2011-03-30T00:00:16 amstan, I couldn 2011-03-30T00:00:20 't find it on the wiki 2011-03-30T00:00:24 could you point me to it? 2011-03-30T00:00:25 https://github.com/aichallenge/aichallenge/issues#issue/12 2011-03-30T00:00:46 nesquik: pyro is working on it... but you could join forces 2011-03-30T00:01:00 ok thanks :) 2011-03-30T00:03:59 woot! 2011-03-30T00:04:03 *** Kingpin13 has quit IRC (Ping timeout: 276 seconds) 2011-03-30T00:04:12 mcsmash: se topic 2011-03-30T00:05:08 *** KP13 is now known as Kingpin13 2011-03-30T00:05:50 mcsmash: here's a visualizer to satisfy your desires: http://home.hypertriangle.com/~alex/aichallenge/ants/visualizer/offline.html 2011-03-30T00:25:49 *** nesquik has quit IRC (Quit: Leaving) 2011-03-30T01:08:19 amstan: I kinda deprecated the offline.html now in favor of the standalone visualizer.jar. McLeopold was the first to complain about that move ^^ 2011-03-30T01:09:59 *** Apophis_ has joined #aichallenge 2011-03-30T01:13:44 *** Apophis has quit IRC (Ping timeout: 250 seconds) 2011-03-30T01:17:18 Frontier: don't make people run jars unless they can't do it any other way 2011-03-30T01:18:34 sigh: what is bad about them? 2011-03-30T01:18:36 html visualiser is so much easier, especially when I'm doing dev by sshing to a different machine 2011-03-30T01:19:26 with html, I can just serve up the webpage and access it local browser :) 2011-03-30T01:19:26 sigh: can you explain your use case a little bit? I never used the html visualizer in the last contest, only ShowGame.jar 2011-03-30T01:20:14 I worked from my laptop, but ssh into my desktop which is a much more powerful machine to actually do dev and run the thing 2011-03-30T01:20:15 sigh: If you run your own webserver, you can probably set up php as well and run the official visualizer.php 2011-03-30T01:20:26 eh? webserver? 2011-03-30T01:20:33 python -m SimpleHTTPServer 2011-03-30T01:20:42 oic ^^ 2011-03-30T01:21:15 even when I'm working on the same machine that's what I do 2011-03-30T01:21:17 now how would you get the replay into the html? 2011-03-30T01:22:20 I haven't looked at your visualiser yet... the one in the viewer directory is just hard coded :P 2011-03-30T01:23:04 what do you mean by hard coded? if you cannot change it, it doesn't work :o) 2011-03-30T01:23:45 Does it load a replay from a predefined http location? 2011-03-30T01:23:59 no, just local fike 2011-03-30T01:24:00 *file 2011-03-30T01:24:12 0.stream :P 2011-03-30T01:24:29 html cannot read local files or did I miss something? 2011-03-30T01:25:20 hmm... lemme check... it might be an artifact of the fact that I'm serving up the entire dir 2011-03-30T01:25:21 but anyway, via http that would be possible using a GET request on that same file 2011-03-30T01:25:58 i guess that is what McLeopold is doing. I saw a GET somewhere in his code. 2011-03-30T01:26:13 ah yes it uses $.get('0.stream',..) 2011-03-30T01:27:33 basically my set up is that I have `python -m SimpleHTTPServer` running on that directory, and I simply go to http://localhost:8000/viewer.html 2011-03-30T01:27:34 It would be easy to add a 'reload' button to the visualizer.jar, that fetches a new replay from your simple http server. I had that in mind anyway to avoid the VM load times 2011-03-30T01:27:48 python -m SimpleHTTPServer is starting a local (very simple) webserver so you don't have to deal with the local file restrictions 2011-03-30T01:27:57 for the html, you just refresh the browser 2011-03-30T01:28:22 same thing, yes 2011-03-30T01:28:43 you dont restart the browser, you just refresh 2011-03-30T01:30:34 even without SimpleHTTPServer you should still be able to do $.get('0.stream',...) if you are just looking at localhost 2011-03-30T01:30:49 I just have that setup because I'm not always looking up localhsot 2011-03-30T01:30:51 *localhost 2011-03-30T01:31:28 localhost doesn't exist when you don't run a server on it :p 2011-03-30T01:31:51 yeah, Frontier is referring to using file:// urls 2011-03-30T01:31:55 sorry, I meant when you are looking at file:///../viewer.html 2011-03-30T01:32:48 if you can view that, then file:///.../0.stream is just as valid 2011-03-30T01:33:08 any local file request is considered a cross site request for javascript and is not allowed to access the content of any such request 2011-03-30T01:33:14 sigh: I think browsers have 'same-origin' restrictions in place there. 