2011-05-10T00:00:01 where does your bot dir go? 2011-05-10T00:00:23 the uploaded file shouldn't be in the compiled dir, just the uploaded one 2011-05-10T00:00:44 well the worker will be getting it from the manager 2011-05-10T00:00:55 and shouldn't the bot dir be outside the submission dir? if you are worried about clobbering stuff? 2011-05-10T00:01:12 no 2011-05-10T00:01:25 the upload get's unpacked into the bot dir 2011-05-10T00:01:34 ooooooohhh 2011-05-10T00:01:43 the only thing that then writes/makes changes there is the compilation process and the bot itself 2011-05-10T00:03:11 does that clear it up? 2011-05-10T00:03:57 thinking 2011-05-10T00:04:09 the worker needs to download an entry file... 2011-05-10T00:04:17 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-05-10T00:04:25 unless you want to try and share the upload dir for a combined setup 2011-05-10T00:05:26 where does the temp worker dir go? That's what I was using downloaded for 2011-05-10T00:05:29 I think I'd rather not, just because I'd like combined setups not to work too differently from separate machines 2011-05-10T00:05:50 the temp dir is created directly in the compiled/ dir 2011-05-10T00:06:55 those are really easy changes to worker 2011-05-10T00:07:42 there are python functions to make sure you get a unique temp directory too without a race condition 2011-05-10T00:07:52 looking that up now since I can't remember the name offhand ;) 2011-05-10T00:08:03 k 2011-05-10T00:08:10 *** Winterfell has quit IRC (Quit: Page closed) 2011-05-10T00:09:56 basically want something like, import tempfile; tempfile.mkdtemp(dir="/home/contest/compiled") 2011-05-10T00:11:00 which will end up creating a directory something like /home/contest/compiled/tmp1c35b 2011-05-10T00:11:17 *** Mugen has joined #aichallenge 2011-05-10T00:11:27 janzert: is there an easy way to run the setup script without messing with install and jail users, so I can test dir changes? 2011-05-10T00:12:14 worker_setup? 2011-05-10T00:13:38 I think it should be safe to run it multiple times and it will skip the parts it thinks are already done 2011-05-10T00:14:06 ok 2011-05-10T00:14:07 that isn't really tested much though, so I wouldn't be too surprised if it runs into a problem 2011-05-10T00:14:13 hmm 2011-05-10T00:14:24 janzert: it'll skip installing /srv/chroot if it sees it's present, it'll jump directly to jail users setup 2011-05-10T00:14:27 but I would guess worst case is that it will either error out or add duplicate lines to some files 2011-05-10T00:15:06 *** Mugen has quit IRC (Client Quit) 2011-05-10T00:15:10 hmm, yeah there isn't any test to check if a jailuser already exists :( 2011-05-10T00:16:23 add an if file_contains("/etc/passwd", username): return check to the top of create_jail_user should fix that 2011-05-10T00:22:52 *** bhasker has joined #aichallenge 2011-05-10T00:30:25 okay, just need to get the bot dir in the submission dir working... 2011-05-10T00:30:45 do you really want to keep the uploaded zip on the worker? 2011-05-10T00:30:49 janzert: ^^ 2011-05-10T00:33:13 I don't think I really care one way or the other at this point 2011-05-10T00:33:33 deleting it after unpacking would be fine but if you don't want to add that I'm fine with that too 2011-05-10T00:33:58 mostly I just wanted to show that it should be outside the bot directory 2011-05-10T00:34:12 ok 2011-05-10T00:35:41 the worker temporary files you're talking about games/ are game replays it is sending to the manager? 