Thursday, 2007-03-29

*** tpb has joined #tp00:00
*** ChanServ sets mode: +o tpb00:00
nashSo this surprise is?00:04
mithronash: when I clean it up a little, i'll show you00:05
* mithro laughs evilly00:11
mithroo btw, I figure out the very strange error you where seeing sometimes with tpclient-pywx00:11
mithroI also figure out why it was being so slow00:11
nashOh good00:11
* nash assumed slow was: a) serial requests & b) python00:12
mithroi had a time.sleep(0.1) in the wrong place00:12
nashUnlike tpe which is a) Hammer the server with everything you damn well can, and b) C00:12
nashWhy is there a sleep at all?00:12
mithrowell, it's probably 10x slower00:12
mithronash: to stop it hammering the CPU while waiting on socket data00:13
nashWhat?  No select()?00:13
nashAnd I often get a "resource unavailable" message when running00:14
nashLots of them00:14
mithronash: that means the client tried to get more data then the socket had ready00:15
mithroselect only tells you data is ready, not how much is there00:16
nashYeah, but read() with return what is there... Why are you getting an error?00:16
mithrohrm...00:17
mithrothere was a reason, but I forget it now00:17
* mithro ponders00:18
mithroI think it was quicker to just let the thing fail instead of checking if data existed first00:19
* nash notes console output is generally devestatingly slow00:20
mithrothe server side uses select00:20
mithronash: they should only output when debug mode is on00:20
nashWhich is hte default?00:20
mithrooff (I think), but tpclient-pywx turns it on00:21
nashYou should fix that.  Any 'known' debug messages are a waste of time.  They just get in the way of real errors00:23
mithronash: i should really be checking the type of that error - IE only output when it's not a socket doesn't have data error00:23
mithronash: there is a lot of things I should really do :P00:26
* nash cares not - it's essentially spam00:26
mithroxstruct should have way better error reporting00:26
mithronash: lots of things to do, so little time :P00:28
mithronash: I would appreciate feedback on the darcs tutorial thing - I'm not happy with the formatting in that email - think i'll convert it to a proper page00:28
* nash hasn't read it yet00:28
mithroon a scale of 1 to 10 I rate it about a 4 :P00:30
*** miChou has quit IRC00:30
cherezdarcs tutorial?!00:39
nashcherez: Personally I think the first line should be "install a revision control system people actually know"... but anyway00:40
mithronash: darcs is about as known as git00:40
* nash disagrees with that00:40
nashhttp://www.googlefight.com/index.php?lang=en_GB&word1=git+&word2=darcs+00:41
nashAs does googlefight ;-)00:42
tpb<http://ln-s.net/M5D> (at www.googlefight.com)00:42
chereznash: git is also British slang00:43
nashcherez: Why does everyone tell me that?00:44
nashcherez: It's also australian slang.00:44
nash:-)00:44
cherezI learned it on the BBC00:44
cherezdon't watch so much Australian TV00:44
nashcherez: Sorry - I just keep getting it told that by people as if it is unusual info...00:44
cherezdarcs isn't slang as far as I know00:44
nashcherez: There isn't much australian tv00:44
nashcherez: Red Dwarf being the canonical reference for the word00:45
cherezand I have seen every episode of Red Dwarf00:45
mithronash: in Americas and Europe, they wouldn't know about git :P00:45
mithronash: fairer fight http://www.googlefight.com/index.php?lang=en_GB&word1=darcs&word2=cogito00:47
tpb<http://ln-s.net/M5I> (at www.googlefight.com)00:47
nashmithro: I can sort of understand it online - but offline too... Besides Linus is a European who lives in America, and he knew exactly what is meant00:47
mithronash: ahh well :P00:47
nashmithro: So cogito, a front end to git does a 3-2 on darcs...00:48
mithroahh well00:48
* nash jsut looked at the list of projects using darcs...00:49
mithrohrm, turns out not to be that fair - cogito is the name of a few companies00:49
mithrohttp://www.google.com/trends?q=cogito%2C+darcs00:50
tpbTitle: Google Trends: cogito, darcs (at www.google.com)00:50
* nash notes a large % are haskell projects00:50
mithronash: ahh well, tailor makes the problem not that huge00:51
nashSo is there a solution for darcs on windows?00:51
cherezcogito is also a fairly common Latin word00:53
mithronash: there are precompiled binaries which work to some degree00:54
* nash had assumed mithro would have chosen a vcs that worked on all the platforms he wanted to target...00:54
cherezhttp://www.googlefight.com/index.php?lang=en_GB&word1=darcs+version+control&word2=git+version+control00:55
tpb<http://ln-s.net/M5S> (at www.googlefight.com)00:55
mithronash: well it does work under windows - you just have to use the command line interface00:55
nashcherez: The adds are great... One for Bazaar, and one for CVS00:55
mithrohttp://www.googlefight.com/index.php?lang=en_GB&word1=darcs&word2=git+version+control00:56
tpb<http://ln-s.net/M5T> (at www.googlefight.com)00:56
mithro:@00:56
mithros/:@/:P00:56
nashmithro: Try darcs vs git cogito00:56