2011-03-30T01:33:31 does it? ok 2011-03-30T01:33:43 file:// urls are genrally considered NOT to be same-origin 2011-03-30T01:35:14 loading http(s) urls into the visualizer works, the question is: would a simple 'f5 to refresh' be ok or are there other reasons not to run a jar? 2011-03-30T01:35:39 (considering setting up simpleHTTPServer to be at least as complicated as installing Java 6) 2011-03-30T01:36:38 hang on... viewer.html is allowed to access the local javascript files? 2011-03-30T01:36:45 * Frontier takes a note to collect use cases before starting actual work 2011-03-30T01:36:49 or whatever 2011-03-30T01:37:07 it must be 2011-03-30T01:37:08 sigh: yes, but the special XmlHTTPRequest is not 2011-03-30T01:37:10 Frontier: except they're going to need install python for the engine currently anyway right? 2011-03-30T01:37:30 in which case write the replay to a .js file which stores the json to a variable 2011-03-30T01:37:34 global variable 2011-03-30T01:37:36 at which point the SimpleHTTPServer setup is a matter of telling them the right commandline to use 2011-03-30T01:38:19 I think if you write the .js then it doesn't need SimpleHTTPServer, and works like the old planetwars one 2011-03-30T01:38:20 apt-get install sun-jre6 , come on ^^ 2011-03-30T01:38:58 so the old planetwars visualizer loaded a js file? 2011-03-30T01:39:12 yes, it's not a big thing. just certainly no easier than the other way 2011-03-30T01:39:24 sort of, it wrote the visualiser file itself 2011-03-30T01:39:47 the html file WROTE a js file? 2011-03-30T01:39:57 * Frontier is puzzled 2011-03-30T01:39:58 no, the engine which generated the replay 2011-03-30T01:40:04 generated a html file 2011-03-30T01:40:29 for example in your case the engine would output offline.html 2011-03-30T01:41:26 (or it injected it into an existing template, can't remember) 2011-03-30T01:41:59 I see. Well, I'm not sure now, if I should work on that or let the engine write the replay into a html template. 2011-03-30T01:42:27 I sounds more like an output option to me, like 'export this file as pdf' 2011-03-30T01:42:31 damnit 2011-03-30T01:42:38 my physics lecturer gives out bonus marks 2011-03-30T01:42:51 that means we need to put additional effort into assignments :\ 2011-03-30T01:43:28 Frontier: actually, in planet wars you did something like this: engine | create_visualiser 2011-03-30T01:43:53 lol, never heard of that option. where did you get this create_visualizer script? 2011-03-30T01:43:58 so engine output the replay, and create_visuliser did the messy stuff (names different of course) 2011-03-30T01:45:39 hang on 2011-03-30T01:45:54 I will find the actual code instead of making up stuff 2011-03-30T01:47:05 didn't deal with the stuff people were using client side much but all I remember from planetwars is something like "plagame.jar .... | showgame.jar" 2011-03-30T01:47:47 janzert: yeah, you could substitute showgame..jar with one which created a html version and opened it in your browser 2011-03-30T01:47:56 ahh 2011-03-30T01:48:30 https://github.com/DanielVF/Planet-Wars-Canvas-Visualizer/blob/master/visualize_localy.py 2011-03-30T01:48:34 there 2011-03-30T01:49:19 it took index.php as a template and created generated.htm 2011-03-30T01:49:47 and launched it in your browser 2011-03-30T01:50:26 (personally I commented out the webbrowser.open line and just used SimpleHTTPServer, but that wasn't a requirement) 2011-03-30T01:50:29 makes sense since that's the way it worked on the server too 2011-03-30T01:50:35 yup 2011-03-30T01:52:41 *** Kingpin13 is now known as Kingpin13|away 2011-03-30T01:52:50 Frontier: you still there? 