2011-05-10T00:35:55 yes 2011-05-10T00:36:15 wait... 2011-05-10T00:36:18 ok, just concerned about what we put in there since we are planning on exposing it publically 2011-05-10T00:36:20 but it couldn't be shared 2011-05-10T00:36:31 yeah 2011-05-10T00:36:52 why can't playgame just return it directly to worker.py and it send it directly without ever hitting a disk? 2011-05-10T00:36:53 I want a place for the worker to store results when it can't communicate with the server 2011-05-10T00:37:07 but it'll only ever have one result 2011-05-10T00:37:11 in that state 2011-05-10T00:37:22 so just keep it in memory? 2011-05-10T00:38:03 yeah, ok 2011-05-10T00:38:58 btw, what is the game directory structure? 2011-05-10T00:39:24 since it's going to have literally a million files or so :} 2011-05-10T00:39:48 maybe more 2011-05-10T00:41:44 janzert: with the mod 1000 thing 2011-05-10T00:42:44 hierarchical unlike the submissions though? 2011-05-10T00:43:41 i.e. games/// (and actually steps of 100 might be better) 2011-05-10T00:44:02 and submissions are division 2011-05-10T00:44:05 I do want them to be permalinks 2011-05-10T00:44:48 why is 100 better? 2011-05-10T00:46:12 I guess even % 1000 is still a maximum of 1998(?) files in a single directory 2011-05-10T00:46:29 *** amstan has quit IRC (Remote host closed the connection) 2011-05-10T00:46:30 s/files/files and subdirectories/ 2011-05-10T00:47:55 just be some long ls output :} 2011-05-10T00:48:12 true 2011-05-10T00:48:20 but I don't think we care too much 2011-05-10T00:48:51 it would be nice to get a url rewrite to make the links look nice 2011-05-10T00:49:58 I would think that should be fairly easy to add to the apache configuration 2011-05-10T00:50:04 yeah 2011-05-10T00:50:17 so, double directories for the games, single for the submissions 2011-05-10T00:50:36 every thing is at the same level, so /home/contest/games/0/0/0.replay 2011-05-10T00:51:17 ahh, I was picturing expanding levels as the gameid increased 2011-05-10T00:51:34 hence the ~2x number of entries per directory 2011-05-10T00:51:46 which is why I clarified :) 2011-05-10T00:52:07 but yeah I doubt we'll ever get to a third level at steps of 1000x 2011-05-10T00:53:02 so fixed two levels deep is almost certainly easier to work with 2011-05-10T00:53:35 it would be /home/contest/games/10000/0/10000000000.replay 2011-05-10T00:53:51 and if it's fixed you might as well use division instead of mod 2011-05-10T00:54:08 I am 2011-05-10T00:56:15 *** GeorgeSebastian has joined #aichallenge 2011-05-10T00:56:15 *** GeorgeSebastian has joined #aichallenge 2011-05-10T00:58:27 *** delt0r___ has joined #aichallenge 2011-05-10T00:59:39 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-05-10T01:12:18 *** GeorgeSebastian has quit IRC (Ping timeout: 240 seconds) 2011-05-10T01:18:56 *** Apophis has joined #aichallenge 2011-05-10T01:21:19 *** nann has quit IRC (Ping timeout: 240 seconds) 2011-05-10T01:23:23 janzert: I'm thinking that the compile stuff in manager can go away 2011-05-10T01:23:37 yes 2011-05-10T01:24:42 then, we don't need a "root_path" var in any of the server_setup files 2011-05-10T01:24:47 *** bhasker has quit IRC (Quit: bhasker) 2011-05-10T01:25:19 the sandbox needs it on the worker 2011-05-10T01:25:40 but, it has a repo path, so we could just make that absolute 2011-05-10T01:25:50 but yeah, looks like it could go away on the manager 2011-05-10T01:26:28 repo_path and root_path are only used once in sandbox, and together 2011-05-10T01:26:38 yeah, just looking at it 2011-05-10T01:27:09 yes I suppose making all the paths absolute and getting rid of root path would work fine 2011-05-10T01:27:16 probably a little better even 2011-05-10T01:28:52 should I delete bayeselo stuff? 