2011-03-15T01:10:02 *** Mathnerd314 has quit IRC (Ping timeout: 260 seconds) 2011-03-15T01:58:08 *** delt0r___ has joined #aichallenge 2011-03-15T01:59:21 *** closedbracket has quit IRC (Remote host closed the connection) 2011-03-15T01:59:21 *** delt0r_ has quit IRC (Ping timeout: 250 seconds) 2011-03-15T03:05:32 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Quit: Lost terminal) 2011-03-15T03:09:56 *** Cyndre_ has joined #aichallenge 2011-03-15T03:10:19 *** Cyndre has quit IRC (Ping timeout: 240 seconds) 2011-03-15T03:10:28 *** Cyndre_ has quit IRC (Read error: Connection reset by peer) 2011-03-15T03:10:59 *** Cyndre_ has joined #aichallenge 2011-03-15T03:12:53 *** Cyndre_ has quit IRC (Read error: Connection reset by peer) 2011-03-15T03:13:24 *** Cyndre_ has joined #aichallenge 2011-03-15T03:17:24 *** Cyndre__ has joined #aichallenge 2011-03-15T03:17:38 *** Cyndre_ has quit IRC (Read error: Connection reset by peer) 2011-03-15T03:43:14 can we make the server support submissions with sources spread out over multiple folders? 2011-03-15T04:10:40 *** amstan has quit IRC (Ping timeout: 250 seconds) 2011-03-15T04:19:29 *** aerique has joined #aichallenge 2011-03-15T04:20:11 *** boegel has joined #aichallenge 2011-03-15T04:27:22 *** Frontier has joined #aichallenge 2011-03-15T05:04:46 *** Cyndre has joined #aichallenge 2011-03-15T05:06:45 *** Cyndre__ has quit IRC (Ping timeout: 248 seconds) 2011-03-15T05:08:22 *** kaemo has joined #aichallenge 2011-03-15T05:23:43 *** boegel has quit IRC (Quit: Leaving) 2011-03-15T05:31:57 *** Naktibalda has joined #aichallenge 2011-03-15T07:09:05 *** mceier has joined #aichallenge 2011-03-15T08:05:33 *** stocha has joined #aichallenge 2011-03-15T08:06:10 hi. how is going the java starter package ? 2011-03-15T08:13:00 *** sketchyd_ has joined #aichallenge 2011-03-15T08:47:34 *** sigh has quit IRC (Remote host closed the connection) 2011-03-15T09:05:00 *** UncleVasya has joined #aichallenge 2011-03-15T09:06:49 *** kaemo has quit IRC (Ping timeout: 276 seconds) 2011-03-15T09:43:18 *** stocha has quit IRC (Quit: Page closed) 2011-03-15T09:53:59 *** Frontier has quit IRC (Read error: Connection reset by peer) 2011-03-15T09:54:32 *** Frontier has joined #aichallenge 2011-03-15T09:58:43 *** delt0r_ has joined #aichallenge 2011-03-15T10:00:22 *** delt0r___ has quit IRC (Ping timeout: 260 seconds) 2011-03-15T10:03:02 *** kaemo has joined #aichallenge 2011-03-15T10:17:32 *** |UncleVasya| <|UncleVasya|!kvirc@178-133-221-89.dialup.umc.net.ua> has joined #aichallenge 2011-03-15T10:18:27 *** UncleVasya has quit IRC (Ping timeout: 260 seconds) 2011-03-15T10:29:33 *** Cyndre has quit IRC (Ping timeout: 246 seconds) 2011-03-15T10:29:58 *** Cyndre has joined #aichallenge 2011-03-15T10:49:26 *** McLeopold has left #aichallenge 2011-03-15T10:50:21 *** closedbracket has joined #aichallenge 2011-03-15T10:50:37 *** rebelxt has joined #aichallenge 2011-03-15T10:51:06 *** rebelxt has left #aichallenge 2011-03-15T11:08:30 aichallenge: Marco Leise epsilon * r0a7a536 / (13 files in 2 dirs): prepared visualizer for closure compiler - http://bit.ly/gwQKLi 2011-03-15T11:23:49 *** UncleVasya has joined #aichallenge 2011-03-15T11:25:01 *** UncleVasya has quit IRC (Client Quit) 2011-03-15T11:26:46 *** antimatroid has quit IRC (Quit: Leaving.) 2011-03-15T11:27:17 *** |UncleVasya| <|UncleVasya|!kvirc@178-133-221-89.dialup.umc.net.ua> has quit IRC (Ping timeout: 260 seconds) 2011-03-15T11:31:20 *** antimatroid has joined #aichallenge 2011-03-15T11:34:00 *** kaemo has quit IRC (Ping timeout: 252 seconds) 2011-03-15T11:36:00 *** superflit has joined #aichallenge 2011-03-15T11:37:59 *** amstan__ has joined #aichallenge 2011-03-15T11:37:59 *** ChanServ sets mode: +o amstan__ 2011-03-15T11:38:06 *** amstan has joined #aichallenge 2011-03-15T11:38:06 *** ChanServ sets mode: +o amstan 2011-03-15T11:43:08 *** amstan__ has quit IRC (Ping timeout: 250 seconds) 2011-03-15T12:10:12 *** Cyndre has quit IRC (Read error: Connection reset by peer) 2011-03-15T12:10:38 *** Cyndre has joined #aichallenge 2011-03-15T12:13:58 *** aerique has quit IRC (Quit: ...) 2011-03-15T12:58:48 antimatroid: see https://github.com/aichallenge/aichallenge/issues#issue/16 2011-03-15T12:58:56 antimatroid: that should answer your qn 2011-03-15T13:07:28 aichallenge: Alexandru Stan epsilon * r3c688a4 / (84 files in 20 dirs): Merge branch 'epsilon' of https://github.com/aerique/aichallenge into epsilon - http://bit.ly/etu9aS 2011-03-15T13:08:27 aichallenge: Alexandru Stan epsilon * rb9c5ab4 / README.md : removed trueskill referrence - http://bit.ly/dNrr4U 2011-03-15T13:11:19 *** Umsturz has joined #aichallenge 2011-03-15T13:16:07 *** closedbracket has quit IRC (Remote host closed the connection) 2011-03-15T13:16:28 *** amstan has quit IRC (Read error: Connection reset by peer) 2011-03-15T13:16:29 *** amstan_ has joined #aichallenge 2011-03-15T13:16:29 *** ChanServ sets mode: +o amstan_ 2011-03-15T13:16:58 *** Umsturz has quit IRC (Quit: Umsturz) 2011-03-15T13:24:36 *** Cyndre has quit IRC (Ping timeout: 255 seconds) 2011-03-15T13:54:27 *** closedbracket has joined #aichallenge 2011-03-15T14:03:41 *** readncode has joined #aichallenge 2011-03-15T14:05:34 *** closedbracket has quit IRC (Ping timeout: 255 seconds) 2011-03-15T14:14:52 *** readncode is now known as closedbracket 2011-03-15T14:18:17 *** Naktibalda has quit IRC (Quit: ChatZilla 0.9.86 [Firefox 3.6.15/20110303024726]) 2011-03-15T14:31:52 *** closedbracket has quit IRC (Ping timeout: 252 seconds) 2011-03-15T14:36:30 *** janzert has quit IRC (Quit: Leaving.) 