2011-03-30T01:52:51 sigh: now we load replays via GET request, but with a different template it still works 2011-03-30T01:53:55 I'll let you handle it ^_^ 2011-03-30T01:54:06 visualizer = new Visualizer(...); 2011-03-30T01:54:07 visualizer.loadReplayData(); 2011-03-30T01:54:29 *** McLeopold has joined #aichallenge 2011-03-30T01:54:57 I don't get to wash my dishes today it seems ^^ 2011-03-30T01:55:03 :) 2011-03-30T01:55:13 Is frontier still around? 2011-03-30T01:55:19 o_O 2011-03-30T01:55:20 hi McLeopold: pls join the discussion 2011-03-30T01:55:21 no 2011-03-30T01:55:28 :P 2011-03-30T01:55:37 Am an in the discussion now? 2011-03-30T01:56:03 I was just trying to convince Frontier to allow people to run the html visualiser locally 2011-03-30T01:56:16 We were talking about sigh running the engine on a remote machine via ssh and viewing the replay over a browser 2011-03-30T01:56:20 So, just some thoughts on the visualizer. Replay files will be resources on the internet. 2011-03-30T01:56:30 correct 2011-03-30T01:56:36 An html viewer should be supported. 2011-03-30T01:56:51 My impression was that the java viewer was only an IE work around. 2011-03-30T01:57:15 mine too, but then the offline.html file disappeared :P 2011-03-30T01:57:16 from making it faster on IE emerged a standalone .replay viewer 2011-03-30T01:57:19 There is no reason to force people to use java or jar files. 2011-03-30T01:57:32 that's what I said :D 2011-03-30T01:57:34 but we already force them to use Python 2011-03-30T01:57:41 That's cool, but we need to give people the js only version. 2011-03-30T01:57:42 same logic I think 2011-03-30T01:58:09 less dependencies the better 2011-03-30T01:58:21 The engine is written as python, but is written to a spec. 2011-03-30T01:59:00 We should have a java engine for those that want it, but we shouldn't force it. 2011-03-30T01:59:09 ok, you need to help me out a bit here. what should the engine do, what should the html look like to make it easy to use. copy&paste replays seem awkward to me 2011-03-30T01:59:34 The engine is capable of producing replay and stream formats. 2011-03-30T01:59:53 Frontier: for now just do what McLeopold's visualiser does 2011-03-30T02:00:01 The visualizer should accept those formats from any resource, url, paste, local file, etc... 2011-03-30T02:00:25 no etc, that are all options ;) 2011-03-30T02:00:29 The java viewer is fine, it just needs to go in a tool package for java users. 2011-03-30T02:00:39 Someone will think of something else 2011-03-30T02:00:43 yup 2011-03-30T02:00:55 The visualizer needs text, and that is all it should care about. 2011-03-30T02:01:18 McLeopold: Hehe, I don't think it is for Java users only. Everyone can use it until other alternatives emerge 2011-03-30T02:01:22 The offline.html should just be a tool to help gather the text from various sources and pass it to the visualizer 2011-03-30T02:02:12 http://www.faqs.org/docs/artu/ 2011-03-30T02:02:17 Text is king 2011-03-30T02:02:43 what is that html 3? 2011-03-30T02:03:12 can you not read html source? 2011-03-30T02:03:28 McLeopold: I think your viewer html is what you want, I don't need to reinvent the wheel. 2011-03-30T02:04:03 the downside of viewer.html is that it requires SimpleHTTPServer 2011-03-30T02:04:11 but it's better than no HTML visualiser :P 2011-03-30T02:05:35 restore offline.html, the wheel was already made 2011-03-30T02:06:00 I only have so much time and I would rather ensure the applet runs stable and I get map scrolling and zoom implemented 2011-03-30T02:06:25 Frontier: I'll do it then, just put offline.html back and I'll get something working 2011-03-30T02:06:40 It is still in the repo history, isn't it? 2011-03-30T02:06:54 ok, fine.. I'll get it back myself :P 2011-03-30T02:07:17 should it go into the viewer folder then, or in a separate tools folder? 2011-03-30T02:08:02 why can't it live in the visualiser folder where offline.html used to be? 2011-03-30T02:08:32 People would download a lot of stuff like the java classes to get it working 2011-03-30T02:08:49 although they only need the js files and images 2011-03-30T02:08:53 eh? 2011-03-30T02:09:12 but the directory structure changed to make it cleaner to work with from ant 2011-03-30T02:09:15 oh, you are talking about the java visualiser? 