| *** tpb has joined #tp | 00:00 | |
| *** ChanServ sets mode: +o tpb | 00:00 | |
| *** StupidIncarnate has quit IRC | 00:09 | |
| *** Agon has joined #tp | 00:13 | |
| *** verhoevenv has quit IRC | 00:49 | |
| *** Erroneous has quit IRC | 01:20 | |
| *** glew has quit IRC | 02:49 | |
| *** tote has quit IRC | 03:30 | |
| *** tote has joined #tp | 03:39 | |
| *** llnz has joined #tp | 04:54 | |
| llnz | hi all | 05:04 |
|---|---|---|
| *** muthusuba has joined #tp | 07:09 | |
| *** Vadtec_ has quit IRC | 07:12 | |
| *** Vadtec_ has joined #tp | 07:23 | |
| *** null_000 has joined #tp | 07:24 | |
| null_000 | morning all | 07:24 |
| *** Q_x has joined #tp | 07:40 | |
| Q_x | hiho | 07:40 |
| *** Q_x has quit IRC | 07:41 | |
| *** Vadtec_ has quit IRC | 08:02 | |
| *** Vadtec_ has joined #tp | 08:09 | |
| *** bisc has joined #tp | 09:34 | |
| *** mithro has joined #tp | 09:49 | |
| bisc | mithro: ping | 10:01 |
| null_000 | mithro: ping | 10:08 |
| * llnz wonders if he can help | 10:09 | |
| null_000 | llnz: issue 97 is blocking my AI development... can't do any serious testing. | 10:10 |
| null_000 | llnz: it looks like the problem is that timestamps of an object are not (always) updated when orders of that object are updated | 10:11 |
| null_000 | is this a bug or a feature? | 10:13 |
| llnz | bug, probably | 10:15 |
| llnz | there is also an issue between order queues and objects | 10:16 |
| null_000 | this might be related... | 10:20 |
| null_000 | since the python client started having this problem when the library "corrected" something with order queue id and object id | 10:21 |
| null_000 | if anyone plans to look into this... information here might help http://code.google.com/p/thousandparsec/issues/detail?id=97 | 10:29 |
| tpb | <http://ln-s.net/75lY> (at code.google.com) | 10:29 |
| llnz | it could be server side, I will have a look tomorrow | 10:46 |
| * llnz wanders off | 10:46 | |
| llnz | later all | 10:46 |
| null_000 | later | 10:46 |
| *** llnz has quit IRC | 10:48 | |
| bisc | null_000: I'm trying to reproduce your bug, but ai client sometimes (and sometimes not) meets an error at end of turn. here's log: | 11:05 |
| bisc | null_000: Now in python mode! | 11:06 |
| bisc | I am 2 . My name is ai | 11:06 |
| bisc | Traceback (most recent call last): | 11:06 |
| bisc | File "./daneel-ai.py", line 267, in <module> | 11:06 |
| bisc | gameLoop(options.filename,turns=options.numturns,uri=options.uri,verbosity=options.verbosity,benchmark=options.benchmark) | 11:06 |
| bisc | File "./daneel-ai.py", line 200, in gameLoop | 11:06 |
| bisc | gameLoopWrapped(rulesfile,turns,connection,cache,verbosity,benchmark) | 11:06 |
| bisc | File "./daneel-ai.py", line 221, in gameLoopWrapped | 11:06 |
| bisc | endTurn(cache,rulesystem,connection) | 11:06 |
| bisc | File "./daneel-ai.py", line 143, in endTurn | 11:06 |
| bisc | m.endTurn(cache,rulesystem,connection) | 11:06 |
| bisc | File "./daneel/mod-minisec.py", line 22, in endTurn | 11:06 |
| bisc | AICode() | 11:06 |
| bisc | File "./daneel/mod-minisec.py", line 229, in AICode | 11:06 |
| bisc | helper.printAboutMe() | 11:06 |
| bisc | File "./daneel/helper.py", line 123, in printAboutMe | 11:06 |
| bisc | for x in enemies(): | 11:06 |
| bisc | File "./daneel/helper.py", line 118, in enemies | 11:06 |
| bisc | players.remove(whoami()) | 11:06 |
| bisc | ValueError: list.remove(x): x not in list | 11:06 |
| null_000 | now this is weird... | 11:07 |
| bisc | null_000: this error seems to be probabilistic. this time only one ai client has managed to get to the end of 4th turn. and I connected with the latest tpclient-pywx with no problems. | 11:08 |
| null_000 | you can avoid running this code by commenting out helper.printAboutMe() in AICode() | 11:08 |
| null_000 | bisc: when I try it it has a 100% success rate ^_^ | 11:09 |