mithronash: darcs win that one :P00:57
nashmithro: my point exactly00:57
mithroahh well, as I said, tailor makes this not a huge problem00:58
cherezI put version control on both as I figured that would filter out most Red Dwarf discussions and I would think they'd both mention version control on about the same fraction of sites00:58
mithroif cogito becomes "better" then darcs, we just convert it00:58
cherezis this tutorial online?00:59
nashActually one area where both git and darcs suck - is new user doco :-(01:00
mithrocherez: it's in the forums or on the mailing list01:04
cherezalright, I'll read it01:04
cherezI still haven't figured out how to generate patches01:04
cherezI was hoping to fix the Python server up at least a bit, and submit the pywx patch01:05
cherezI figured out my problem01:05
cherezI have 2.4 and 2.6 installed, and it defaults to 2.4 unless 2.6 is selected before importing01:05
nashOhh... patches ;-)01:05
mithrocherez: yay!01:05
cherezso I was going to import wxversion and specify 2.6 inside a try01:06
cherezoh, and pyOpenSSL01:06
cherezit doesn't detect it on either of my systems01:06
mithrocherez: ahh01:06
cherezI think it's trying "import pyOpenSSL"01:06
cherezbut the package is named OpenSSL on both my Ubuntu and Gentoo boxes01:07
cherezso a "try: import OpenSSL as pyOpenSSL" might work01:07
cherezI've been too busy with my mother to try it though01:08
cherezand I will probably be computerless after tonight until saturday or sunday01:08
mithrocherez: :/01:08
cherezmy mother requires too much babysitting for me to do homework and TP01:09
mithrocherez: well, hope you have better luck with your mother01:09
cherezthanks01:09
cherezI promised my girlfriend we'd see a movie before I go back to the dorm, so I can't leave yet01:10
mithrobblr01:16
cherezthe library!01:22
cherezI can go to the public library to work on TP and print docs to read01:23
cherezwoohoo, I get to hack code with Notepad over Explorer FTP01:23
* nash thinks hacking code with a 2B may be easier then that01:27
cherezhacking by dial-uping with a telephone would be easier01:58
*** matthewd has quit IRC02:00
*** matthewd has joined #tp02:01
*** mithro has quit IRC02:40
*** mithro has joined #tp02:40
cherez'lo, mithro02:44
tpbaloril has quit worldforge (Ping timeout: 620 seconds)02:50
mithronash: ping?02:52
nashyou rang?02:53
mithronash: here comes the suprise :)02:53
tpbaloril has joined on worldforge02:54
nashaloril has joined worldforge is not what i expected...02:55
* mithro uploads02:55
mithrolousy slow wireless :/02:55
* nash notes impact is lessened02:55
tpb<[email protected]> hehe02:56
mithroyeah, maybe I should have waited for it to upload first :P02:56
CIA-17[email protected] * tpsai-py/ (50 files):02:56
CIA-17Big update.:02:56
CIA-17 AI now uses a "task method". This means it creates a list of tasks which need to02:56
CIA-17 be done and then assigns assets to do the tasks. This stops the AI from02:56
CIA-17 assigning two colonisation fleets to the same planet.02:56
CIA-17 AI still has no concept of time. It will rewrite all the orders every turn. This02:56
CIA-17 is bad, but oh well.02:56
CIA-17[email protected] * tpsai-py/ (25 files):02:57
CIA-17AI Actually does something.:02:57
CIA-17 AI connects and issues orders.02:57
CIA-17 Only builds frigates, and then only works if it was the first player connected.02:57
CIA-17 (Not smart enough to send a probe and figure out which is actually frigate.)02:57
CIA-17 Does stupid things like sending multiple coloniser ships to one location (as has02:57
CIA-17 no concept of future action).02:57
mithronash: there you go :P02:58
mithroshould give smith a run for is money02:58
mithros/is/his02:59
nashCool03:00
nashDoes it handle 10 second turns?03:00
* nash notes he is going to be up late adding a few new features03:00
nashFirst question: Are there any other entries?03:01
mithronash: yes03:01
nashCool03:01
nashhow many?03:01
mithroalthough unlike smith, he isn't a persistant AI03:01
mithroIE you run him once for each turn03:01
mithroshould be pretty easy to add persistance support however03:02
mithronash: nobody has said anything about submitting an entry yet03:02
mithroso I thought we should atleast give smith something to crush :)03:02
nashmithro: I need to add my "build warships" patches then ;-)03:03
nashAlso I'm deciding how defensive to be...03:03
nashAnd whether I should write to the rules, or write for a general game03:03
* nash prefers to write for the game...03:03
nashAlso, one thing I have is the ability to have multiple AIs in the binary - you just use --ai=<AI Name> to select03:04
nashSo if I want to submit two AIs - do I need to do the whole damn thing twice?03:04
mithrothe tpsai-py is very simple in how it works03:04
nashSo is smith03:04
mithroit creates a list of "tasks"03:04
nashAnd does them03:05
mithroand then works out which assets can be used to fullfil the tasks03:05
mithroand then issues the orders03:05