2011-03-30T02:09:15 then can we pull the js out of the java package? 2011-03-30T02:09:21 there is js, java and data 2011-03-30T02:09:42 but the offline.html will look for img/... instead of data/img 2011-03-30T02:10:10 sigh: are you familiar with the visualize_locally.py from the last contest? 2011-03-30T02:10:21 could we add an option for that in playgame.py? 2011-03-30T02:10:26 I pasted it to the channel just before you got here 2011-03-30T02:10:37 McLeopold: The Java version compiles the JS, images and Java into one jar. So I think it belongs together somehow, but we can propably seperate things a bit 2011-03-30T02:10:47 McLeopold: that's kind of what I was thinking 2011-03-30T02:10:48 except maybe it launches the SimpleHTTPServer? 2011-03-30T02:11:28 I don't like the idea of it launching anything 2011-03-30T02:11:29 the ant script can pull from other directories if it needs to 2011-03-30T02:11:46 McLeopold: If you use visualize_locally.py, I could let the ant script zip up a package with the html template to use 2011-03-30T02:12:25 We should provide as many tools as possible on a central download page of ai-contest.com 2011-03-30T02:13:13 sigh: how come you don't want it to launch things? 2011-03-30T02:13:32 even if it was an option? 2011-03-30T02:13:56 the engine shouldn't launch random processes 2011-03-30T02:14:09 Not the engine, playgame 2011-03-30T02:14:41 playgame is just a testing tool, like the offline.html 2011-03-30T02:14:57 hmmm... maybe 2011-03-30T02:15:14 the engine.py is in worker, and doesn't even know anything about ants 2011-03-30T02:15:18 in which case it should just run webbrowser.open like visualise_locally.py did 2011-03-30T02:15:42 yeah, but it would have to write a special html file, which we don't want to do 2011-03-30T02:15:43 * janzert wishes ants didn't depend on stuff in worker 2011-03-30T02:16:05 janzert: the ants.py is written to an interface of sorts 2011-03-30T02:16:12 who doesn't want to write a html file? 2011-03-30T02:16:15 other than that, it is completely separate 2011-03-30T02:16:20 me 2011-03-30T02:16:21 McLeopold: what's wrong with the process that planetwars had? 2011-03-30T02:16:40 it's just going to make it harder to package it up 2011-03-30T02:16:43 nothing, I'm just striving for elegance 2011-03-30T02:17:32 janzert: yes, but easier when we have multiple games going 2011-03-30T02:17:40 launching SimpleHTTPServer in the foreground, forcing people to navigate to http://localhost:8000/offline.html 2011-03-30T02:17:43 is not elegent 2011-03-30T02:17:47 * janzert doesn't believe in multiple games ;) 2011-03-30T02:18:21 *sigh* another bike shed :p no offense, i start of a few myself 2011-03-30T02:18:22 or the 'interface' could just be a convention all the games conformed to 2011-03-30T02:18:31 sigh: okay, maybe, but it's more elegant that writing a temp file to the same location everytime 2011-03-30T02:18:50 the user doesn't care 2011-03-30T02:18:54 McLeopold: think PDF exporter ;) 2011-03-30T02:18:55 janzert: that's the idea 2011-03-30T02:19:20 well with that idea the game doesn't import anything from worker :) 2011-03-30T02:19:31 all the user knows is that there shell is taken out of action and that they have to manually navigate to some url 2011-03-30T02:19:32 sigh: ultimately, it would be nice for a user testing locally to treat his system as a separate tcp server 2011-03-30T02:19:59 which is why having playgame.py manage replay files and an html interface for browsing them seems elegant 2011-03-30T02:20:24 that's fine if you can set it up so that someone who doesn't know python, html, or anything about the web can run it 2011-03-30T02:20:58 McLeopold: either way, it is not usable 2011-03-30T02:20:59 McLeopold: In that case we are close to a full ai-contest server installation 2011-03-30T02:21:15 http://www.boingboing.net/2011/03/29/machete-shooting-rif.html completely off topic but...sweet :) 2011-03-30T02:21:16 what's not usable? 