-04-22T08:00:25