nashSmith is even simpler - he just watches for interesting things, and reacts03:05
nashInteresting things are:03:05
nash - Objects which he can order03:05
mithroi will be intrested to see which one works better :P03:06
mithroi'm a bit disappointed because it shows how easy it is/was to write an AI03:06
nashWell smith needs to have some concept of things to do03:06
nashmithro: Indeed - most of my work was gui stuff03:07
mithroyet nobody else seems to have taken up the challenge03:07
nash 302  891 6885 ai_smith.c03:07
nash  39   98  817 ai_util.c03:07
nash 341  989 7702 total03:07
mithro[email protected]:~/oss/tp/tpsai-py$ wc -l *py03:11
mithro  173 things.py03:11
mithro  272 tpsai-py03:11
mithro  445 total03:11
nashI've written just shy of 5000 lines of code... 350 for the AI ;-)03:11
mithroanyway I have a meeting an luck03:11
nashOkay 0 talk to you soon03:12
mithrowe can have battles on the weekend after you submit :P03:12
nashYou on this afternoon?03:12
nashmithro: Love to03:12
nashhappy to this afternoon if you like03:12
mithronash: well i have things I should be working on  :P03:13
nashokay - just wanted to ask you some questions this afternoon - not play AI games03:13
CIA-17[email protected] * libtpproto-py/tp/netlib/objects/Object.py : Object now has a slightly better repr method.03:14
CIA-17[email protected] * libtpproto-py/tp/netlib/common.py :03:14
CIA-17Remove a sleep which was causing slowness.:03:14
CIA-17 This sleep was triggering when there was no data on the socket, however there03:14
CIA-17 was still frames which had not been fully processed yet. This makes the library03:14
CIA-17 a little less friendly however.03:14
CIA-17[email protected] * libtpproto-py/tp/netlib/common.py :03:14
CIA-17(Bad) Fix for a problem nash was having.:03:14
CIA-17 When getting order descriptions you get a recursive call which can cause keys to03:14
CIA-17 disappear. This should make sure it no longer crashes in this case.03:14
CIA-17[email protected] * libtpproto-py/tp/netlib/ (xstruct.py xstruct.py xstruct.py xstruct.py):03:14
CIA-17Added better error checking and reporting.:03:14
CIA-17 Now check that you arn't trying to send a negative number to an unsigned type.03:14
CIA-17 Errors are now a bit more informative (instead of things like "Can't mask03:14
CIA-17 string").03:14
CIA-17[email protected] * libtpproto-py/tp/netlib/ (client.py objects/Order_Insert.py):03:14
CIA-17Fixed the second method of adding an order.:03:14
CIA-17 The connection.order_insert(100, -1, BUILD_TYPE, *arguments) now works. Every other03:14
CIA-17 client uses the connection.order_insert(100, -1, Order()) method.03:14
CIA-17[email protected] * libtpproto-py/tp/netlib/objects/OrderDesc.py :03:14
CIA-17Raised a better error.:03:14
CIA-17 This will let new developers figure out the IS for a String problem quicker.03:14
nashmithro: So yeah - are you going to be online - even if doing other things?03:15
nashAlso did you see my email to the tp-devel list?03:15
nashAbout URI schemes03:15
nashAnyway - off to lunch03:17
nashback soon03:17
mithronash: replied to your email04:10
nashmithro: Cool04:26
mithronash: that all okay?04:44
nashnot quite - I am replying now... doesn't affect me at the moment - but I see some issues04:44
*** zipola has quit IRC04:57
cherezman, it is hard to write a decent goodbye note05:06
cherezthat must be a major suicide deterrent05:06
dmpayton_WoWcherez: Hows it coming? =/05:07
cherezalmost done05:08
cherez"I've been concerned about your drinking for a long time and it has become a horrible problem for both you and those around you. I and others have tried to help you, but I truly don't believe you want to be helped. I think it is time to take control of my life and not let your drinking affect me, so I have decided to leave until you seek help and overcome this problem. I love you and don't want this to drive us apart, but I see no other way."05:11
mithrocherez: is there a local alchohol victim support group?05:18
cherezI don't know, I'll look into it05:21
mithrobecause I would highly recommend talking to them05:22
*** Porthos_ has joined #tp05:23
*** Porthos_ has left #tp05:23
*** Porthos_ has joined #tp05:23
mithrocherez: there are plenty which have nothing to do with religion as well (if you prefer that)05:29
jothamhaha @ suicide note comment05:36
*** tpb has joined #tp06:00
*** ChanServ sets mode: +o tpb06:00
cherezwell then, I'm going06:27
cherezI guess I'll see you guys in a few days06:28
cherezthanks06:28
jothamgood luck bro06:28
cherezthanks everyone06:28
*** cherez has quit IRC06:30
dmpayton_WoWMan, I hope everything works out for him.07:08
nashyeah07:14
nashhope so too07:14
nashANyway - be back later07:15
*** nash has quit IRC07:15
mithrowell i'm heading home07:38
*** mithro has quit IRC08:00
*** nash has joined #tp08:49
* nash waves08:51
*** sijmen has joined #tp09:10
sijmenmorning!09:10
*** tuna-fish has quit IRC09:14
*** tuna-fish has joined #tp09:16
nashEvening09:23
daxxarMornin' :)09:45
nashSeems I'm out voted09:45
sijmenyay daxxar09:45