2011-03-15T14:37:09 *** closedbracket has joined #aichallenge 2011-03-15T14:39:08 *** janzert has joined #aichallenge 2011-03-15T14:51:35 amstan_: Do you plan to allow binary uploads? 2011-03-15T15:22:00 aichallenge: McLeopold epsilon-new-schema * r8a4c151 / (13 files in 3 dirs): Added api_game_result.php. Updated worker. - http://bit.ly/goFrOe 2011-03-15T15:22:01 aichallenge: McLeopold epsilon-new-schema * re6ff9cc / (61 files in 9 dirs): Merge branch 'epsilon-new-schema' of github.com:aichallenge/aichallenge into epsilon-new-schema - http://bit.ly/fr0564 2011-03-15T15:28:01 McLeopold: Did you upload some cool stuff? What does that api_game_result do? 2011-03-15T15:28:21 aichallenge: McLeopold epsilon-new-schema * rde27c2e / (4 files): Added bot and map for starter package. - http://bit.ly/en8ksB 2011-03-15T15:29:37 *** McLeopold has joined #aichallenge 2011-03-15T15:29:45 amstan_: ping 2011-03-15T15:31:33 @later tell amstan ready for merge, will not be clean 2011-03-15T15:31:33 McLeopold: Ready to serve, my lord. 2011-03-15T15:31:36 *** McLeopold has left #aichallenge 2011-03-15T15:41:34 *** jbroman has quit IRC (Quit: Scotty, beam me up!) 2011-03-15T15:43:25 *** pgpaskar_ has quit IRC (Ping timeout: 260 seconds) 2011-03-15T15:44:01 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-03-15T15:44:09 *** Zepp has quit IRC (Ping timeout: 276 seconds) 2011-03-15T15:56:27 *** McLeopold has joined #aichallenge 2011-03-15T16:03:35 *** jbroman has joined #aichallenge 2011-03-15T16:05:05 aichallenge: McLeopold epsilon * r8a4c151 / (13 files in 3 dirs): Added api_game_result.php. Updated worker. - http://bit.ly/goFrOe 2011-03-15T16:05:06 aichallenge: McLeopold epsilon * re6ff9cc / (61 files in 9 dirs): Merge branch 'epsilon-new-schema' of github.com:aichallenge/aichallenge into epsilon-new-schema - http://bit.ly/fr0564 2011-03-15T16:05:07 aichallenge: McLeopold epsilon * rde27c2e / (4 files): Added bot and map for starter package. - http://bit.ly/en8ksB 2011-03-15T16:05:09 aichallenge: Scott Hamilton epsilon * r5d42aa0 / (94 files in 10 dirs): Merge branch 'epsilon-new-schema' into epsilon - http://bit.ly/eFVNwS 2011-03-15T16:05:30 *** pgpaskar_ has joined #aichallenge 2011-03-15T16:06:04 *** Zepp has joined #aichallenge 2011-03-15T16:10:16 *** kaemo has joined #aichallenge 2011-03-15T16:15:56 *** Accoun has quit IRC () 2011-03-15T16:24:47 @later tell amstan actually, the merge went ok, it's the merge with sigh that will be tough 2011-03-15T16:24:47 McLeopold: Aye, aye, sir 2011-03-15T16:34:14 *** Accoun has joined #aichallenge 2011-03-15T16:46:39 *** sigh has joined #aichallenge 2011-03-15T16:49:07 *** zoo has joined #aichallenge 2011-03-15T16:59:35 *** sigh has quit IRC (Remote host closed the connection) 2011-03-15T17:05:14 *** McLeopold1 has joined #aichallenge 2011-03-15T17:06:15 janzert: ping 2011-03-15T17:10:33 McLeopold1: yes? 2011-03-15T17:11:12 what do we want as the maximum number of files per directory? 10,000? 2011-03-15T17:11:22 I'm thinking about replay storage. 2011-03-15T17:12:26 and I suppose submission storage as well 2011-03-15T17:12:29 10,000 is probably a safe number. I think we ran into problems at 32,768 before 2011-03-15T17:12:47 ok 2011-03-15T17:13:10 10000 is easy because you can calc the top level dir easily without a computer 2011-03-15T17:23:16 *** boegel has joined #aichallenge 2011-03-15T17:27:09 Should you manage to store replays as files on the same machine that runs the web server, please let me know. It's no big deal then to change the visualizer.php into a static html file and download the replay through JavaScript 2011-03-15T17:28:50 Frontier: what if the visualizer html and js came from a separate subdomain, could the visualizer be completely static html then as well? 2011-03-15T17:29:08 or is some information still needed dynamically from the database? 2011-03-15T17:29:55 that is probably phrased rather badly, in regards to the subdomain 2011-03-15T17:30:12 janzert: I'm not 100% sure about sub-domains, but every detail the visualizer needs can be stored in the replay file. 2011-03-15T17:30:35 I mean the visualizer and replays are served from a common subdomain of ai-contest.com 2011-03-15T17:30:43 say replay.ai-contest.com or some such 2011-03-15T17:31:20 hmm, and js could use say a #! anchor to download the correct replay I suppose? 