Monday, 2009-04-06

*** tpb has joined #tp00:00
*** ChanServ sets mode: +o tpb00:00
*** tansell_laptop has joined #tp00:30
*** tansell-laptop has quit IRC00:30
llnzbbl, home time00:37
*** llnz has quit IRC00:37
*** shrink has quit IRC00:37
*** shrink has joined #tp00:37
*** JLafont has quit IRC00:49
*** llnz has joined #tp01:32
Landonnash: replied to your comment01:53
nashLandon: vcool - will look shortly (in meeting)01:58
CIA-1llnz libtpproto-cpp * rc4a9b4faa9f5 /tpproto/ (object.cpp object.h):02:01
CIA-1Connect up the ObjectParameters and ObjectParameterGroups in Object.02:01
CIA-1ObjectParameters should now be usable.02:01
llnzbbs, dinner02:02
* llnz is back02:38
CIA-1nash galaxie * r546736ac8b0e /Makefile: Makefile: Fix the main make rule02:46
CIA-1nash galaxie * rf237a364d7b3 / (Makefile galaxie.c gui_window.h server.c): Don't build EWL/GUI for now.02:46
nashGee that thing is fast02:46
nashit messaged here before git had printed it had pushed02:46
llnzwhen it works, it works well02:47
nashdidi`: You areoung?02:47
nashdidi`: You around even02:47
nashdidi`: If so this tag is for you: Gui-Free-Build02:47
* llnz works on the research frame proposal02:55
Landon ah darn, jmtan isn't here :\ I don't suppose anyone would know how to use pdb with the 3d client, would they?02:55
Landonthe client sucks my focus when I hit the point where I set the trace02:56
tansellLandon, jmtan would know03:05
tansellI would recommend emailing him03:06
tanselland then putting the result in the wiki :)03:06
*** DTRemenak has joined #tp03:16
* llnz updates the research frame proposal - nearly there!03:18
llnzdo we want/need choices between prerequsites, for example (a || b) && c, currently only have 'and' preerequsite list03:20
tansellI think an OR could be useful03:20
nashOR is very useful03:21
Landon"guns" OR "diplomacy" => "world peace" :p03:21
llnzlist of (list of choices) ? wasteful in the normal case... only one item in list of choices means no choice and twice the bytes03:21
* nash is off home03:22
llnzcya nash03:22
nashllnz: Bytes are nothing03:22
*** nash has quit IRC03:23
llnzmaybe have two lists, normal prereqs and list of list of choices03:23
tansellreusing things is good :)03:25
llnztansell: the overall proposal is reusing things03:26
llnzthis is inside the technology description03:27
tanselldo you have exclusions? IE You can have A || B but not both03:27
llnzi have conflicts, can't research if have any of [x, y, z]03:28
tansellI would expect that this stuff has a lot in common with the design stuff03:29
llnzi don't really want scheme in here, because it will be run often03:29
llnzit's a much simpler problem03:30
llnznearly linear03:30
tansellrun often?03:33
llnzthe design stuff is only done on add design and modify design03:33
llnzin research, it may need to be run everytime the player requests a Research frame03:33
tansellthis stuff would only be run when a person requests a new research?03:34
llnzit may be cachable though03:34
llnzno, any research03:34
tansellwell it's not going to change03:34
llnzlike when they want to see what it currently set03:34
llnzas well as when they set their new research03:35
tansellwhich is the same as designs right?03:35
llnzno, as i said, it's only run once - when the design is added or modified03:36
llnzall later requests use the stored values03:36
llnzstored in the database03:39
tansellso why can't research the same way?03:40
llnzi suppose it could...03:41
llnzbut still no reason to use scheme03:41
llnzscheme integration is ugly and if I could drop it I would03:42
llnzif it wasn't going to be so much code I would replace guile and mzscheme with something written in c++ to calculate it directly03:43
llnza mini scheme interpreter03:43
tansellscheme has been well tested and it means we don't have to maintain yet another language03:47
llnzi don't think we need another language03:48
tanselland can be calculated both client/server side03:48
llnzapt gets on without needing to embed a programming language in each deb package03:48