2011-05-10T01:29:28 I'm going to move it into a bayeselo folder for now 2011-05-10T01:29:44 no don't move it get rid of it if it's been fully replaced 2011-05-10T01:29:59 only leave old stuff it is something that we still need to write a replacement for 2011-05-10T01:30:30 just so we can easily look at how it used to be done when writing the replacement 2011-05-10T01:30:35 but, for future contest, we might use it 2011-05-10T01:30:39 and remind us that we still need to do it :) 2011-05-10T01:30:57 if we ever do another 1v1 game 2011-05-10T01:30:59 at that point someone can dig it out of the history 2011-05-10T01:31:00 *** mcstar has joined #aichallenge 2011-05-10T01:31:03 k 2011-05-10T01:31:20 but I think the structure has changed so much that it will be of little use anyway 2011-05-10T01:31:39 can gmail.py be removed from manager? 2011-05-10T01:32:02 and right now even if we went back to a 1v1 game I'd rather stick with an incremental rating system like trueskill just for the performance benefits 2011-05-10T01:32:37 hmm, I can't think of any time the manager itself needs to send an email 2011-05-10T01:32:56 the worker sends the emails on submission compilation/testing failure right? 2011-05-10T01:33:03 I've never even looked at the emails we send 2011-05-10T01:33:07 and php sends it on upload failures 2011-05-10T01:34:01 how about successful submission emails, I suppose the worker sends those at well? 2011-05-10T01:34:33 account confirmation and submission success/failure emails are the only ones I can think of offhand 2011-05-10T01:35:46 right now, the worker doesn't do emails 2011-05-10T01:36:14 I've never setup a test smtp/sendmail server to look at it 2011-05-10T01:36:52 yeah, we'll need to get that working one way or the other at some point 2011-05-10T01:36:55 looks like I just import gmail and call one function 2011-05-10T01:37:19 if you haven't noticed the gmail module is really misnamed 2011-05-10T01:37:37 *** dr-` has joined #aichallenge 2011-05-10T01:37:40 yeah, I was just thinking that :) 2011-05-10T01:37:42 gmail limits to sending a few hundred emails per day and that wasn't enough for the contest 2011-05-10T01:37:52 hard coded to smtp.gmail.com:587 2011-05-10T01:38:18 where? 2011-05-10T01:38:22 does amazon have a relay we can use? 2011-05-10T01:38:25 it should be pulling it out of server_info 2011-05-10T01:38:51 they introduced a "simple mail service" a few months ago with 2000 free messages sent per day 2011-05-10T01:39:01 but I haven't looked into the details yet 2011-05-10T01:39:31 what if we just didn't do email? 2011-05-10T01:39:35 my understanding is we can configure the server so our code can still do a straight smtp send with the local server 2011-05-10T01:39:46 it'd be a lot easier :) 2011-05-10T01:39:58 I'm not sure how much of an uproar it would create though 2011-05-10T01:40:11 also we probably want to keep the account confirmation emails at least? 2011-05-10T01:40:19 why? 2011-05-10T01:40:28 that's only used to confirm you own the email address 2011-05-10T01:40:49 if we don't send emails, we don't need to confirm anything 2011-05-10T01:41:07 hmm, and don't collect emails at all 2011-05-10T01:41:17 but then we'll have no way to contact participants 2011-05-10T01:41:26 so, the user would just need a submission status page 2011-05-10T01:41:32 I guess bring it up with amstan 2011-05-10T01:41:43 I'm not sure what we need to contact them for. 2011-05-10T01:42:22 they won, they are killing our servers and we're throwing the present submission out, etc. ;) 2011-05-10T01:42:25 but yeah 2011-05-10T01:42:40 submission status page 2011-05-10T01:42:45 I'm not too opposed to doing away with it 2011-05-10T01:42:59 okay, if amstan agrees, it would be an easier setup 2011-05-10T01:43:15 actually I just thought of something, I have a feeling j3camero will be opposed to dropping it 2011-05-10T01:43:16 hmm... 2011-05-10T01:43:22 lost password 2011-05-10T01:43:32 yeah, that is another problem 2011-05-10T01:44:23 the google sponsorship comes from their headhunters and j3camero