* nash informs planet the sun is in the wrong spot...09:45
sijmenhmm09:46
sijmenmaybe you should file a bug09:46
daxxarnash: Let me know when the bug is resolved09:46
nashThe but?09:46
nashThe bug?09:46
sijmenI didn't know the solar system generator was done though09:46
daxxarnash: The sun-bug.09:47
nash:-009:47
daxxarAnyway, that thing with cherez, was that his mother?09:47
sijmenhuh?09:48
daxxarReal shame, hope it works out. :|09:48
nashdaxxar: I believe so09:49
nashYeah09:49
daxxarCan't be easy..09:49
sijmenWhat happened?09:49
daxxarIt seems she has a alcohol problem.. :(09:50
sijmenand.. his mother?09:50
daxxarYes, she's the one who has it09:51
sijmenooh09:51
sijmen:|09:51
daxxarBut I'm not really familiar with it, I've just seen him talk about it, so it's kinda out of my place to talk about it09:51
daxxarLet's just give him our thoughts and hope it works out. :-)09:52
sijmenyah09:52
daxxar(or prayers, if you're into religion)09:52
daxxarWell, I'll be heading to a project meeting. :-)09:56
sijmenthis late? :o09:56
daxxarHave a pleasant morning / day, sijmen / nash.09:56
daxxarIt's 11.59AM ;)09:56
sijmenheh ;)09:56
sijmengbye09:56
nashTalk to you later09:56
*** jotham has quit IRC10:04
*** jotham has joined #tp10:04
sijmenwhat's that "Remote closed the connection" thing?10:05
sijmendoes it mean you closed your bash/ssh session or something?10:05
nashsijmen: Or the client10:34
sijmenokay10:47
*** llnz has joined #tp10:53
llnzthat was lucky10:54
sijmenso am i10:54
llnzinternational traffic from my ISP is *very* slow and unreliable today10:55
CIA-17Lee Begg <[email protected]> * tpserver-cpp/tpserver/player.cpp : Update copyright date in player.cpp10:58
*** dmpayton_WoW is now known as dmpayton11:31
nashraid done?11:32
llnzthe tpserver-cpp on llnz.dyndns.org (ie, my testing server) has done 10153 turns11:52
llnzabout 70.5 days11:53
nashnot bad11:53
llnz(it runs with persistence, because it gets restarted quite often)11:53
nashAhh11:54
nash70 days up would be impressive11:54
nashStill there must be a lot of cruft in that universe11:56
llnzsee the metaserver11:56
llnz14 players, 63 objects (no generated universe except of three star systems)11:57
CIA-17Lee Begg <[email protected]> * tpserver-cpp/ (83 files in 3 dirs):11:57
CIA-17Shifting turn processing from tpserver to rulesets:11:57
CIA-17 This gives rulesets great flexibility about how the turn is processed. Both11:57
CIA-17 MinisecTurn and MTSecTurn will get rewritten to make the most of it. Also made11:57
CIA-17 slight improvement to reducing magic numbers.11:57
nashQuestion: Where is the game name supposed to be in the url over http?11:57
* llnz really doesn't know11:58
nashmithro's email earlier was contradictory11:58
* nash notes games really need to be sorted out11:58
llnzi think it should be at the end (host:port/gamename)11:58
nashBut then how do you do http://host/path/to/server/ ?11:59
llnzyou don't11:59
llnzit's not a url, it's a uri11:59
llnzthe actual request in http is for "/<12 random hex chars>"12:00
llnzor something like that12:00
nashTrue12:01
llnzhttp://www.thousandparsec.net/tp/dev/documents/protocol3.php#HTTPTunneling12:01
tpb<http://ln-s.net/M:4> (at www.thousandparsec.net)12:01
llnzwe passed 80k lines of code today (or yesterday)12:04
* llnz notes that about 12k is libltdl12:05
nashouch12:05
llnzsilly distros, if they hadn't "fixed it" then i would need to include it in the tpserver-cpp tree12:07
*** mithro has joined #tp12:08
nashIndeed12:09
llnzhi mithro12:10
mithrohowdy12:10
sijmenhi12:11
mithrosup?12:13
nashWhat is the default port again?12:14
mithro692312:14
nashIs that something?12:14
llnzmy /etc/services doesn't list anything for that port12:15
nashokay12:16
sijmenhow was that number chosen?12:16
nashsijmen: I think that was the way12:17
llnzi chose it12:17
llnznot exactly randomly12:17
sijmenthen how did you came up with it?12:17
llnz6900's is fairly common for games12:18
llnzand 23 is my lucky number (birthday)12:18
llnz:-)12:18
sijmenhah okay12:18
llnzi think i did check /etc/services before telling everyone though12:19
* llnz notes his email hasn't reached the devel list yet12:21
* llnz wonders if mithro has read the cvs list yet12:23
nashBTW: You should register 6923 and 6924 with IANA12:23
mithrollnz: reading now12:23
llnznash: my email to devel (about an hour ago) suggested that12:23
llnzseems my isp isn't doing well internationally today12:24
mithrollnz / nash: feel free to do it :P12:24
mithrollnz: you see my commits?12:24
llnzmithro: i skimmed them, why?12:24
* mithro points to the tpsai-py :P12:25
llnzah, cool12:25
nashbbs12:25
mithroit should give nash's AI a run for it's money12:25
llnz:-)12:27
nashmithro: I still wish you'd make the deadling 23:59 on Saturday12:28
nashI'm out tomorrow12:28
mithronash: aah well :P12:30
* nash had assume it was saturday night12:31
mithrohrm, oh well12:31
mithronash: well atleast we got two AI out of the AI Comp :)12:32
mithrodmpayton: http://freshmeat.net/projects/ajaxstub/?branch_id=68789&release_id=25048512:34