2011-03-15T17:36:44 janzert: It looks like XmlHttpRequests work only on the exact same hostname. If both the html and the replays are on replay.ai-contest.com that works. An anchor is one possibility, although I currently use the anchor to center the visualizer. ;-) The current scheme which uses a parameter (?game_id=xyz) is working here, too. The script can parse the adress and get the information. 2011-03-15T17:37:17 *** antimatroid has quit IRC (Ping timeout: 252 seconds) 2011-03-15T17:37:37 If you ever loaded the visualizer and it produced an error you might have noticed the game_id is already displayed in the debug information. 2011-03-15T17:37:48 you could load the file in an iframe, couldn't you? 2011-03-15T17:38:00 *** Naktibalda has joined #aichallenge 2011-03-15T17:38:10 what is that? something from apple? 2011-03-15T17:38:22 I hope your kidding 2011-03-15T17:38:24 *** Chris_0076 has joined #aichallenge 2011-03-15T17:39:12 so long as s3 static web serving doesn't choke on parameters being fed to it that would probably the way to go 2011-03-15T17:41:56 janzert: If you know a s3 static web site you could just try it out. 2011-03-15T17:42:37 I set up a test one last week but already removed it (and the corresponding dns entry) 2011-03-15T17:43:07 I could set it back up but it would take a couple hours for the dns changes to propagate again 2011-03-15T17:43:17 *** antimatroid has joined #aichallenge 2011-03-15T17:44:40 I'm 95% sure it will work ^^ 2011-03-15T17:46:02 actually just realized testing it with the default bucket url should be fine 2011-03-15T17:49:55 yeah, looks like s3 is fine with arbitrary parameters in the url 2011-03-15T17:52:10 btw, it seems javascript can't access information in an iframe from another domain 2011-03-15T17:54:23 yeah, cross-site-scripting was a big security problem for a while 2011-03-15T17:57:59 http://softwareas.com/cross-domain-communication-with-iframes 2011-03-15T17:59:15 *** delt0r___ has joined #aichallenge 2011-03-15T18:01:06 *** delt0r_ has quit IRC (Ping timeout: 260 seconds) 2011-03-15T18:04:58 *** sigh has joined #aichallenge 2011-03-15T18:05:11 *** boegel has quit IRC (Quit: *poof!*) 2011-03-15T18:06:00 that doesn't seem to help? 2011-03-15T18:06:28 you can't read anything out of the iframe 2011-03-15T18:06:48 just set (parent or child) urls 2011-03-15T18:25:09 *** superflit has quit IRC (Quit: superflit) 2011-03-15T18:27:31 McLeopold: oh yeah, amstan wants to move any ants stuff out of manager (which is just map_analyzer I think) 2011-03-15T18:27:57 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has quit IRC (Ping timeout: 255 seconds) 2011-03-15T18:31:50 antimatroid: ok 2011-03-15T18:39:58 *** suicidolt has joined #aichallenge 2011-03-15T18:44:33 *** zoo has quit IRC (Quit: Page closed) 2011-03-15T19:03:02 janzert: what do you think about older submission getting less games? 2011-03-15T19:05:27 I have mixed feelings about it but in any case I don't think jeff would ever go for it 2011-03-15T19:05:38 I could be surprised though 2011-03-15T19:06:20 well, with truescore, old submissions should stabalize 2011-03-15T19:06:25 trueskill 2011-03-15T19:07:11 if there the servers were to stay up after the finals I think it should be done for sure at that point 2011-03-15T19:07:21 My thought is that a submission gets to be a seed if the sigma is above a certain value 2011-03-15T19:07:54 When a submission is submitted, all submissions get an increase in sigma 2011-03-15T19:08:12 so, if enough are submitted, then even old submissions will be seeded again 2011-03-15T19:08:27 but if not, then they only get games when chosen to sort out the new submissions 2011-03-15T19:09:29 that might work fine, I'm not sure all the side effects are immediately apparent though 2011-03-15T19:09:48 resubmitting the same thing to get more games 2011-03-15T19:10:00 * janzert nods 2011-03-15T19:10:08 but we could check the submission hash 2011-03-15T19:10:27 *** closedbracket has quit IRC (Remote host closed the connection) 2011-03-15T19:10:33 I think that hack was only to get a few games right away. 2011-03-15T19:10:58 what hack is that? 2011-03-15T19:10:59 If we gear the system to give them a few game right away on a new submit, we may prevent that from happening as much. 2011-03-15T19:11:24 uh, hack as in gaming the system 2011-03-15T19:11:45 oh resubmitting constantly 2011-03-15T19:11:48 yeah 2011-03-15T19:12:13 wouldn't giving more games to a new resubmit encourage it more? 2011-03-15T19:12:30 I'm thinking it would be the opposite 2011-03-15T19:12:39 Why would someone want more games? 2011-03-15T19:13:00 so they can see their submission play more games 2011-03-15T19:14:01 *** Frontier has quit IRC (Ping timeout: 260 seconds) 2011-03-15T19:14:20 Well, this is more turning off old submission. 2011-03-15T19:14:22 I don't understand how more games would ever be something people wouldn't want 2011-03-15T19:14:42 right, so long as it takes a reasonable time to kick in 2011-03-15T19:14:47 So, if everyone resubmits constantly, we are back to normal. 