usually sends an email out to the top competitors towards the end of the contest asking if they'd like to be contacted by a recruiter 2011-05-10T01:44:52 so without an email address google might not be as enthused about the sponsorship 2011-05-10T01:45:06 I guess I didn't make the cut last time :p 2011-05-10T01:45:23 heh 2011-05-10T01:45:42 I'm not actually sure he did it last time 2011-05-10T01:45:58 I did get one with the tron contest 2011-05-10T01:46:29 I guess it'd be something we'd need to find out from him 2011-05-10T01:46:48 okay, well, at least for lost password issues, email is good, so let's keep it 2011-05-10T01:46:56 but even without that password reset does seem like a good reason to keep it as well 2011-05-10T01:47:05 * janzert nods 2011-05-10T01:47:33 I only submitted twice last contest, which is why I don't remember getting too many emails 2011-05-10T01:48:44 submissions_path is going away in the server_info files, to be replaced with upload_dir and compiled_dir 2011-05-10T01:48:51 err *_path 2011-05-10T01:49:12 sounds fine 2011-05-10T01:57:30 janzert: do you care where log files go 2011-05-10T01:58:12 a consolidated log directory would be nice, but it's fine if that doesn't happen 2011-05-10T02:18:29 janzert: I'm thinking we don't want MultiViews for the apache setup 2011-05-10T02:23:27 the current apache config is a barely modified default configuration, there is pretty certainly a lot that should change 2011-05-10T02:23:37 feel free to work on it 2011-05-10T02:25:10 i got the email for tron 2011-05-10T02:25:24 McLeopold: we weren't in the top 50 even towards the end though? 2011-05-10T02:25:29 i'm not sure he sent it out for pw either 2011-05-10T02:26:48 i look forward to the day we have a server up and running again for tron 2011-05-10T02:26:51 it was such a sweet ai game 2011-05-10T02:27:06 janzert: i hope you'll dust your old bot off for that :P 2011-05-10T02:27:14 i would write a new bot for that 2011-05-10T02:30:38 *** FireFly has joined #aichallenge 2011-05-10T02:39:16 *** Mathnerd314 has quit IRC (Ping timeout: 240 seconds) 2011-05-10T02:39:41 *** Ice_Harley has joined #aichallenge 2011-05-10T02:40:06 heh, well the source for my tron bot is out there for anyone that wanted to try it :) 2011-05-10T02:40:31 same for a lot of people 2011-05-10T02:40:41 i guess that's one issue with having old contests up 2011-05-10T02:40:47 people might be reluctant to release their code 2011-05-10T02:41:27 i'm not really sure people get much out of the code being released, but it's nice that the top ones are usually available for anyone who wants a look 2011-05-10T02:44:29 Hi, what with "r" lines? 2011-05-10T02:44:39 a mistake 2011-05-10T02:44:45 ignore it 2011-05-10T02:45:09 mistake, that will be fixed? 2011-05-10T02:45:18 yes 2011-05-10T02:45:31 it was an idea, but we decided it added to much complexity 2011-05-10T02:45:42 to the spec 2011-05-10T02:46:09 i am slow to fixing the c++ bot 2011-05-10T02:46:17 sorry, they are going to be removed 2011-05-10T02:47:10 it's a pity. I think "r" lines are useful 2011-05-10T02:47:42 *** JaycobColeman has joined #aichallenge 2011-05-10T02:49:38 they are 2011-05-10T02:50:06 but if you can do a good bot, you can calculate it yourself by calculating your vision 2011-05-10T02:52:33 I will, but i'll spend part of MY second on it 2011-05-10T02:52:56 :) 2011-05-10T02:52:58 yep 2011-05-10T02:58:44 Is anyone here doing heritage health prize? 