2011-03-30T02:22:30 sigh: I don't think I followed your train of thought... 2011-03-30T02:23:49 I have a feeling browsing replays is a nice feature, but you are making it more complicated than neccesary for most users. Look at how simple the starter packs are 2011-03-30T02:24:29 Ideally you just unzip and run a command (or pipe the output of the engine into it) 2011-03-30T02:24:44 I just don't want to make it more complicated for those that can cope with good tools. Obviously we want sane defaults, but that doesn't mean we should limit ourselves. 2011-03-30T02:25:06 McLeopold: the only thing that you have a problem with is writing a html file? 2011-03-30T02:25:45 Yeah, because I think the replay files should remain text so that you can easily pass them around to different tools 2011-03-30T02:26:02 McLeopold: again, think PDF 2011-03-30T02:26:17 it is an export option, not an format to be parsed 2011-03-30T02:26:23 the replays were plain text in planetwars 2011-03-30T02:26:26 So, if the replay files was written somewhere for storage, *and* a temp html files was written, that would be ok. I just think we could do slightly better. 2011-03-30T02:26:38 it was visualise_locally which created html 2011-03-30T02:27:21 either way, if you insist I would rather visualise_locally.py so SimpleHTTPServer stuff, I still think it makes it more effort on the user's part than planetwars was 2011-03-30T02:27:26 McLeopold: for me that option seems ok. You tell visualize_locally to keep a copy of the original data somehwere and let it write the html 2011-03-30T02:27:59 Okay, a http server could be separated from playgame, but both could assume some type of local storage structure. 2011-03-30T02:28:18 also automatically running SimpleHTTPServer hijacks the port, and assumes that the user is not doing anything else with it 2011-03-30T02:28:29 *** Stocha has joined #aichallenge 2011-03-30T02:28:36 add a port command-line option 2011-03-30T02:28:38 sigh: stop, I've changed my mind already :P 2011-03-30T02:29:07 no, now I'm arguing tht visualise_locally shouldn't use SimpleHTTPServer either :P 2011-03-30T02:29:10 the official server will have replays in ROOT/games/... 2011-03-30T02:29:29 sigh: yes, I agree, make a separate tool 2011-03-30T02:29:39 ok :) 2011-03-30T02:29:41 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-03-30T02:29:58 replay_server.py 2011-03-30T02:30:03 yup 2011-03-30T02:30:34 ideally you would run that in the background, and not start it for every new game 2011-03-30T02:30:37 which would need the js stuff and some of the offline.html stuff 2011-03-30T02:30:42 yeah 2011-03-30T02:30:52 yes, that is actually elegent :) 2011-03-30T02:31:01 *elegant 2011-03-30T02:31:06 \o/ I knew we could do it 2011-03-30T02:31:14 :D 2011-03-30T02:31:32 okay, gotta go to bed now 2011-03-30T02:31:43 goofd night 2011-03-30T02:31:44 oh, did you see my map generator? 2011-03-30T02:31:59 the labyrinth ? 2011-03-30T02:32:04 yeah 2011-03-30T02:32:12 McLeopold: amazing voronoi diagrams 2011-03-30T02:32:15 I'll get it symmetric tomorrow 2011-03-30T02:32:42 shame :p 2011-03-30T02:33:03 *** McLeopold has left #aichallenge 2011-03-30T02:33:05 you will give the spec of you map generator also right ? Is it easy to implement ? 2011-03-30T02:33:06 sigh: Do you agree with me, that we should package this html visualizer for download? 