llnzand that is exactly the level of complexity we are dealing with in Research03:53
tansellllnz, the apt stuff is pretty much turning complete03:57
tansellyou can write a suduko solver in it03:57
llnzin the deb metadata? depends and conflicts fields?03:58
llnzactually, i suppose you could04:00
tpb<> (at
tansellmy theory on config options is to embed a "real" language now04:13
tansellotherwise you'll end up keep adding features until you have some horror of a real language anyway04:14
llnzwell... that's fine for python, perl and sh04:15
JLPgood morning everyone04:32
llnzhi JLP04:33
* llnz ponders06:12
*** BCarlyon has quit IRC06:33
* llnz wanders off07:00
llnzlater all07:00
*** llnz has quit IRC07:00
*** JLafont has joined #tp07:20
*** ruzette has joined #tp07:32
*** ruzette has quit IRC07:38
*** peres has joined #tp08:27
*** jmtan has joined #tp08:39
Landonhey jmtan, I'm about to leave for class in a few minutes, but I was wondering how you used pdb with the 3d client08:43
Landonwhen I hit a set_trace, all input except the mouse stops working _everywhere_08:44
jmtandid you try alt-tab? sorry it seems the input capturing has different behaviour depending on your os08:44
jmtanwhat os are you running on?08:45
jmtani'll email you, gotta go for dinner too08:46
LandonI'm running ubuntu08:47
Landonalt-tab works to get out of it before it starts the trace08:47
Landonbut after, nothing08:47
jmtanwhich window manager are you using?08:47
jmtanthis is a weird solution, but you can try switching workspaces08:48
LandonI'm just using gnome right now08:50
Landonyeah, I take advantage of that often, but it doesn't work to take focus away (even when it';s running normally)08:50
LandonI've tried going to a tty and back as well08:50
jmtanare you using compositing?08:52
LandonI need to be going now, I'll check back after class08:56
*** verhoevenv has joined #tp09:03
*** didi` has quit IRC09:11
*** bddebian has joined #tp10:01
*** sanchan89 has joined #tp10:14
*** mithro has quit IRC10:19
*** sanchan89 has quit IRC10:24
*** sanchan89 has joined #tp10:35
*** peres has quit IRC10:38
*** mithro has joined #tp10:41
sanchan89I am having some trouble downloading the tp-pyogre-client for windows. It says after sometime that the source file could not be read and quits10:47
jmtanhi sanchan, could you state the error message?10:48
sanchan89I used the firefox default dloader. It didn't detect the file size but somewhere around 15MB later, it gave the message "source file cuold not be read" and aborted10:50
jmtanit seems that the download must have cut off somewhere, the full zip should be around 22 megs10:51
*** peres has joined #tp10:51
jmtani mean, the setup exe10:51
jmtancould you try redownloading the file?10:52
sanchan89O! The part file is 20M. I tried a couple of times again. Using the DTA plugin, it resulted in Network TimeOut Exception. Something similar happend a couple of days back. I'll try again10:53
*** sanchan89 has quit IRC11:02
*** sanchan89 has joined #tp11:02
*** sanchan89 has quit IRC11:07
*** mithro has quit IRC11:26
*** Landon has quit IRC12:08
*** Noughmad has joined #tp12:11
*** vivri has joined #tp12:46
*** jmtan has left #tp12:46
* vivri used to be vi1985.12:46
verhoevenvHeya vivri.12:52
* verhoevenv used to be Iwanowitch12:52
vivriverhoevenv: hey! just dropped in to say hi, havent been here in a while12:57
ezodhey vivri12:59
vivrihi ezod.12:59
*** peres has left #tp13:37
*** didi` has joined #tp13:53
*** verhoevenv_ has joined #tp14:34
*** vivri has quit IRC14:39
*** verhoevenv has quit IRC14:56
*** jernejovc has joined #tp15:05
*** Erroneous has joined #tp15:13
*** Landon has joined #tp15:26
*** alanp has joined #tp15:45
alanphow are you?16:06
didi`alanp: Fine.16:08
didi`alanp: And you?16:08
didi`alanp: Actually, I am going away for a while. See you later.16:09
alanppretty good16:10
alanpi'll see you later16:10
*** llnz has joined #tp16:13
llnzmorning all16:23
ezodhey llnz16:24
llnzhi ezod16:24
ezodi'll get on the libtpproto-cpp doc soon16:25