2011-05-10T02:59:54 no, but i was mistaken as the guy running it :P 2011-05-10T03:00:09 haha, that's nice 2011-05-10T03:00:46 a1k0n: are you doing it? I thought you brought it up the other day 2011-05-10T03:12:06 *** antimatroid has quit IRC (Read error: Connection reset by peer) 2011-05-10T03:12:15 *** antimatroid has joined #aichallenge 2011-05-10T03:38:38 *** aerique has joined #aichallenge 2011-05-10T03:51:06 *** berak has joined #aichallenge 2011-05-10T03:53:42 *** ltriant has quit IRC (Quit: Computer has gone to sleep) 2011-05-10T04:04:58 *** FireFly has quit IRC (Quit: swatted to death) 2011-05-10T04:13:40 *** Naktibalda has joined #aichallenge 2011-05-10T04:23:08 *** JaycobColeman has quit IRC (Quit: JaycobColeman) 2011-05-10T05:34:26 *** Kingpin13 has joined #aichallenge 2011-05-10T05:34:29 *** Kingpin13 has joined #aichallenge 2011-05-10T05:38:07 *** sigh has joined #aichallenge 2011-05-10T05:44:25 *** okayzed has quit IRC (Ping timeout: 240 seconds) 2011-05-10T05:45:27 *** okayzed has joined #aichallenge 2011-05-10T05:57:45 *** MuTa-ofd00m has joined #aichallenge 2011-05-10T06:00:18 *** kaemo has joined #aichallenge 2011-05-10T06:26:38 *** needsch has joined #aichallenge 2011-05-10T06:31:06 *** MuTa-ofd00m has quit IRC (Quit: Bye) 2011-05-10T06:56:00 *** berak has quit IRC (Ping timeout: 240 seconds) 2011-05-10T07:53:17 *** _0x47 <_0x47!~slash@p5B25FD2F.dip.t-dialin.net> has joined #aichallenge 2011-05-10T07:54:37 *** dr- has quit IRC (Read error: Operation timed out) 2011-05-10T08:08:53 *** amstan has joined #aichallenge 2011-05-10T08:08:53 *** ChanServ sets mode: +o amstan 2011-05-10T08:20:39 *** boegel has joined #aichallenge 2011-05-10T08:29:25 *** amstan has quit IRC (Ping timeout: 260 seconds) 2011-05-10T08:53:09 *** FireFly has joined #aichallenge 2011-05-10T08:58:57 *** delt0r_ has joined #aichallenge 2011-05-10T08:59:56 *** sigh has quit IRC (Remote host closed the connection) 2011-05-10T09:00:29 *** delt0r___ has quit IRC (Ping timeout: 252 seconds) 2011-05-10T09:03:24 *** Blargs has joined #aichallenge 2011-05-10T09:21:29 *** berak has joined #aichallenge 2011-05-10T09:32:17 *** amstan has joined #aichallenge 2011-05-10T09:32:17 *** ChanServ sets mode: +o amstan 2011-05-10T09:39:54 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-05-10T09:43:33 *** _0x47 <_0x47!~slash@p5B25FD2F.dip.t-dialin.net> has quit IRC (Read error: Operation timed out) 2011-05-10T09:44:32 *** _0x47 <_0x47!~slash@p3EE0B3EF.dip.t-dialin.net> has joined #aichallenge 2011-05-10T10:01:24 *** amstan has joined #aichallenge 2011-05-10T10:01:24 *** ChanServ sets mode: +o amstan 2011-05-10T10:08:01 *** amstan_ has joined #aichallenge 2011-05-10T10:08:01 *** ChanServ sets mode: +o amstan_ 2011-05-10T10:08:42 *** amstan has quit IRC (Ping timeout: 248 seconds) 2011-05-10T10:31:11 *** boegel has quit IRC (Ping timeout: 240 seconds) 2011-05-10T10:32:06 *** kaemo has quit IRC (Ping timeout: 258 seconds) 2011-05-10T10:37:33 *** Ice_Harley has quit IRC (Ping timeout: 252 seconds) 2011-05-10T10:41:14 *** amstan_ has quit IRC (Ping timeout: 276 seconds) 2011-05-10T10:52:33 *** MuTa-ofd00m has joined #aichallenge 2011-05-10T10:59:01 *** McLeopold has left #aichallenge 2011-05-10T11:01:19 *** MuTa-ofd00m has quit IRC (Ping timeout: 240 seconds) 2011-05-10T11:07:08 *** contestbot_ has joined #aichallenge 2011-05-10T11:07:30 *** contestbot has quit IRC (Ping timeout: 260 seconds) 2011-05-10T11:11:30 *** MuTa-ofd00m has joined #aichallenge 2011-05-10T11:16:35 *** fmeyer has joined #aichallenge 2011-05-10T11:20:33 *** aerique has quit IRC (Quit: ...) 2011-05-10T11:23:57 *** fmeyer has quit IRC (Quit: leaving) 2011-05-10T11:38:26 *** smiley1983 has joined #aichallenge 2011-05-10T12:20:29 *** GeorgeSebastian has joined #aichallenge 2011-05-10T12:20:29 *** GeorgeSebastian has joined #aichallenge 2011-05-10T12:30:04 *** amstan_ has joined #aichallenge 2011-05-10T12:30:04 *** ChanServ sets mode: +o amstan_ 2011-05-10T12:53:55 *** amstan_ has quit IRC (Read error: Operation timed out) 2011-05-10T12:54:56 *** Frontier has joined #aichallenge 2011-05-10T13:02:12 *** bhasker has joined #aichallenge 2011-05-10T13:06:50 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86.1 [Firefox 4.0.1/20110413222027]) 2011-05-10T13:11:54 *** GeorgeSebastian has quit IRC (Quit: Leaving) 2011-05-10T13:13:22 *** Frontier has quit IRC (Quit: Leaving.) 