2011-03-15T19:15:03 *** Frontier has joined #aichallenge 2011-03-15T19:15:05 before it kicks in just be sure to rotate through the players 2011-03-15T19:15:34 rotate how? 2011-03-15T19:15:37 and I wouldn't mind seeing it but I still have a feeling jeff will be opposed 2011-03-15T19:16:08 always choose the seed to be the player least recently used as a seed 2011-03-15T19:16:39 What I'm suggesting is to stop using old submissions as a seed, but still allow it to be matched. 2011-03-15T19:16:51 * janzert nods 2011-03-15T19:17:19 Okay, the userid, id still gets chosen, but they don't have a new submissions, so they get passed. 2011-03-15T19:17:52 right, only use the userids that have a submission that is eligible 2011-03-15T19:19:08 I think I'd just leave it out for now though and then after the contest has run for a few weeks check what the sigma is for submissions that are still active and were submitted at the beginning 2011-03-15T19:19:27 then use that as the cutoff 2011-03-15T19:20:17 *** McLeopold1 has quit IRC (Read error: Connection reset by peer) 2011-03-15T19:35:23 aichallenge: Marco Leise epsilon * r60626ff / (4 files in 2 dirs): graphical changes to the visualizer - http://bit.ly/hBnfa2 2011-03-15T19:36:25 *** McLeopold1 has joined #aichallenge 2011-03-15T19:41:36 *** Cyndre has joined #aichallenge 2011-03-15T19:45:39 *** davidjliu has joined #aichallenge 2011-03-15T19:45:43 I'm pretty sure the choose the seed based on the person who played a game least recently. 2011-03-15T19:46:08 So, that should roughly equalize the games, even when truskill just wants to play the middle of the field over and over. 2011-03-15T19:46:49 Plus, I'm not sure we'll ever reduce to "just" trueskill. That doesn't happen until the graph is fully connected. 2011-03-15T19:46:59 janzert: McLeopold1: ^^ 2011-03-15T19:50:28 aichallenge: Marco Leise epsilon * r7e1168d / (62 files in 12 dirs): moved ants/www/visualizer to ants/visualizer - http://bit.ly/hTCjKE 2011-03-15T19:53:02 bss03: yes, that is the current system 2011-03-15T19:53:24 McLeopold was proposing to change it so old submissions would never be chosen as the seed 2011-03-15T19:55:15 janzert: do you know php? 2011-03-15T19:55:53 not proficiently 2011-03-15T19:56:08 enough to read and stumble through writing it though 2011-03-15T19:56:20 I'm struggling with the dumbest thing. I can't get a php page to write a file. 2011-03-15T19:57:09 I can't tell if it's windows, apache or php that's got the safety on. 2011-03-15T19:57:36 I guess I can't remember ever directly writing to a file from php myself 2011-03-15T19:58:00 I want to take the posted game result and create a replay file. 2011-03-15T19:58:12 * janzert nods 2011-03-15T19:58:34 hmm, actually I guess the old code should have done that as well 2011-03-15T19:58:43 it put it in the database 2011-03-15T19:58:52 so, no permission issues there 2011-03-15T19:58:54 oh yeah :} 2011-03-15T19:59:04 but here's the kicker, I can create a directory 2011-03-15T19:59:10 just not a file 2011-03-15T19:59:46 just as a double check, you're opening the file in a write mode correct? 2011-03-15T20:00:16 e.g. fopen(filename, "w") 2011-03-15T20:00:52 yea 2011-03-15T20:01:06 And I've tried relative and absolute as well. 2011-03-15T20:01:19 I'm doing absolute for the mkdir() call, which works. 2011-03-15T20:02:13 hmm, I see in safe mode the directory must have the same owner as the script being executed 2011-03-15T20:02:21 not sure how that works out on windows though 2011-03-15T20:02:41 maybe try disabling safe mode once just to see 2011-03-15T20:03:10 hmm, I see safe mode is actually deprecated now 2011-03-15T20:04:14 yeah, and I don't think it is on... 2011-03-15T20:05:01 do you get any error or the file just doesn't get written? 2011-03-15T20:06:19 *** mceier has quit IRC (Quit: leaving) 2011-03-15T20:06:57 let me check 2011-03-15T20:09:55 janzert: crap, I'm an idiot 2011-03-15T20:10:46 sometimes a problem just has to be talked through :) 2011-03-15T20:11:22 *** suicidolt has quit IRC (Read error: No route to host) 2011-03-15T20:11:47 what was the error? 2011-03-15T20:11:51 lurkers want to know 2011-03-15T20:12:18 *** amstan_ is now known as amstan 2011-03-15T20:12:21 hmm 2011-03-15T20:12:38 McLeopold1: sorry about that, i only got home at like 5 2011-03-15T20:12:44 McLeopold1: well ... that was the plan 2011-03-15T20:12:52 but then my car broke down, so i did all the towing fun 2011-03-15T20:12:55 aichallenge: McLeopold epsilon * r48d6a23 / (3 files in 2 dirs): Added writing replay file to api_game_result.php - http://bit.ly/gbJRji 2011-03-15T20:12:58 aichallenge: McLeopold epsilon * r0849192 / (61 files in 12 dirs): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://bit.ly/fVvsuC 2011-03-15T20:13:18 sigh: Luckily I don't think I committed the error. :) 2011-03-15T20:14:48 amstan: okay, the plan for tonight is to pull in sigh's changes to the engine, which are big 2011-03-15T20:14:55 having not worked with git before is it normal on git to get all these commits merging a branch into itself? 