tpb<http://ln-s.net/M:Q> (at freshmeat.net)12:34
nashmithro: You also have an extra partial client, a couple of new platforms (Soon), and another infrastructure library for building AIs12:35
nashAnd a new developer ;-)12:35
mithroyeah12:35
mithronot a total loss12:35
llnznot a total loss by a long way12:38
* llnz ponders what to do next12:38
nashllnz: tp0412:38
llnznash: i was thinking about that12:38
nashBTW: Is there a url scheme for the metaserver>12:38
nashSo I can do the equiv of ./client <metaserver URI to browse>12:38
mithrourl scheme?12:38
sijmenmaybe there should be a metaserver browser.. maybe using some kind of metaserver.12:38
nashSo in tp04 you can talk toa metaserver to get a list of games it knows about12:39
llnzmithro: he means something like "tpmeta+http://"12:39
nashI want my client to be able to handle that...12:39
mithrohttp://metaserver.thousandparsec.net/12:39
tpbTitle: Thousand Parsec : Metaserver Server Listing (at metaserver.thousandparsec.net)12:39
nashYeah12:39
nashNo, thats a human interface12:39
nashI mean client interface12:39
sijmenHey maybe that's not a bad idea. Metaservers referencing each other so you'd get a mesh of metaservers12:40
mithrohttp://metaserver.thousandparsec.net/?action=get12:40
nashmithro: Ugly... and not quite what I mean...12:40
nashI want something like tpm://hostname/12:40
nashm for meta here12:40
mithronash: feel free to suggest something12:41
nashI shall have to12:41
mithrometaserver was only designed/setup in a few hours12:41
nashYes, but tp04 has a lot of features in it12:42
nashTo handle servers12:42
nashAlso need to start looking at a proper game setup protocol12:42
nashA lobby ;-)12:42
*** mithro has quit IRC12:42
*** llnz has quit IRC12:42
*** jotham has quit IRC12:42
*** tuna-fish has quit IRC12:42
*** sijmen has quit IRC12:42
*** Demitar_ has quit IRC12:42
*** CIA-17 has quit IRC12:42
*** fizzle_ has quit IRC12:42
*** Demitar has quit IRC12:42
*** DystopicFro has quit IRC12:42
*** daxxar has quit IRC12:42
*** TBBle_ has quit IRC12:42
*** dmpayton has quit IRC12:42
*** JLP has quit IRC12:42
*** brlcad has quit IRC12:42
*** nash has quit IRC12:42
*** matthewd has quit IRC12:42
*** Epyon has quit IRC12:42
*** mithro has joined #tp12:43
*** llnz has joined #tp12:43
*** jotham has joined #tp12:43
*** tuna-fish has joined #tp12:43
*** sijmen has joined #tp12:43
*** nash has joined #tp12:43
*** matthewd has joined #tp12:43
*** Demitar has joined #tp12:43
*** CIA-17 has joined #tp12:43
*** daxxar has joined #tp12:43
*** DystopicFro has joined #tp12:43
*** Demitar_ has joined #tp12:43
*** dmpayton has joined #tp12:43
*** TBBle_ has joined #tp12:43
*** brlcad has joined #tp12:43
*** JLP has joined #tp12:43
*** fizzle_ has joined #tp12:43
*** Epyon has joined #tp12:43
nashSo I can start a game with a few friends or other poeple without everyone just jumnping in part way through the game.12:43
mithronash: hrm, well once we have EOT support added it won't be such problem12:43
nashNope, it's a start, but doesn't help.  In many games the server needs # of players when it sets up the universe12:44
nashCreating a <player> home star doesn't work12:44
mithrowhy?12:44
nashRFTS and Stars! are the obvious examples12:44
nashWell for starters peoples homeworlds are secret...12:44
nash[in both of those games]12:45
nashSo it needs to be there at the begining12:45
nashSlotting them in to a random world or creating one doesn't work12:45
mithroahh well, figure out something then :P12:45
mithrothis is not an area i'm all that intrested in12:46
nashWhat - playing stars?12:47
mithrotrying to figure out how to do a lobby properly12:47
nashdoign is properly is anohter issue12:47
nashAs an aside... Posix regex interface in C... sucks12:49
mithrohe he :P12:49
nashNever used them before in C...12:49
* nash was half tempted to do popen("perl -e '/.../'");12:50
llnzis supporting tp03 with binary version needed? (ie same header format as tp04)12:50
mithrollnz: it would be nice12:51
mithrollnz: as it's going to be a fairly slow process for libtpproto-py to be updated12:51
mithro(unless one of the students gets into it)12:51
llnzit doesn't really add anything except allowing that header to be different12:53
mithro[email protected]:~/oss/tp/tpsai-py$ time ./tpsai-py > /dev/null12:54
mithroreal    0m0.772s12:54
mithrouser    0m0.532s12:54
mithrosys     0m0.048s12:54
mithrollnz: up to you12:54
mithronash: not to bad, 770ms to start up python, download the complete universe (with the aicomp config), and issue all the orders12:56
nashmithro: Sure beats the hell out of the python client ;-)12:56
mithroreal    0m2.553s12:57
mithrothats without the redirect to null12:57
mithro1.8s just to output to the console!?12:58
nashHow much does it pront12:58
mithroquite a bit12:58
nashconsoles (especially gnome-term and friends) are slow...12:58
mithro565 lines12:58
nashsounds about right12:59
mithro./tpsai-py | wc -l12:59
mithro56512:59
nashTry running with >& /dev/null12:59