*** Noughmad-laptop has joined #tp17:01
*** alanp has quit IRC17:16
*** jernejovc has quit IRC17:23
*** bddebian has quit IRC17:30
*** alanp has joined #tp17:30
*** Noughmad-laptop has quit IRC17:51
*** Noughmad has quit IRC18:09
*** alanp has quit IRC18:20
*** jnengland77 has joined #tp19:00
*** nash has joined #tp19:02
didi`nash: You commit did the trick. It now compiles.19:18
nashdidi`: :-)19:19
nashIt's essentially in library form19:19
nashAs an aside however... You really need to convince about Neural nets... or change your strategy19:19
didi`nash: Hummm... Really? Maybe I am not succeeding being clear.19:22
nashmay well be...19:24
didi`nash: Well. Help me doing it better. :-)19:24
nashOkay, I'm not an expert on the field by any stretch of hte imagination.19:25
didi`nash: But it would be great, isn't it? ;-)19:26
nashBut the number of inputs & outputs scares me.19:26
nashCan you control that? and keep computation power down to a reasonable level?19:26
*** bddebian has joined #tp19:26
nashYes, but will it work is a more fundamental question ;-)19:26
nashdidi`: We'd love you to do it in your own time... but for SoC we need to see something we know will work19:26
didi`nash: Like I stated in the proposal, this is a difficult game to predict. I think no one can really say if something you take a long or short time. With AI it is maybe even worse.19:28
nashdidi`: I don't think you've nailed down a game either?19:28
didi`nash: I know it's a risky game. But I really think it is worth.19:28
didi`nash: Sorry. "nailed down"?19:28
nashdidi`: But ballpark... minisec normally has say 100 stars, with 2 types of build objects, 4 players say, and say 100 ships... thats 100000 combinations striaght way19:29
nashWhich ruleset are you targetting?19:29
didi`nash: Minisec.19:29
didi`nash: It is not the combinations who matters but what data you will use. Like say, you will not eat each one of the 100 ships, but you will say "player B has 100 ships". That is one data, not 100.19:31
nashBut position has to be taken into account19:31
nashAnd type of ship19:32
didi`nash: Not necessary.19:32
nashYou need to react differently to a 100 battleships then 100 frigates19:32
didi`nash: It _could_ be.19:32
didi`nash: Yes, probably the types of the ships is a better data than its position.19:33
didi`nash: Maybe, just a quadrantic of the map is important, not all of then.19:33
didi`nash: Passed, like say, 20 turns of flight, I am no more interest.19:33
nash100 ships engaging 100 enemy ships is not important; 200 ships (2 different fleets) converging on homeworld are different19:34
didi`nash: Indeed. But you have little data here too: X fragates, Y battleships, Z scouts coming. That could be it.19:35
nashdidi`: You need to put this up on a wiki, and start to give a better plan19:36
didi`nash: Ok.19:36
nashProblem is there are a _lot_ of questions to answer.  And you have only done some.19:37
nashAnd I mean a lot19:37
didi`I will happily answer them.19:37
nashBasically an adventurous proposal needs strong support19:37
nashdidi`: But you need to start going:  Here are the inputs19:37
nash"Extraordinary Claims need Extraordinary Evidence"19:38
nashThis will include things like estimates of size and computational time.19:38
didi`nash: The last thing will be a problem. I know AI. It is not a thing that you can predict this way. But I will try my best.19:40
nashThen you need to explain why... and why it WON'T be a problem19:40
didi`nash: Ok.19:40
nashIf your AI takes 2 hours of computation to generate a turn in the simplest game... it's essentially useless19:41
cherezUnless the game is Diplomacy. :P19:46
nashcherez: If it takes 2 hours to compute a new move in diplomacy there is a problem...19:50
cherezI wish I'd gotten to play Diplomacy with you rather than the guys I usually played with.20:11
zzorndepends on the ai20:17
zzornif the search space is too large, heuristics are generally used20:17
*** nash has quit IRC20:23
*** nash has joined #tp20:24
*** mithro has joined #tp20:27
mithrotote: ping?20:27
mithroGreywhind: ping?20:28
Greywhindmithro: pong20:35
mithroGreywhind: so still waiting for a response to the GSoC comments20:35