2011-03-30T02:34:14 We can add a makefile in the ants directory that zips up your Python scripts and the stuff from the visualizer directory (html template, js, images) 2011-03-30T02:36:45 *** sigh has quit IRC (Ping timeout: 246 seconds) 2011-03-30T02:37:37 *** Stocha has quit IRC (Ping timeout: 252 seconds) 2011-03-30T02:54:55 *** amstan has quit IRC (Remote host closed the connection) 2011-03-30T02:57:17 *** Eruonen has joined #aichallenge 2011-03-30T03:04:00 *** smellyhippy has quit IRC (Ping timeout: 264 seconds) 2011-03-30T03:07:58 *** phirenz has quit IRC (Ping timeout: 248 seconds) 2011-03-30T03:08:11 *** smellyhippy has joined #aichallenge 2011-03-30T03:08:11 *** smellyhippy has joined #aichallenge 2011-03-30T03:10:06 *** phirenz has joined #aichallenge 2011-03-30T03:20:05 *** Kingpin13|away is now known as Kingpin13 2011-03-30T03:27:27 *** SubmittedDenied has joined #aichallenge 2011-03-30T03:27:34 Hey all 2011-03-30T03:27:44 *** aerique has joined #aichallenge 2011-03-30T03:27:45 *** phirenz is now known as phire 2011-03-30T03:31:41 *** sigh has joined #aichallenge 2011-03-30T03:42:22 Frontier: I don't mind as long as it is easy to use 2011-03-30T03:44:47 Accessibility is what is really great about this contest. 2011-03-30T03:45:09 *** FireFly has joined #aichallenge 2011-03-30T03:45:24 I still have to implement the color blind mode 2011-03-30T03:47:18 I'll overlay the player's letter A, B, C... over the ants 2011-03-30T03:48:24 Just a side note only loosely related to the ease of use of the tools :) 2011-03-30T03:48:42 is there a limit to the size of an ants game? 2011-03-30T03:52:34 aichallenge: sigh fast-vision * rf63f0b1 / ants/ants.py : Implemented incremental updates for vision - http://bit.ly/h0A519 2011-03-30T03:55:10 *** delt0r___ has joined #aichallenge 2011-03-30T03:56:48 sigh: I should also incrementally load the replay or calculate vision. 2011-03-30T03:57:11 :) 2011-03-30T03:57:26 *** delt0r_ has quit IRC (Ping timeout: 276 seconds) 2011-03-30T03:57:36 It was pretty naive to precalculate every players vision for every turn. Most of the times that data wont be used 2011-03-30T03:58:29 *** woudshoo has joined #aichallenge 2011-03-30T03:59:50 aichallenge: sigh fast-vision * r9640155 / ants/ants.py : Cleanup and comment new vision code - http://bit.ly/i5TaOe 2011-03-30T04:00:06 Frontier: feel free to look at what the engine does in the fast-vision branch 2011-03-30T04:00:15 janzert: you there? 2011-03-30T04:00:59 Is overlaying the letter on the ant the best idea? 2011-03-30T04:01:03 sort of 2011-03-30T04:01:29 janzert: check out the ants.py in the fast-vision branch :D 2011-03-30T04:02:03 it's the incremental update idea? 2011-03-30T04:02:21 in a game it took down time for vision from 20ms per turn to 0.01ms/turn 2011-03-30T04:02:47 hmm, I guess it may be worth it then 2011-03-30T04:02:59 what is the total time per turn at that point? 2011-03-30T04:03:07 back in a bit 2011-03-30T04:03:08 sorry, 0.37ms/turn 2011-03-30T04:05:44 *** Stocha has joined #aichallenge 2011-03-30T04:06:16 hard to tell total time per turn, there's no function for it 2011-03-30T04:07:37 and the game doesn't get much shorter because 2 seconds is spent sleeping waiting for bots to reply :) 2011-03-30T04:09:18 hi all, just curious, but what is the rationale for showing water only the first time it is seen? It does not seem to add anything to the gameplay. 2011-03-30T04:09:51 woudshoo: reduce the amount of data that needs to be transfered 2011-03-30T04:10:16 that's really the only issue there 2011-03-30T04:11:10 sigh: I see, 5000 water * 8 bytes can add up 2011-03-30T04:11:16 yup 2011-03-30T04:11:41 and it is trivial for a bot to keep track of it 2011-03-30T04:12:51 Yes, it is trivial to keep track, I just started playing with implementing a bot and that is definitely not the hard part. 2011-03-30T04:13:37 Also, I just updated yesterday, and the offline.html viewer is gone :-( It looked cool eventhough the playgame did not generate output that could be parsed by offline.html 2011-03-30T04:15:12 woudshoo: it will be back, don't worry :) 2011-03-30T04:17:06 for now, you can just: git checkout b08cc7 ants/visualizer/offline.html 2011-03-30T04:18:23 sigh: thank you, now I just should write a script to wrap the 0.replay in the format offline.html expects. (Or fix/change playgame.py, but I don't know any python) 2011-03-30T04:19:09 woudshoo: yeah, we'll get around to implementing that, everything is still in dev 2011-03-30T04:19:34 woudshoo: python code teaches itself 2011-03-30T04:19:42 sigh: I know, I just couldn't wait to start playing around. 