mithrohrm?12:59
mithrostill around 700ms12:59
mithroif i turn debugging on it goes to 11 seconds!13:00
nashBTW... 6923 is in the middle of the bittorrent range...13:04
llnzbittorrent will use any available port13:05
sijmenthat's probably why they blocke dit here then13:05
nashllnz: Only between 6800 and 6999 was what I was just reading13:05
llnzhumm...13:06
nashAzerus or whatver it is called uses the same range13:06
sijmenAzereus13:08
nashthank yo13:08
nashu13:08
nashstupid keyboard13:08
*** Epyon has quit IRC13:08
nashmithro: Regarding the "Your AI must work from an arbitrary location"13:10
mithroyes?13:10
nashI assume that means you may do something like make install PREFIX=/usr/pants13:10
nashBut you won't just do a cp tpai /usr/pants and expect it to work?13:10
nashActually, I realised... since it's running without GUI... I don;t need any data13:11
nashDon't worry for now.. I'll sort that out later ;-)13:12
nashThe data path I need is only for the GUI...13:13
nashSo... who cares at the moment ;-)13:13
mithro:)13:15
mithrojc, the ai has exponential growth :P13:19
nashmithro: I've been saying that for ages13:20
nashIt's why minisec is an AI game.  And really pointless13:20
nashmithro: And it's extremely sensite to initial conditions.13:20
nashGet a setup where one AI is close to a few planets, and one is a bit away.13:21
nashAfter 20 turns see the difference13:21
mithroprobably13:21
nashIt's a scary difference13:21
* mithro is going to leave it running for a bit13:22
mithrotpclient-pywx is now able to keep up with 20 second turns13:22
llnzonce all planets are colonised, the real test/fun begins13:22
nash#set turn length (in seconds)13:22
nashturn_length = 1013:22
nash#minisec ruleset options13:22
nashminisec_min_systems = 5013:22
nashminisec_max_systems = 10013:22
nash#which gives about 20 on average13:22
nashminisec_min_planets = 113:22
nashminisec_max_planets = 1013:22
nash#which gives about 2 on average13:22
nashI've been testing with that ;-)13:22
nashllnz: Not for me.  With losing 2 days, I won't have time to merge in my agressive fleet patches :-(13:23
llnznash: how good is your ai at defending?13:23
nashHaving said that... it would take a long time to take out the fleets and fleets of frigates that are going to be above each world13:23
llnz:-)13:23
nashllnz: If there is nothing to colonise it keeps building frigates13:23
nashmemory usage is kinda scary when you leave it over night...13:24
mithronash: mine stops13:24
mithroit'll continually build frigates+battleships to try and take out other fleets13:24
mithrohttp://www.thousandparsec.net/~tim/Screenshot-1.png13:25
tpb<http://ln-s.net/MA5> (at www.thousandparsec.net)13:25
mithroafter 20 turns13:25
nashI've got a world attack patch, but tuning it to take an appropriate number of ships was a PITA13:26
llnzwow13:26
mithrollnz: he he, a human has no hope against an AI in Minisec :)13:27
nashllnz: Have you seen http://tranquillity.nash.id.au/tp/13:27
tpbTitle: Thousand Parsec E-Client Screenshots (at tranquillity.nash.id.au)13:27
nashmithro: I said that ages ago...13:27
llnzoh,13:27
llnzdoes the guest have no system or ships?13:27
mithrollnz: it has a scout, which the AI ignores (it has a threat level of 0)13:28
llnzodd13:28
mithrotpserver-cpp> tpserver-cpp: minisec.cpp:481: virtual void MiniSec::checkGame(): Assertion `fleet->totalShips() > 0' failed.13:28
mithroAborted (core dumped)13:28
mithro:/13:28
llnzmithro: update13:29
nashAhh... that's still around...13:29
mithronash: nah, running old code13:29
llnzi thought i had fixed that13:29
nashokay13:29
* mithro pulls13:29
mithrollnz: http://tranquillity.nash.id.au/tp/Screenshot-6.png :)13:29
tpb<http://ln-s.net/MA-> (at tranquillity.nash.id.au)13:29
llnzyeah, just saw that one13:30
nashmithro: Do you still have a sources list for the e libraries13:30
mithroyes13:30
mithrono13:30
nashGive me a moment - I'll get a 8 way shot13:30
nashIt's insane13:30
mithrothey disappeared when I updated to Fesity13:30
llnznash: Screenshot-13.png is 404 (not found)13:31
nashTo be frank, between little competition (so I'm just fighting myself at the moment) and a weak ruleset, my inspiration to fix the AI is limited13:31
nashllnz: okay - will fix13:31
mithronash: well in 4 months you'll have plenty of rulesets to fight with13:31
nashYeah, then I'll have fun13:32
nashllnz: Try noe13:32
nashtry now - it's a good one13:32
mithronash: i find your colors very hard to see13:32
nashmithro: Good thing my interface is themeable then ;-)13:33
llnznash: impressive13:33
mithronash: no colorizing the ships?13:33
nashmithro: Not yet13:34
nashAlso all ships are the same.13:34
nashI haven't had time to fix that13:34
nashThe ships are named now13:34
nashOf course... that makes it even harder to see at times...13:35
mithromine just makes them all called "A robot army!" :P13:35
nashstatic const char *shipnames[] = {13:35
nash"Endeavour",13:35
nash"Spore",13:35
nash"Mayflower",13:35
nash"Colonial",13:35
nash"Vanquish",13:35