Greywhindmithro: i'd already responded to all but the last one by updating the page20:35
Greywhindthe last one, i've been having trouble trying to figure out how to get my client to connect to my local server20:36
Greywhindand due to school work, i haven't had much time at all to look into it20:36
Greywhindbut i might be able to look now20:36
Greywhindany idea if this could cause problems:20:44
tpbTitle: Nopaste - No description (at
tansellGreywhind, you should be able to run it without installing20:46
Greywhindi am running it without installing20:46
didi`nash: Is there a special place to create the wiki page at the TP wiki?20:46
Greywhindwhat's the default port for the server?20:46
tansell6923 i think20:48
tanselldidi`, depends what for20:48
didi`tansell: a proposal20:48
tanselldidi`, if it's a gsoc proposal, put it under the Google_Summer_of_Code/Proposals200920:49
didi`tansell: Thanks.20:50
tpb<> (at
Greywhindyeah, it's still saying it can't connect to the server, and i've tried using the specific port, my internal ip,, and opening my firewall20:51
Greywhindwhat version of tpserver-cpp should I be running for tp0420:52
tansellGreywhind, if you have that error20:52
tansellit means it has not start up the network yet20:52
tansell./tpserver/tpserver-cpp -d -C ./quickstart-sample.conf20:52
Greywhindgives the same errors about guile and libtpmzscheme20:52
Greywhindwhich is odd, since i know i installed one of the two after it refused to compile20:53
Greywhindlet me try recompiling, see if i missed something20:53
tansellGreywhind, you clone from git right?20:53
llnzquickstart-*.conf don't work unless the server is installed20:53
llnzyou will need to edit them to make them work without installing20:54
Greywhindi can install, i guess20:54
tansellllnz, oh?20:54
llnztansell: because of where the library/modules are20:55
llnzthe quickstart files assume tpserver-cpp will find them, which it can't until they are installed20:55
Greywhindok, installing it fixed the error. thanks, llnz20:57
Greywhindstill says it can't connect to the host, though20:58
*** alanp has joined #tp21:01
tpbTitle: Nopaste - No description (at
ezodhey alanp21:02
alanpwhat's up?21:02
alanpI immediately regret my decision not to go to Detroit21:02
*** verhoevenv_ has quit IRC21:02
ezodbig party?21:02
Vadtecevening all21:04
ezodGreywhind: i may be missing something here, but could you use the single player stuff to do your server config and connect the client to it?21:04
*** konr has quit IRC21:05
Greywhindno, ezod - i'm on the tp04 branch21:08
Greywhindsingleplayer hasn't been merged into it yet21:08
ezodsingleplayer didn't touch too many files, maybe we could merge them?21:09
Greywhindit's definitely on my to-do list21:09
Greywhindbut i'd prefer to see what the current state is first21:09
Greywhindif that's possible21:09
Greywhindso i know if i break anything :P21:09
llnzGreywhind: what is quickstart-sample?21:19
Greywhindnot sure, it's a default config21:20
Greywhinddoes the same thing if i run quickstart-risk-random.conf though21:20
Greywhindmaybe i'm running the wrong server version?21:23
Greywhindwhat's the one with tp04?21:23
mithroit should just be a21:24
mithrogit clone git://
Greywhindok, that's what i did21:24
Greywhindare there some confs that don't use tp04? maybe i'm running one of those?21:25
mithrocd tpserver-cpp; ./; ./configure --prefix=/opt/tpserver-cpp; make -j3; sudo make install21:25
mithroGreywhind: this is not a tp04 thing21:25
mithrocd /opt/tpserver-cpp/bin21:25
mithro./tpserver-cpp -d -C ~/tpserver-cpp/quickstart-blah.conf21:25
Greywhinddid you take a look at the log of the tpserver run i put up?21:26
tpbTitle: Nopaste - No description (at
*** heineckm has joined #tp21:26
mithroGreywhind: so that looks like it almost worked21:35
mithroGreywhind: there should be more stuff about loading a ruleset and stuff21:35
mithroGreywhind: your not using sample.conf right?21:35
mithroyou are using the quickstart config?21:35
Greywhindmithro: i was using quickstart21:36