2011-03-30T04:21:01 :) 2011-03-30T04:22:20 antimatroid: Well, I have heard good things about it, but in my copiuous free time *cough* *cough* I haven't found time to look at it. 2011-03-30T04:23:43 BTW, who should I contact about the common lisp interface/infrastructure? 2011-03-30T04:24:40 *** SubmittedDenied has quit IRC (Quit: Page closed) 2011-03-30T04:25:43 sigh: it is a nice impressive speed up and should probably go in, I just am getting the feeling that it is probably at the point of optimizing something that isn't the bottleneck 2011-03-30T04:26:33 janzert: it was the bottleneck before... but I only did it to make it harder for McLeopold to beat it 2011-03-30T04:26:42 :) 2011-03-30T04:27:58 woudshoo: me, but i'm dumbing it down as per McLeopold's instructions 2011-03-30T04:28:31 janzert: and in case you missed me saying it, it was 0.37ms/turn not 0.01ms :P 2011-03-30T04:28:47 heh 2011-03-30T04:29:00 still 50 times faster :) 2011-03-30T04:29:50 *** Stocha has quit IRC (Ping timeout: 252 seconds) 2011-03-30T04:29:53 aerique: Thanks for the work, curious what are you dumbing down? 2011-03-30T04:30:21 woudshoo: https://github.com/aichallenge/aichallenge/issues/3#comment_934055 2011-03-30T04:30:56 aerique: Just a question, on MacOSX (my laptop) I regularly have that my bot times out during the setup face because loading the sbcl core/image takes too long with a cold cache. Is that a dange in the real competition? 2011-03-30T04:31:53 woudshoo: no, we always use compiled bots on the game server (exactly because of the problem you mentioned) 2011-03-30T04:31:54 I think the timer is set to 3 seconds for the warm up phase. 2011-03-30T04:32:05 woudshoo: (at least for CL that is) 2011-03-30T04:33:07 aerique: I use locally the compiled bot. The resulting executable is about 40MB and if I run it the first time, it can take more than 3 seconds. (Somehow my Mac is sometimes really slow with disk access.) But if the experience of planetwars was that it is not a problem I will not worry about it. 2011-03-30T04:33:31 woudshoo: ah ok. it has indeed not been a problem with planet wars 2011-03-30T04:33:39 woudshoo: for local you can extend the time using the --loadtime option to playgame.py 2011-03-30T04:35:18 woudshoo: you could also try using the proxy-bot so you don't have to recompile every time (but i haven't got OS X so it isn't tested there) 2011-03-30T04:35:33 *** Naktibalda has joined #aichallenge 2011-03-30T04:35:52 woudshoo: mega1 used it as well but i think he's on linux as well 2011-03-30T04:36:36 aerique: I was just trying the proxy bot this morning (in the train), but I did not use your starter pack and started from scratch, so I was just trying to get sockets working. 2011-03-30T04:39:16 woudshoo: you're from Leiden? (i'm in The Hague) 2011-03-30T04:41:19 aerique: Yes, working in Leiden (should get back to it, my boss doesn't pay me to do this) living in Rijswijk 2011-03-30T04:42:55 woudshoo: cool, i live in Delft.. anyway, back to work for me as well 2011-03-30T04:43:09 sigh: Thank you for the --loadtime option 2011-03-30T04:47:15 *** Stocha has joined #aichallenge 2011-03-30T04:47:48 just put the rational for not sending already seen water into the spec; So not all 5000 participants inquires :p 2011-03-30T04:51:16 Stocha: were you the color blind person? 2011-03-30T04:52:26 Stocha: done 2011-03-30T05:00:36 woudshoo: There is now a really cool standalone replay viewer. Just install 'ant' go into the visualizer directory and type 'ant run-visualizer'. The offline.html is deprecated. It will be replaced by scriptable Python solution. 2011-03-30T05:01:26 So in the long run you can use the Java visualizer or the Python script which will generate a html page embedding the replay 2011-03-30T05:03:12 *** Stocha has quit IRC (Ping timeout: 252 seconds) 2011-03-30T05:07:56 *** needsch has joined #aichallenge 2011-03-30T05:16:57 Frontier: I tried it, but the compile fails on MacOSX 2011-03-30T05:19:19 Frontier: It fails with: compile-rhino:...