2011-03-15T20:15:04 After that, I plan to hand off every todo to anyone I can 2011-03-15T20:15:12 janzert: well.. you're pretty much merging 2 branches 2011-03-15T20:15:24 janzert: since 2 people work at the same time on the same branch without pulling eachother's changes 2011-03-15T20:15:28 janzert: there has to be a merge 2011-03-15T20:15:29 a merge is a commit, or action that should be tracked 2011-03-15T20:15:56 but why do you have to merge it into itself? 2011-03-15T20:16:04 janzert: it's not into itself 2011-03-15T20:16:07 I just don't remember ever seeing this under bzr or hg 2011-03-15T20:16:18 janzert: look at this: https://github.com/aichallenge/aichallenge/network/ 2011-03-15T20:16:31 amstan: I'll be back in 2 hours or so 2011-03-15T20:16:38 McLeopold1: perfect 2011-03-15T20:16:38 then we can go over what's left 2011-03-15T20:16:43 *** McLeopold1 has left #aichallenge 2011-03-15T20:17:01 janzert: Frontier and McLeopold1 started from 5d42 2011-03-15T20:17:17 janzert: then they did stuff without pushing or pulling 2011-03-15T20:17:32 and then one of them pulled before pushing so the merge was created 2011-03-15T20:17:47 ok 2011-03-15T20:19:14 sigh: are you going to be available in 2 hours? 2011-03-15T20:19:50 I'm going to be online, but I won't available for anything indepth... I'm at work 2011-03-15T20:29:58 sigh: when are you? 2011-03-15T20:30:59 umm... in about 8 hours, my timezone doesn't match up well with you guys :P 2011-03-15T20:31:06 darn 2011-03-15T20:31:13 where are you from? 2011-03-15T20:31:17 Australia 2011-03-15T20:31:22 same tz as antimatroid 2011-03-15T20:31:40 except he stays up at ungodly hours 2011-03-15T20:32:06 melbourne, australia gmt+11 2011-03-15T20:33:47 hmm 2011-03-15T20:34:02 what about morning here? 2011-03-15T20:34:33 what's your timezone (gmt offset)? 2011-03-15T20:34:37 -5 2011-03-15T20:35:33 I think I go to sleep about 8am your time 2011-03-15T20:36:00 darn 2011-03-15T20:36:30 wait.. what? so 8 hours from now is 6 am from me 2011-03-15T20:36:37 then you go to sleep at 8 am 2011-03-15T20:36:53 hmm... I calculated wrong 2011-03-15T20:37:22 no.. i did darn 2011-03-15T20:37:23 4 am 2011-03-15T20:37:30 that's kinda better 2011-03-15T20:37:34 then you have 4 hours after work 2011-03-15T20:38:43 yeah, although it's a bit more if I just hit the ground running as soon as I get home from work :P 2011-03-15T20:39:24 also my weekend mornings matches up well, which is your friday and saturday around this time 2011-03-15T20:45:29 *** Cyndre has quit IRC (Quit: Leaving) 2011-03-15T20:47:39 *** _flag <_flag!~flag@69-165-173-172.dsl.teksavvy.com> has joined #aichallenge 2011-03-15T20:48:31 *** Cyndre has joined #aichallenge 2011-03-15T21:10:19 sigh: can you explain the structure of your branches? what's sym-food, epsilon and replay supposed to do? 2011-03-15T21:13:27 *** davidjliu has quit IRC (Ping timeout: 252 seconds) 2011-03-15T21:14:33 *** smellyhippy has quit IRC (Ping timeout: 252 seconds) 2011-03-15T21:16:07 all the changes I've been talking to McLeopold about are in epsilon 2011-03-15T21:16:20 Is playgame.py currently broken? 2011-03-15T21:16:37 the other ones are branches I had on my local repo which I uploaded to show code to others 2011-03-15T21:16:38 Frontier: maybe... mcleopold merged so.. 2011-03-15T21:16:41 don't worry about them 2011-03-15T21:16:55 sigh: cool 2011-03-15T21:17:13 * amstan happly submits github bug: https://github.com/aichallenge/aichallenge/network 2011-03-15T21:17:19 err: http://support.github.com/discussions/graphs/272-network-graph-overlapping 2011-03-15T21:21:03 *** amriedle has joined #aichallenge 2011-03-15T21:22:14 contestbot: seen bss03 2011-03-15T21:22:14 amstan: bss03 was last seen in #aichallenge 1 hour, 35 minutes, and 15 seconds ago: janzert: McLeopold1: ^^ 2011-03-15T21:22:19 bss03: ping 2011-03-15T21:22:40 amstan: pong 2011-03-15T21:22:54 bss03: how's the matchmaking changes going? 2011-03-15T21:23:10 I'm in and out for around 90 minutes right now. 2011-03-15T21:23:26 I haven't worked on them since... Thursday, I think. 2011-03-15T21:23:32 I replied to your stuff on Issue 7. 2011-03-15T21:23:49 bss03: yeah, i noticed 2011-03-15T21:23:54 Where you trying to link to an interface I should be modeling? 2011-03-15T21:24:02 bss03: the files i was linking.. i was almost sure they existed in epsilon 2011-03-15T21:24:07 I've just been concerntrating on the SQL, no matching an API. 2011-03-15T21:24:16 *not 2011-03-15T21:24:17 bss03: so i changed the links from epsilon-new-schema 2011-03-15T21:24:24 bss03: ok.. in that case 2011-03-15T21:24:55 I can cerntainly do that (match an interface) though. The SQL is useable right now, just not ideal. 