2011-05-10T13:18:24 *** kaemo has joined #aichallenge 2011-05-10T13:26:50 aichallenge: McLeopold setup-test * r36e4a9d / (27 files in 5 dirs): changes to setup path structure - http://bit.ly/iDP6QQ 2011-05-10T13:38:10 aichallenge: Scott Hamilton epsilon * r124c95e / INSTALL.md : Edited INSTALL.md via GitHub - http://bit.ly/jWN9hu 2011-05-10T13:38:31 aichallenge: Scott Hamilton epsilon * r82e2d7f / INSTALL.md : Edited INSTALL.md via GitHub - http://bit.ly/j6WpMU 2011-05-10T13:40:20 aichallenge: Scott Hamilton epsilon * r326d6a6 / INSTALL.md : Edited INSTALL.md via GitHub - http://bit.ly/jNhSLM 2011-05-10T13:40:36 btw, if you want the contest user like the automatic worker setup would create it the command is 'useradd --disabled-password --gecos "" contest' 2011-05-10T13:48:05 err rather that should the adduser command 2011-05-10T13:52:52 *** mcstar has quit IRC (Remote host closed the connection) 2011-05-10T13:57:20 *** Naktibalda has joined #aichallenge 2011-05-10T13:57:57 *** bhasker has quit IRC (Quit: bhasker) 2011-05-10T13:59:21 *** bhasker has joined #aichallenge 2011-05-10T14:43:34 aichallenge: McLeopold setup-test * rae404a3 / setup/server_setup.py : fixes to server_setup - http://bit.ly/lY0MJ4 2011-05-10T14:43:34 *** McLeopold has joined #aichallenge 2011-05-10T15:12:39 aichallenge: McLeopold setup-test * rc5d1679 / (3 files): updates to setup - http://bit.ly/jO7FeS 2011-05-10T15:27:56 *** mcstar has joined #aichallenge 2011-05-10T15:28:49 *** mcstar has joined #aichallenge 2011-05-10T15:49:30 *** berak has quit IRC (Read error: Connection reset by peer) 2011-05-10T15:53:34 *** amstan_ has joined #aichallenge 2011-05-10T15:53:34 *** ChanServ sets mode: +o amstan_ 2011-05-10T16:01:46 *** amstan_ has quit IRC (Ping timeout: 248 seconds) 2011-05-10T16:06:31 *** Accoun has quit IRC () 2011-05-10T16:21:00 *** Accoun has joined #aichallenge 2011-05-10T16:25:30 *** Frontier has joined #aichallenge 2011-05-10T16:34:18 *** Mathnerd314 has joined #aichallenge 2011-05-10T16:40:59 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86.1 [Firefox 4.0.1/20110422205559]) 2011-05-10T16:55:07 *** amstan_ has joined #aichallenge 2011-05-10T16:55:07 *** ChanServ sets mode: +o amstan_ 2011-05-10T16:55:45 *** Mathnerd314 has quit IRC (Ping timeout: 240 seconds) 2011-05-10T16:58:54 *** delt0r___ has joined #aichallenge 2011-05-10T17:00:13 *** delt0r_ has quit IRC (Ping timeout: 240 seconds) 2011-05-10T17:00:39 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-05-10T17:04:40 *** Frontier has quit IRC (Quit: Leaving.) 2011-05-10T17:09:58 *** teuneboon has joined #aichallenge 2011-05-10T17:10:04 hi all 2011-05-10T17:12:08 *** Frontier has joined #aichallenge 2011-05-10T17:26:19 *** amstan_ has quit IRC (Quit: Konversation terminated!) 2011-05-10T17:27:58 *** teuneboon has quit IRC (Quit: Page closed) 2011-05-10T17:42:14 *** smiley1983 has quit IRC (Ping timeout: 241 seconds) 2011-05-10T17:43:30 *** smiley1983 has joined #aichallenge 2011-05-10T17:50:42 *** bhasker has quit IRC (Quit: bhasker) 2011-05-10T17:51:03 *** ltriant has joined #aichallenge 2011-05-10T18:03:23 antimatroid: you around 2011-05-10T18:03:25 *** smiley1983 has quit IRC (Ping timeout: 260 seconds) 2011-05-10T18:04:11 McLeopold: for boolean options in the setup scripts you probably want to use the get_choice function 2011-05-10T18:05:04 *** smiley1983 has joined #aichallenge 2011-05-10T18:06:10 *** _0x47 <_0x47!