nash"Victory",13:35
nash"Sirius",13:35
nash};13:35
nashFollowed by #id13:35
JLPahoy people13:36
llnzhi JLP13:36
nashheyo13:36
nashGenerating new screenshot with 8 AIs13:36
nashI'll wait to about turn 1013:36
CIA-17Lee Begg <[email protected]> * tpserver-cpp/ (52 files in 2 dirs):13:37
CIA-17TP04 support in frame, and reformating:13:37
CIA-17 Also added support for padding strings (and data and bytes) to 32bit boundaries, for13:37
CIA-17 the alignment filter.13:37
llnz:-)13:37
nashhmmm13:38
mithronash: it would be cool to do an animated version?13:38
mithrocould you automatically dump a screenshot each turn?13:38
nashI could actually13:39
nashBut not tonight... no time :-/13:39
nashI could also have it just save the star map13:39
mithronash: an animated picture would be soo cool :)13:40
nashmithro: I'll look over the weekend13:40
nashIt's fairly easy... Might make a new 'logger' interface for that however13:41
nashHmm... this is not good13:42
nashsince I updated the server 8 AIs don't work properly13:42
nash4 are fine13:42
nashThere is a 14 and 15 screenshot there now...13:47
mithrohrm, not seeing them13:48
nashupdating index13:49
nashokay - look now...13:52
mithrowow13:53
mithrohttp://tranquillity.nash.id.au/tp/Screenshot-17.png13:53
tpb<http://ln-s.net/MAc> (at tranquillity.nash.id.au)13:53
nashthat's 4 AIs13:54
nashI've done 16.13:54
nashI had a larger map.13:54
nashTurn gen was taking about 30 seconds13:54
llnzouch13:55
nashI had to give them 20 seconds to do turn entry...13:55
llnzapprox number of objects?13:55
mithronash: fixing up the AI improve libtpproto-py quite a bit13:56
nashllnz: 400 * 5 worlds13:57
nashProbably a few times that again ships13:57
nashDuplicated 17 times... (1 server + 16 AIs)13:57
llnzso circa 8000-10000 objects13:57
nashYeah13:58
nashProbably something like that13:58
llnzwhat's tpserver-cpp's memory usage like?13:58
*** Iwanowitch has joined #tp13:59
mithronash: I wonder how tpclient-pywx would handle that many objects14:00
llnzmithro: removing the delay from libtpproto-py made tpclient-pywx go nice and fast14:01
nashmithro: Badly... My machine was hurting14:01
nashI stopped when I got a CPU over temp limit message14:01
llnzouch14:02
nashllnz: My machine has heat problems to be fair14:02
mithrollnz: i think the config parsing is a bit borked?14:02
mithro#minisec ruleset options14:03
mithrominisec_min_systems = 100014:03
mithrominisec_max_systems = 500014:03
mithroyet it only generated 150 systems14:03
mithroany idea?14:03
llnzit should only generate one for each name in the list of system names14:04
llnzso it should only create 45 systems14:04
mithro:/14:04
mithroso I should make the list bigger?14:04
llnzit may have also reached the minisec_total_planets limit14:04
nashmithro: Highlight - cut and paste, s/"/"X/ for the second14:05
nashRepeat but with Y14:05
mithrollnz: it should probably read from a text file?14:07
nashOr have a generator ;-)14:07
nashmithro: Add more planets ;-)14:08
nashI use 1-10, but higher numbers can be quite silly for objects ;-)14:08
mithrowhere did I put matthew's planet name generator14:08
CIA-17Lee Begg <[email protected]> * tpserver-cpp/tpserver/ (frame.cpp frame.h): Better name for enablePaddingStrings14:09
CIA-17Lee Begg <[email protected]> * tpserver-cpp/tpserver/playertcpconn.cpp :14:09
CIA-17Allow incoming tp04 connections:14:09
CIA-17 I think it set it up well.14:09
CIA-17Lee Begg <[email protected]> * tpserver-cpp/tpserver/ (4 files):14:09
CIA-17Added paddingfilter property to playerconnection:14:09
CIA-17 Currently not set, but it is there.14:09
llnzthe server at llnz.dyndns.org will take tp04 connections14:10
nashllnz: I'm so not adding support for that now..14:11
llnzhehe14:11
llnzi guess i need to write more of the protocol.xml soon14:13
nashyes14:14
mithroand the xslt14:18
llnzyes14:19
llnzanyway14:20
* llnz wanders off14:20
llnzlater all14:20
nashllnz: night14:20
*** llnz has quit IRC14:20
nashBTW: Writing submission email now14:20
nashmithro: Enjoy14:21
mithro5000 systems is a bad idea :P14:21
nashmithro: :-)14:22
nashmithro: We may need to start work on more ondemand server downloading ;-)14:23
mithro:)14:23
nashAnyway, got my submission yet?14:23
*** Epyon has joined #tp14:24
nashmithro: ?14:25
mithroyes14:25
nashSeem to be okay?14:25
mithroyeah14:26
mithrojust downloading now14:26
nashcool14:26
mithronash: sums match14:27
* nash isn't surprised14:27
nashAlso you can cg-clone from that public git tree14:28
nashThe E libs work?14:29
* nash wants to go to bed soon14:29
nashreal soon14:29
nashhas to be up i too few hours14:30
mithronash: i'll try later, so feel free to head to bed14:30
nashokay14:30
nashnight14:30
*** TBBle_ is now known as TBBle14:31
mithrohowdy TBBle, any luck? :P14:50
*** sijmen has quit IRC14:50
TBBleSorry, been flat out at work all week. And I've just noticed my laptop clock had stopped, it's 90 minutes later than I thought it was. >_<14:51
mithro:/14:51
*** zipola has joined #tp15:56
daxxarAaah. What a magnificent day.16:06