Greywhindi also ran it with quickstart-risk-random.conf and the same result21:36
mithroGreywhind: that is all the output you are getting?21:37
mithroit's -C21:37
mithronot -c21:37
Greywhinddamn :P21:37
mithrollnz: why was it not complaining about not finding a config file?21:38
Greywhindthat seems to have made a difference21:38
Greywhindyep - now i can connect21:38
Greywhindthanks :)21:38
Greywhindit seems to be showing all the planets and wormholes correctly on the starmap21:39
tansellGreywhind, so that probably means you are not actually using the tp04 branch :)21:40
tansellif you check the output21:40
Greywhindtansell: heh. i probably ran the wrong client21:40
tansellit'll tell you what version the client/libraries are using21:40
Greywhindtansell: yep, switched to the wrong dir. silly me.21:40
Greywhindat least now i know what the map should look like21:41
Greywhindok, yeah, it doesn't show any objects on the map21:42
Greywhindbut it has all of them in the list on the right21:42
tansellI need to check I don't have any unpushed changes21:42
tansellbecause I think i managed to get something to display on the map21:42
Greywhindmessages appear to work21:43
Greywhindthere are no images for systems or planets21:43
Greywhindit seems that the "you have no objects" dialog is broken - it comes up when i do have objects21:44
Greywhindno orders panel comes up when i select a planet i own21:45
Greywhindobjects without orders window doesn't show anything either21:45
Greywhindso i'm guessing the client is not detecting object ownership at all21:45
tansellsounds about right21:45
Greywhindsearch for objects by name works21:46
Greywhindmessage filters seem to work still21:47
Vadteccan anyone recommend a book on C++ that covers things like the std container classes (deque, list, map, etc) and C++ generally? the only C++ book i have is from 1994 so it doesnt cover anything useful21:49
tansellGreywhind, messages didn't change21:51
Greywhindwell, that does explain why they work perfectly! :P21:51
llnzmithro: because he set it with -C21:56
Greywhindi've added a comment about the errors and problems that are evident so far21:56
Greywhindllnz: it didn't warn that the config was not found even when i tried incorrectly to set it with -c21:57
llnzGreywhind: odd21:57
tpbTitle: Nopaste - No description (at
llnzVadtec: there is a good STL reference online21:57
Greywhindtansell: i have to go for now21:58
mithroVadtec: there are multiple good ones online21:58
tpbTitle: Standard Template Library Programmer's Guide (at
Greywhindi'll be back21:58
llnzGreywhind: odd, i'll have a look21:58
tansellwell heading to lunch21:59
Greywhindthanks for helping me get the server to work :P21:59
Greywhindi feel stupid for not noticing the capital C21:59
Greywhindoh well21:59
Vadteci know there are good ones online22:01
Vadteci just prefer books for when im on the go and dont have access to the net22:01
Vadtecand yes, ive been using the sgi site22:02
llnzVadtec: i have "C++ from the Ground Up" in front of me here, not too bad, barely covers stl though22:04
llnzactually.... it does cover stl, i hadn't noticed that22:07
Vadtecpass me the ISBN22:09
Vadtecill see if my local hastings has a copy22:11
llnzby Herbert Schildt22:11
Vadtecgood author22:11
ezodyeah, i think i originally learned c++ from a herbert schildt book22:11
ezodit was designed for people who are already good at c, pretty cool book22:11
*** paulmcnett has joined #tp22:27
llnzhi paulmcnett22:34
*** mithro has quit IRC22:39
*** mithro has joined #tp22:40
*** Erroneous has quit IRC22:46
jnengland77hi all22:53
llnzhi jnengland7722:53
jnengland77mithro: neat google talk.22:55
tpbTitle: YouTube - Gaming For Freedom (at
jnengland77I watched it the other day, but have been quite busy.22:59
mithrojnengland77: that is quite old now23:13
jnengland77Yeah I know. It was still interesting.23:14
jnengland77I liked the Austrian bit. :P23:16
*** bddebian has quit IRC23:37
*** tansell__laptop has joined #tp23:49
*** tansell_laptop has quit IRC23:49

Generated by 2.13.1 by Marius Gedminas - find it at!