2011-03-15T21:26:26 bss03: well.. there's a file from elo, that should be replaced with your stuff 2011-03-15T21:27:18 create_matchups.py? 2011-03-15T21:27:28 bss03: yes 2011-03-15T21:29:09 I'm not exactly sure I can match that API or that we want to. 2011-03-15T21:29:23 ok, that's fine 2011-03-15T21:29:36 we can do a new api if it's needed 2011-03-15T21:29:39 But, I can implement an add_matches that's at least mostly the same, modulo necessarily schema changes. 2011-03-15T21:30:59 But, yeah, that's what I'm gonna replace eventually. I can prioritize that over getting the SQL just right. 2011-03-15T21:31:25 That should make it easier for people with larger data sets or a full-fledge test server to look at the results of the matching algo. 2011-03-15T21:34:25 No promises on when; I never seem to have enough time after work. 2011-03-15T21:36:30 *** McLeopold has left #aichallenge 2011-03-15T21:39:46 amstan: could you close the new-schema branch now that it's been merged back into epsilon? 2011-03-15T21:39:55 janzert: sure 2011-03-15T21:45:54 Bot 2 timed out, lol. The HunterBot needs some optimization ;) 2011-03-15T21:47:08 *** closedbracket has joined #aichallenge 2011-03-15T21:56:46 aichallenge: janzert epsilon * rb2f8621 / sql/0_schema.sql : Fix copy-paste errors in new schema - http://bit.ly/fJ9Xwv 2011-03-15T21:56:46 aichallenge: janzert epsilon * rcc3499c / sql/0_schema.sql : Add missing comma in schema - http://bit.ly/h8Ztlt 2011-03-15T21:56:47 aichallenge: janzert epsilon * ra4471f2 / setup/server_setup.py : Update server_setup.py to work with new schema - http://bit.ly/fHkLo6 2011-03-15T21:56:48 aichallenge: janzert epsilon * rfb2cf42 / (3 files in 2 dirs): Merge branch 'epsilon' of github.com:aichallenge/aichallenge into epsilon - http://bit.ly/eAvF44 2011-03-15T22:03:44 lawls 2011-03-15T22:04:06 amriedle: hmm? 2011-03-15T22:04:14 the comma error 2011-03-15T22:04:26 always seems to happen i guess 2011-03-15T22:04:32 oh yeah 2011-03-15T22:05:00 i have gotten to the point where commas are stupid :P 2011-03-15T22:05:03 I was rather surprised that the schema wasn't the actual output from a mysqldump command 2011-03-15T22:05:06 and should not be needed 2011-03-15T22:06:10 janzert: use to bad programming habits? 2011-03-15T22:06:51 it's a way to guarantee that you have a schema file actually matching the database you're using at least :) 2011-03-15T22:07:57 contestbot: later tell mcleopold the ants replay format link from the wiki(main page) is broken 2011-03-15T22:07:57 amstan: Yep. 2011-03-15T22:09:10 janzert: it matters on what you consider deployment :P 2011-03-15T22:31:02 *** amstan_ has joined #aichallenge 2011-03-15T22:31:02 *** ChanServ sets mode: +o amstan_ 2011-03-15T22:34:29 *** McLeopold has joined #aichallenge 2011-03-15T22:34:31 *** amstan has quit IRC (Ping timeout: 250 seconds) 2011-03-15T22:38:37 *** McLeopold has quit IRC (Client Quit) 2011-03-15T22:50:51 *** McLeopold has joined #aichallenge 2011-03-15T22:52:08 *** Naktibalda has quit IRC (Ping timeout: 255 seconds) 2011-03-15T22:54:47 hey all, I'm here and ready to work 2011-03-15T22:56:01 McLeopold: woot 2011-03-15T22:56:20 Python has zip? Where? 2011-03-15T22:56:22 so, I need to update the schema file, I forgot to include that in the push 2011-03-15T22:56:29 python has everything 2011-03-15T22:56:49 import zipfile 2011-03-15T22:57:01 I need to zip current_players [1..] 2011-03-15T22:57:08 I mean the functional programming zip. 2011-03-15T22:57:29 oh 2011-03-15T22:57:36 zip [1,2,3] ['a','b','c'] = [(1,'a'),(2,'b'),(3,'c')] 2011-03-15T22:57:53 how come you're in python? 2011-03-15T22:58:08 zip([1,2,3], ['a','b','c']) 2011-03-15T22:58:16 McLeopold: Working on the matchups stuff. 2011-03-15T22:58:47 McLeopold: I think amstan wants be to wrap the existing code in an API so it easier to test. 2011-03-15T22:58:48 it's a builtin function btw 2011-03-15T22:58:48 I though we were going all sql? 2011-03-15T22:58:52 McLeopold: so how stable/complete is the code? 2011-03-15T22:58:59 *** amstan_ is now known as amstan 2011-03-15T22:59:20 once I get the sql change pushed, it's stable, but not complete 2011-03-15T22:59:36 todos... 2011-03-15T22:59:40 McLeopold: notice I checked in some schema type fixes earlier 2011-03-15T22:59:45 *** chris__0076 has joined #aichallenge 2011-03-15T22:59:47 err, typo 2011-03-15T23:00:04 Get a trueskill script to read a game from the database and update the mu/sigma of the game and the submission 2011-03-15T23:00:18 janzert: I pulled, so I've got them 2011-03-15T23:00:21 I need to add the player_id column; it's not in my temporary table. 2011-03-15T23:00:41 McLeopold: how do you install it? 2011-03-15T23:00:45 That's why need the zip([0..],current_players) 2011-03-15T23:01:47 *** Chris_0076 has quit IRC (Ping timeout: 246 seconds) 2011-03-15T23:01:54 enumerate(current_players) is more idomatic, if anyone cares 2011-03-15T23:01:56 McLeopold: Speaking of mu and sigma, are our initial values for those available from python? 2011-03-15T23:02:06 I know they are defaults for the resp. columns. 