~slash@p3EE0B3EF.dip.t-dialin.net> has quit IRC (Quit: Leaving.) 2011-05-10T18:14:36 janzert: yeah, I got that fix, just haven't committed yet :) 2011-05-10T18:14:47 which is why I'm in a branch 2011-05-10T18:15:12 I'm testing stuff on a 10.04 vm right now 2011-05-10T18:15:21 I mean, 11.04 2011-05-10T18:18:27 *** Bleep has joined #aichallenge 2011-05-10T18:18:52 *** Bleep is now known as Guest5502 2011-05-10T18:19:46 great 2011-05-10T18:20:14 janzert: I've got a question about starter packs 2011-05-10T18:20:37 I keep going back and forth between minimalist starter packs and ones with a full set of tools 2011-05-10T18:20:39 yes? although I'm probably the wrong one to ask about them 2011-05-10T18:20:53 right now, you're the only one to ask :) 2011-05-10T18:21:02 full set of tools meaning? 2011-05-10T18:21:10 *** sigh has joined #aichallenge 2011-05-10T18:21:16 depending on the language, plus the engine 2011-05-10T18:21:49 what about full starter packs, but the quick start guide says to download, extract and upload the MyBot.tgz located inside the pack. 2011-05-10T18:22:10 And we reject full starter packs. 2011-05-10T18:22:37 so, each starter pack can have a build script that produces the smaller MyBot file 2011-05-10T18:23:03 I know jeff wants *really* easy, and I'm wondering if that would be easy enough. 2011-05-10T18:23:18 I'm still a little confused I guess 2011-05-10T18:24:07 how so? 2011-05-10T18:24:11 mainly I think jeff wants to be able to take the downloaded package and a) run a game with it locally without other dependancies and b) be able to upload it to the server and have an entry 2011-05-10T18:24:30 what is "the smaller MyBot file"? 2011-05-10T18:25:04 we don't want users uploading hugh sets of files that really aren't part of the bot 2011-05-10T18:25:17 I think splitting the download into two files one with the generic engine, viewer and such and one for each language starter bot is a good idea 2011-05-10T18:25:55 Okay, but the starter pack creators also want to create tools for their language 2011-05-10T18:27:06 So each language would have their own tools files. 2011-05-10T18:27:24 I think that is probably fine, so long as they keep a fairly orderly directory structure that makes it obvious what is part of the bot 2011-05-10T18:27:58 also I would try and keep the main packs c++, java, python fairly minimal 2011-05-10T18:28:09 our build script needs to create 2 zip files then, and know which parts to take from the starter pack dir 2011-05-10T18:28:15 those will cover almost all the entries anyway 2011-05-10T18:28:21 yeah 2011-05-10T18:28:24 they are right now 2011-05-10T18:28:26 especially the ones by newbies 2011-05-10T18:28:48 the more esoteric languages I expect will have users more likely to understand what they have 2011-05-10T18:29:02 I disagree with that 2011-05-10T18:29:19 I wouldn't mind trying out lisp or haskell this time around, but the starter packs were beasts 2011-05-10T18:29:59 but you would still understand pretty quickly what parts of it were tools and what was the actual bot I imagine 2011-05-10T18:30:09 I can't look at the lisp starter and learn lisp 2011-05-10T18:30:18 I really had a hard time with the lisp one, so no 2011-05-10T18:30:34 I can't look at most any lisp and learn lisp ;) 2011-05-10T18:30:52 perl is almost easier to learn by reading :P 2011-05-10T18:31:23 anyway, when it comes to building starter packs and tool sets, I'm thinking each language will get one of each... 2011-05-10T18:31:39 And we need to impose some structure on the starter bot writers 2011-05-10T18:32:11 people did fine without that in the past 2011-05-10T18:32:27 consistent structure in starter bots that is 2011-05-10T18:32:51 perhaps, or maybe people just gave up and didn't complain 