[javac] javac: invalid target release: 1.6 2011-03-30T05:25:01 Frontier, I'm getting a Build error with the visualizer 2011-03-30T05:25:28 "package sun.plugin.javascript does not exist" 2011-03-30T05:25:35 actually, that looks easy to fix 2011-03-30T05:27:23 or not 2011-03-30T05:32:17 *** woudshoo` has joined #aichallenge 2011-03-30T05:32:32 http://pastie.org/1734442 2011-03-30T05:35:52 *** boegel has joined #aichallenge 2011-03-30T05:36:08 *** woudshoo has quit IRC (Ping timeout: 276 seconds) 2011-03-30T05:39:20 aichallenge: aerique epsilon * r4ee3264 / (28 files in 7 dirs): 2011-03-30T05:39:21 aichallenge: Trim starter bot as per McLeopold's instructions. 2011-03-30T05:39:21 aichallenge: See: http://github.com/aichallenge/aichallenge/issues/3#comment_934055 - http://bit.ly/hsPUTW 2011-03-30T05:42:28 @later tell McLeopold, made vision about 50x faster - see fast-vision branch: https://github.com/aichallenge/aichallenge/blob/fast-vision/ants/ants.py#L215 2011-03-30T05:42:28 sigh: As you wish. 2011-03-30T05:42:52 hm... that comma will screw things up wont it 2011-03-30T05:43:28 @later tell sigh, this is a test 2011-03-30T05:43:28 sigh: I think that worked... 2011-03-30T05:43:29 probably 2011-03-30T05:43:41 no, that worked 2011-03-30T05:43:59 depends on how smart the parsing algrothem is 2011-03-30T05:44:20 seems it can handle that 2011-03-30T05:44:25 *** Kingpin13 has quit IRC (Quit: You're a kitty!) 2011-03-30T05:51:33 aichallenge: aerique epsilon * rd688d79 / ants/dist/starter_bots/common_lisp/test/run-tests.lisp : Update CL unit tests. - http://bit.ly/gXT1DY 2011-03-30T06:01:45 * Naktibalda has to copy these unit tests :) 2011-03-30T06:07:12 *** needsch has quit IRC (Quit: Leaving.) 2011-03-30T06:19:36 They are in need of improvement :-| 2011-03-30T06:26:52 *** p4p4p4 has joined #aichallenge 2011-03-30T06:47:41 *** Stocha has joined #aichallenge 2011-03-30T06:48:47 @seen Frontier 2011-03-30T06:48:47 Stocha: Frontier was last seen in #aichallenge 1 hour, 47 minutes, and 21 seconds ago: So in the long run you can use the Java visualizer or the Python script which will generate a html page embedding the replay 2011-03-30T06:49:02 @later tell Frontier i am colorblind yes 2011-03-30T06:49:02 Stocha: I think that worked... 2011-03-30T06:49:31 *** Stocha has quit IRC (Client Quit) 2011-03-30T07:03:16 *** needsch has joined #aichallenge 2011-03-30T07:03:54 *** needsch has quit IRC (Client Quit) 2011-03-30T07:12:22 Frontier: ping 2011-03-30T07:13:09 @later tell Frontier how do I start the visualiser off in expanded mode 2011-03-30T07:13:09 sigh: Yes master! 2011-03-30T07:20:18 @later tell Frontier, nevermind figured it out, but now I need to know why you have that annoying "Accept the next dialog to have colorized button graphics." message 2011-03-30T07:20:18 sigh: Yes master! 2011-03-30T07:20:50 sigh: start it in expanded mode? expand it and then click on save... theoretically 2011-03-30T07:21:04 Frontier: nah, I figured it out 2011-03-30T07:21:13 anyway... why is that security message there? 2011-03-30T07:21:16 can I take it out? 2011-03-30T07:21:31 it's in ImageManager.js 2011-03-30T07:21:39 You can tell Firefox to always accept. 2011-03-30T07:22:16 how? 2011-03-30T07:22:16 If you ignore it the buttons for fog of war will all be the same color. I did not put that there without reason. A shame of you to think so ^^ 2011-03-30T07:22:28 Isn't there a checkbox? 2011-03-30T07:22:30 *** Stocha has joined #aichallenge 2011-03-30T07:22:34 I just get the alert 2011-03-30T07:22:38 huh? 2011-03-30T07:22:45 but you use firefox? 2011-03-30T07:22:48 and it just says "OK" 2011-03-30T07:22:49 yes 2011-03-30T07:22:50 which version? 2011-03-30T07:23:11 4? 2011-03-30T07:23:14 Mozilla/5.0 (Macintosh; U; Intel Mac OS X 10.4; en-GB; rv:1.9.2.16) Gecko/20110319 Firefox/3.6.16 2011-03-30T07:23:21 3.6.16 2011-03-30T07:23:58 I'm now on windows, where I don't have the same arsenal of browsers installed, but that sure looks strange to me. 2011-03-30T07:24:09