2011-03-15T23:02:09 * janzert was about to say the same as sigh 2011-03-15T23:02:18 defaults in sql 2011-03-15T23:02:30 amstan: no install, I just keep working on my system 2011-03-15T23:02:56 But, I need to use BETA which is derived from initial sigma. 2011-03-15T23:03:11 Todo: get a submission to add a counter to the database, every so often increment all the sigma's by the counter 2011-03-15T23:03:21 McLeopold: well.. what are the steps required to install, assuming i have all dependencies already 2011-03-15T23:03:27 McLeopold: just add the cronjobs? 2011-03-15T23:03:47 amstan: I haven't worked on cron jobs at all, janzert has done all the install scripts 2011-03-15T23:04:05 So, the latest thing I've done is create worker.py 2011-03-15T23:04:18 ok, i guess i should be "interogating" him then 2011-03-15T23:04:27 for install, yes 2011-03-15T23:04:41 the install script don't do anything with cron jobs on the main server yet since I don't know what needs to be run with the new structure 2011-03-15T23:04:53 McLeopold, so is there are way to SELECT the default? Like should I just query sys_tables to get the initial sigma? 2011-03-15T23:05:02 it just installs all the packages and points apache at the right spot basically 2011-03-15T23:05:20 bss03: we will put values in server_info.py then 2011-03-15T23:05:27 janzert: so it installs the website pretty much? 2011-03-15T23:06:08 *** amriedle has quit IRC (Quit: Lost terminal) 2011-03-15T23:06:14 *** chris__0076 has quit IRC (Remote host closed the connection) 2011-03-15T23:06:23 McLeopold. Okay. I'll add Beta there, too. It is independently tunable, but the TrueSkill paper suggested initial_sigma / 2 as an example. 2011-03-15T23:06:26 oh wow: "src_repo": "git://github.com/aichallenge/aichallenge.git", 2011-03-15T23:06:26 gets the source repository if needed and sets up the directories and server_info.(py, php) files 2011-03-15T23:06:48 creates the contest user 2011-03-15T23:07:00 creates the database loads the schema and initial data 2011-03-15T23:07:03 wow, I like ubuntu desktop :) 2011-03-15T23:07:08 I think that's about it :) 2011-03-15T23:07:33 janzert: pretty cool 2011-03-15T23:07:45 janzert: and what do you mean by "Do NOT run this on a machine that you use for anything else" 2011-03-15T23:08:00 what kind of invasive changes? 2011-03-15T23:08:12 well, the default is better now but initially the default was what is now --take-over-server 2011-03-15T23:08:28 oh, ok 2011-03-15T23:08:29 meaning it creates users, points apache default website at the contest site,... 2011-03-15T23:09:00 the default may still be a little invasive though so I haven't changed the comment yet 2011-03-15T23:09:39 hmm 2011-03-15T23:09:52 how invasive? it' 2011-03-15T23:10:10 the big thing might be if mysql isn't installed on the server it will install it with a blank root password 2011-03-15T23:10:17 not sure how people would feel about that 2011-03-15T23:10:19 oh, that's fine 2011-03-15T23:10:26 for my case.. 2011-03-15T23:10:55 i'm just wondering if i can run this on my server for deving without breaking my other sites 2011-03-15T23:11:00 or.. anything else 2011-03-15T23:11:02 it set up a http://ai-contest.com website with apache 2011-03-15T23:11:12 but it won't take over the default site 2011-03-15T23:11:20 cool 2011-03-15T23:12:02 you can change that url with the website_hostname option but it's not exposed on the command line yet 2011-03-15T23:12:44 I think other than that so long as it's run from a full git repo it shouldn't do anything really invasive anymore 2011-03-15T23:13:14 it does expect to be run as root still and will try and pick up the username that you sudo'd from 2011-03-15T23:14:24 janzert: so it doesn't really do anything that i don't already have installed 2011-03-15T23:14:48 as in this: http://home.hypertriangle.com/~alex/aichallenge/website/ 2011-03-15T23:14:49 not if you already have the database set up and apache configured 2011-03-15T23:15:04 yep, probably not 2011-03-15T23:15:06 :( ok, lol 2011-03-15T23:15:14 McLeopold: k.. so back to you, how do i install these? 2011-03-15T23:15:24 I'm not sure there is anything beyond that that is currently working 2011-03-15T23:15:35 what do you mean by "these"? 2011-03-15T23:15:45 McLeopold: manager+worker 2011-03-15T23:15:55 they just kinda sit there 2011-03-15T23:16:05 for the worker, play with worker.py 2011-03-15T23:16:24 it does compiles and games 2011-03-15T23:16:38 \o/ for compiling on workers 2011-03-15T23:16:59 what's with https://github.com/aichallenge/aichallenge/blob/epsilon/worker/0_schema.sql ? 2011-03-15T23:16:59 python worker.py -h 2011-03-15T23:17:17 that can be removed when we get the jail users going 2011-03-15T23:17:40 It needs to be added back into sandbox.py 2011-03-15T23:17:59 McLeopold: do i have to execute that or something? 2011-03-15T23:18:20 sandbox? no, it would be called by engine.py, which is called by worker.py 2011-03-15T23:18:44 McLeopold: Will transactions work properly in this MySQL DB? If not, I guess we are fine with a race condition where match is populated before match_player is ready? 