| null_000 | bisc: this not the only way to get this error, I get it all the time... but this way has proven to be the quickest and most certain one (at least on my computer) | 11:14 |
| bisc | null_000: After commenting this line ai clients seem to work stable. I think I have reproduced it. The pywx client freezes on downloading the new turn info? | 11:15 |
| null_000 | an error message should pop up (with no text) and the console should print an error "object isn't new enough" | 11:17 |
| *** muthusuba has quit IRC | 11:18 | |
| null_000 | which version of the client are you using? tp04? | 11:18 |
| bisc | no, haven't still got anything like that. moreover, the last time the connect was successful. | 11:18 |
| bisc | yes. it's several commits ahead of official tp04, but these commits are GUI-connected | 11:19 |
| bisc | do you recieve this error in 100% of attempts? | 11:19 |
| null_000 | are you connecting as guest? I've sometimes been able to connect with other usernames. | 11:20 |
| null_000 | yes 100% | 11:20 |
| bisc | no, I'm using "guest2" login, as the config for guest is broken (don't remember the password :)) | 11:20 |
| null_000 | this might be it | 11:21 |
| null_000 | the server returns different modification times for different users (don't know why) | 11:21 |
| null_000 | i usualy login with guest (and the password guest) | 11:21 |
| null_000 | I think that is the default password | 11:22 |
| bisc | null_000: yes, that's correct password. and I got the discussed error: The object isn't new enough to update the dictionary with! Current 2010-06-12 15:21:53, update 2010-06-12 15:21:50 | 11:23 |
| null_000 | this is it | 11:23 |
| bisc | so. if the server sends an outdated object, it seems to me that he's responsible for the bug, no? | 11:24 |
| null_000 | not sure | 11:24 |
| null_000 | since the tp03 works | 11:25 |
| null_000 | it might be something in the tp04 client... | 11:25 |
| null_000 | but if I would have to guess I would say it's the server | 11:26 |
| bisc | tp03 could have no check for such error, or something. I'll look through the code, but most probably we should ask mithro about who violated the protocol | 11:26 |
| bisc | null_000: to get your testing running again, I'd advise just using another username. I've just had 2 successful connects for "guest2" | 11:31 |
| null_000 | the problem is that I get this error after a while (say 20 turns) | 11:32 |
| null_000 | I'm now using a dirty hack that modifies the modtime in case there should be an error... this might cause some other problems but at least I can test for a while | 11:34 |
| bisc | another way to temporary avoid this issue is to comment lines 23-24 in libtpclient-py/tp/client/ChangeDict.py | 11:36 |
| null_000 | I've tried that... but I think some data got corrupt later on... | 11:37 |
| null_000 | bisc: thanks for all the help... | 11:40 |
| bisc | null_000: that's ok, haven't really helped you yet. llnz or mithro will do, I suppose. | 11:41 |
| *** muthusuba has joined #tp | 12:15 | |
| *** shenki has joined #tp | 14:19 | |
| *** bisc has quit IRC | 14:50 | |
| *** null_000 has quit IRC | 15:20 | |
| *** muthusuba has quit IRC | 15:55 | |
| *** muthusuba has joined #tp | 16:10 | |
| *** tuna-fish has joined #tp | 16:16 | |
| *** bisc has joined #tp | 16:29 | |
| *** tuna-fish has quit IRC | 16:40 | |
| *** muthusuba has quit IRC | 16:53 | |
| *** glew has joined #tp | 18:41 | |
| *** Epyon has joined #tp | 20:17 | |
| *** bisc has quit IRC | 21:24 | |
| *** verhoevenv has joined #tp | 21:41 | |
| *** verhoevenv has quit IRC | 22:12 | |
| *** Greywhind has joined #tp | 22:39 | |
| *** Greywhind has quit IRC | 23:52 | |
Generated by irclog2html.py 2.17.2 by Marius Gedminas - find it at https://mg.pov.lt/irclog2html/!