IwanowitchIs it?16:07
daxxarFrom my subjective view, yes. :-)16:07
IwanowitchAny particular reasons, then? :)16:08
daxxarGreat weather (mild, sunny), last day before I take easter vacation (and travel to a cabin with some friends), finished the last piece of obligatory work for a subject that's really lame, I met with a girl I fooled around with on saturday, and I've got *GREAT* tasting beer here. ;D16:08
IwanowitchSuppose that's okay, then :P16:09
daxxarGreat big step up from the last week(s), indeed. I guess the whole weather thing has a great effect on ones mood. :)16:09
daxxarOh, and my PS2 arrived ;p16:10
daxxarHow's your day then, Iwanowitch? :)16:11
IwanowitchOh, not bad... Doing some schoolwork... Numercial methods for fluid flow, and it's not as fun as it sounds :P16:11
daxxarHehe ;p16:12
IwanowitchThe weather is gray, though... Was a lot better the previous days.16:12
daxxarWhat field are you studying in?16:12
daxxarAnd where do you live? :)16:12
IwanowitchMaths with minor informatics.16:13
IwanowitchBelgium.16:13
daxxarAh, ok. :-)16:13
daxxarBachelors degree?16:13
IwanowitchYep.16:14
daxxarNeat. Second year?16:14
IwanowitchSomewhere in between second and third year now.16:14
daxxarAh, okay. :)16:14
IwanowitchI got some courses from the second year I failed last time.16:14
IwanowitchSo it's a bit mixed up at the moment :)16:15
daxxarHehe, yeah, I can relate to that. ;-)16:15
daxxarI was just going to say I've never been to belgium, but I actually have. ;-)16:15
daxxarVisited Antwerpen a few years ago, on our way to Paris. :)16:16
IwanowitchI don't know Antwerpen too well myself, sorry :P16:17
daxxarHehe, me neither, was just a quick visit on our way to Paris ;-)16:17
daxxarHad belgian waffles (?) and visited, uhm, sj√łmannskirka.16:17
daxxarRoughly translates to sailorschurch. =P16:17
daxxarsailors church*16:18
IwanowitchHmm, doesn't ring a bell :P16:19
*** Iwanowitch_ has joined #tp16:27
*** Iwanowitch has quit IRC16:27
*** Iwanowitch_ is now known as Iwanowitch16:30
IwanowitchDammit, Processing is cool. 5 minutes of coding and I have the Lorenz butterfly on my screen.16:39
daxxarProcessing?18:36
IwanowitchIt's a Java-like language/tool thing. Makes it easy to create small graphical programs.18:42
Iwanowitch(among other things)18:42
Iwanowitchhttp://processing.org/18:42
tpbTitle: Processing 1.0 (BETA) (at processing.org)18:42
IwanowitchIf someone asks for a language to start programming with, I'm going to point them to Processing from now on.18:43
*** clayasaurus has joined #tp18:50
*** Todd has joined #TP18:51
*** zipola has quit IRC19:53
*** fizzle_ has quit IRC20:10
*** Todd has quit IRC20:38
*** nash has quit IRC21:02
* dmpayton yawns21:37
dmpaytonMornin'21:37
clayasaurusmornin''21:41
clayasauruson the 1st day of google of code, someone gave to me...21:41
dmpaytona perl script that was bug-free...21:43
dmpayton?21:43
clayasaurus:-P21:43
*** clayasaurus has quit IRC21:56
*** mithro has quit IRC22:43
*** Demitar has quit IRC23:01
*** nash has joined #tp23:03
*** nash has left #tp23:03
*** nash has joined #tp23:04
* nash waves23:05
*** dmpayton has quit IRC23:12
* Iwanowitch waves back to nash23:23
Iwanowitchkind of silent in here...K23:23
IwanowitchWith that K at the start of my sentence.23:24
nashIwanowitch: How is life?23:24
IwanowitchWaiting, waiting... :P23:25
nashNot for SoC I hope?23:28
IwanowitchOh yes.23:29
IwanowitchNah, seriously, nothing much is happening... Bit of schoolwork.23:29
IwanowitchNext two weeks are vacation, though.23:29
nashWell start working then ;-)23:33
IwanowitchMeh, at 1:38 AM I can't do a lot of good things I think :P23:35
nashAre you on the mailing lists?23:36
IwanowitchYeah, on the tp one... Need to subscribe to tp-devel, too.23:39
nashSubscribe23:39
nashI can give you a long list of unsolved problems too... if you want to start discussions23:40
*** Demitar has joined #tp23:43
* nash waves23:44
IwanowitchOkay, I subscribed. A list of problems you say?23:44
nashOkay: First issue - protocol for creating a game.23:45
nashI want to be able to connect to a server: Using a game wizard or something start/create a new game23:46
nashSecond: Lobby; I want to get a few of my friends together on a common serve,r and start a new game of X23:46
nashThird: URI scheme to tell my client to talk to a particular metaserver23:46
IwanowitchOkay. Protocol for creating a game... I suppose this needs credentials checking, list of installed modules exchange, list of parameters exchange23:49
nashHave you seen the tp04 draft?23:50
nashAnd notes?23:50
IwanowitchThe discussion page, yes, don't know if there's anything more?23:51
Iwanowitch(this one: http://www.thousandparsec.net/tp/dev/documents/protocol4.php)23:51
tpb<http://ln-s.net/JqO> (at www.thousandparsec.net)23:51
nashThat is the main one.23:51
nashThere is a draft generated from the XML source23:52
nashIt's not so interesting23:52

Generated by irclog2html.py 2.5 by Marius Gedminas - find it at mg.pov.lt!