2011-03-15T23:18:47 engine.py modifies the sys.path to get the ants.py game, which I'm not thrilled with, but it works 2011-03-15T23:19:01 McLeopold: If not, is there are status flag I'm missing? 2011-03-15T23:19:32 amstan: that file is the schema used for the worker database, currently only used to track what jailusers are actively in use 2011-03-15T23:19:35 bss03: I think we should use transactions, I'm attempting it when I move the matchup data to the game data 2011-03-15T23:19:58 janzert: have we agreed on jail user management yet? 2011-03-15T23:20:09 well.. do i need to set that up if i'm using it? 2011-03-15T23:20:16 :) not really, but I was leaving it up to you basically 2011-03-15T23:20:25 amstan: jail users aren't hooked up, so no 2011-03-15T23:20:40 although I did think of another reason for central control of the jailusers 2011-03-15T23:20:45 do tell 2011-03-15T23:20:51 McLeopold: Okay, I'll call commit just before inseting into match/match_player and then again after I am done with them. 2011-03-15T23:20:55 janzert: do you do any sort of schema setup in your setup.py? 2011-03-15T23:21:12 bss03: btw, in a few minutes, it will become matchup/matchup_player 2011-03-15T23:21:21 I was looking into a chroot environment for the jailusers and we won't want to create hundreds of those I don't think 2011-03-15T23:21:32 amstan: yes 2011-03-15T23:21:42 janzert: oh, that link I sent a few days ago? 2011-03-15T23:21:52 McLeopold: actually no 2011-03-15T23:22:02 just ubuntu chroot in general 2011-03-15T23:22:03 I think chroot is probably the way to go, anyway 2011-03-15T23:22:39 yep 2011-03-15T23:23:13 McLeopold: Okay. I'm not gonna commit anything tonight. It's already past bedtime. 2011-03-15T23:23:16 http://stackoverflow.com/questions/4060772/sqlite3-concurrent-access 2011-03-15T23:23:21 bss03: please do commit 2011-03-15T23:23:29 if you have stuff to commit 2011-03-15T23:23:53 janzert: I think sqlite3 would work, mysql is a big hammer to use for the workers, and sqlite3 comes with python 2011-03-15T23:24:02 McLeopold: There's two commits in my branch that could be merged (they work) the rest of my local tree is not ready. 2011-03-15T23:24:17 that would be fine or just go to a file based lock of some sort 2011-03-15T23:24:28 I think zeta already implemented that actually 2011-03-15T23:24:28 bss03: are you working in epsilon, and not epsilon-new-schema? 2011-03-15T23:24:35 McLeopold: I'll probably rebase tomorrow evening before I start work on this anyway. 2011-03-15T23:24:46 janzert: would you be willing to do the jail user stuff? 2011-03-15T23:24:57 McLeopold: Just rebased onto epsilon from where I was. No confilcts or anything. 2011-03-15T23:25:08 bss03: cause epsilon-new-schema is not used anymore 2011-03-15T23:25:08 ehh, don't count on me doing it but I may be able to if no one else gets to it first 2011-03-15T23:25:48 amstan: what about you? are you comfortable with setting up jail users with chroot? 2011-03-15T23:26:09 McLeopold: nope, i have no idea what i'm doing yet 2011-03-15T23:26:40 hmm... what about jbroman? 2011-03-15T23:27:00 oh yeah, he can definatelly handle it 2011-03-15T23:27:19 Do we have a target date yet? I heard something about Fri + 2 weeks? 2011-03-15T23:27:46 bss03: would be nice 2011-03-15T23:27:56 but probably won't happen 2011-03-15T23:28:00 but would be nice, so let's try 2011-03-15T23:28:32 Kk. 2011-03-15T23:31:05 amstan: do github issues have an assignment feature? it seems like it's just a forum and not an issue tracker 2011-03-15T23:31:08 *** amstan__ has joined #aichallenge 2011-03-15T23:31:08 *** ChanServ sets mode: +o amstan__ 2011-03-15T23:31:17 McLeopold: nope 2011-03-15T23:31:28 McLeopold: it pretty much is a forum, with tags 2011-03-15T23:31:43 can they be "closed" 2011-03-15T23:32:37 McLeopold: yes 2011-03-15T23:32:53 but not by anyone, right? 2011-03-15T23:33:00 McLeopold: not sure 2011-03-15T23:36:28 *** amstan__ has quit IRC (Ping timeout: 252 seconds) 2011-03-15T23:44:47 McLeopold: epsilon new schema doesn't even exist anymore 2011-03-15T23:44:55 McLeopold: anyway.. so.. what's up? what can i do? 2011-03-15T23:45:22 umm... 2011-03-15T23:46:09 can we start a todo list on etherpad? 2011-03-15T23:46:26 just for now, for brainstorming 2011-03-15T23:46:30 ok, sure 2011-03-15T23:47:06 http://ietherpad.com/sMmBH9htlr 2011-03-15T23:54:12 contestbot: seen nullkuhl 2011-03-15T23:54:12 amstan: nullkuhl was last seen in #aichallenge 3 weeks, 4 days, 2 hours, 41 minutes, and 19 seconds ago: allright you will see it soon :) 2011-03-15T23:54:33 McLeopold: yep ^ 2011-03-15T23:54:38 ouch 2011-03-15T23:58:42 * janzert wonders if api_* should move to api/* 2011-03-15T23:58:55 McLeopold: well, what about the build systems? are we gonna support that? 2011-03-15T23:58:58 janzert: +1 2011-03-15T23:58:59 would make my directory listing easier to find stuff in :P 2011-03-15T23:59:50 actually probably worker_api/* would be even better