2011-05-10T18:33:17 personally I'd rather idiomatic starter bots, i.e. written the way a programmer experienced in the language would naturally write it 2011-05-10T18:33:39 yes, that is good for learning 2011-05-10T18:34:00 but we still need a way to separate tools and extra stuff from what should be submitted in a zip file 2011-05-10T18:34:05 that was something I really disliked about the python and java starter packs in the past 2011-05-10T18:34:29 oh yeah, the python starter was written in java :p 2011-05-10T18:34:52 1 class per file and such 2011-05-10T18:34:54 well, that just entails telling the authors to have a tools directory and a bot directory right? 2011-05-10T18:35:13 something like that, which is why I wanted your opinion 2011-05-10T18:35:25 yeah, and while I'm not as familiar with Java people were complaining about the same sort of thing with that one 2011-05-10T18:35:29 or some way for them to declare what goes where 2011-05-10T18:35:54 ok, I don't have a real problem with that sort of structure 2011-05-10T18:36:10 I thought you meant the structure of the bot program itself 2011-05-10T18:36:14 no 2011-05-10T18:36:24 just enough for our build scripts to be smart 2011-05-10T18:36:57 it might be best to have a per bot, either makefile or build script 2011-05-10T18:37:05 probably 2011-05-10T18:37:44 require something that will produce the 2 zips if not in a dir structure? 2011-05-10T18:38:11 have a recommended dir structure and a required script 2011-05-10T18:39:26 I'd keep the language neutral stuff (engine, visualizer, etc.) as its own download too 2011-05-10T18:40:10 and all a person has to download is the engine pack and a starter bot, they can get the language specific tools if or when they want 2011-05-10T18:40:56 I'm thinking the tools pack can be rolled in with the language tools. 2011-05-10T18:41:16 That way if someone produces a java engine and java playgame, then they can include that instead. 2011-05-10T18:41:32 This might be stuff for zeta. 2011-05-10T18:41:42 that complicates the language packaging though and I want everyone to have the same "official" engine 2011-05-10T18:41:55 hmm... 2011-05-10T18:42:09 what if we had 2 official engines? 2011-05-10T18:42:32 that's the what happened in the past and I'd rather avoid that 2011-05-10T18:42:32 we'd have to have pretty good unit tests for that 2011-05-10T18:42:41 why avoid? 2011-05-10T18:43:13 the problem is that we always end up with bugs/slight differences in behavior on edge cases that confuse people until we get it fixed 2011-05-10T18:43:30 I can see that, thus the unit tests 2011-05-10T18:43:55 even with unit tests I think we'll always miss something 2011-05-10T18:44:18 *** sigh has quit IRC (Remote host closed the connection) 2011-05-10T18:44:19 yeah, it always happens that way, but I don't think we shouldn't try, at least not in the future 2011-05-10T18:44:35 *** amstan has joined #aichallenge 2011-05-10T18:44:35 *** ChanServ sets mode: +o amstan 2011-05-10T18:44:38 but what do we gain by having more than one? 2011-05-10T18:44:54 maybe someone doesn't want to install python 2011-05-10T18:45:25 anymore the only people that don't have it already is windows 2011-05-10T18:45:51 a c++ engine could be much faster, allowing for faster training/testing 2011-05-10T18:45:55 and it's an easy install compared to most other languages 2011-05-10T18:46:20 I'd let the community create that 2011-05-10T18:46:20 this worker install takes forever... 2011-05-10T18:46:26 yep 2011-05-10T18:46:54 its gotta do a full OS install and then the compiled languages take quite a while 2011-05-10T18:47:16 finding ppa's with precompiled packages for all the languages would be really nice 2011-05-10T18:47:23 ppa? 2011-05-10T18:47:39 *** FireFly