--- Log opened Thu Sep 28 00:00:13 2017
20170928 00:07:44-!- sevu [~Shiki@p5485450F.dip0.t-ipconnect.de] has quit [Remote host closed the connection]
20170928 00:11:05-!- vultraz [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20170928 00:21:22-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170928 00:21:55-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 00:30:42-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20170928 00:36:59-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20170928 01:17:04-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170928 01:17:37-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 02:04:49-!- vultraz [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20170928 02:09:46-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170928 02:10:20-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 02:16:32-!- Elsi [~Elsi@luwin.ulrar.net] has quit [Ping timeout: 246 seconds]
20170928 02:19:19-!- Elsi [~Elsi@luwin.ulrar.net] has joined #wesnoth-dev
20170928 02:27:49-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20170928 02:27:57-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20170928 02:44:00-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20170928 03:41:58-!- lobby [~wesnoth@wesnoth/bot/lobby] has joined #wesnoth-dev
20170928 03:41:58-!- Topic for #wesnoth-dev: 1.13.10 tagging pending vultraz fixing the right click bug | Wesnoth Developers Channel | >>> Want to help? Go here: http://r.wesnoth.org/t42911 (and thanks!) <<< | Discord Server: https://discord.gg/tSmJS2E | Logs: http://irclogs.wesnoth.org | Bug tracker: http://bugs.wesnoth.org
20170928 03:41:58-!- Topic set by JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] [Sun Sep 24 04:44:49 2017]
20170928 03:41:58[Users #wesnoth-dev]
20170928 03:41:58[ aeth           ] [ DeFender1031 ] [ heirecka   ] [ matthiaskrgr] [ Polsaker ] [ timotei_ ] 
20170928 03:41:58[ AI0867         ] [ elias        ] [ higgins    ] [ mattsc      ] [ pydsigner] [ tomreyn  ] 
20170928 03:41:58[ aidanhs        ] [ EliDupree    ] [ Ivanovic   ] [ minzbonbon  ] [ Ravana_  ] [ vincent_c] 
20170928 03:41:58[ APic           ] [ Elsi         ] [ iwaim      ] [ molgrum     ] [ Rhonda   ] [ vn9711   ] 
20170928 03:41:58[ Appleman1234_  ] [ Elvish_Hunter] [ janebot    ] [ Necrosporus ] [ senkwich ] [ vultraz  ] 
20170928 03:41:58[ atarocch       ] [ esr          ] [ Jetrel_bot ] [ new_one     ] [ shadowm  ] [ wedge009 ] 
20170928 03:41:58[ ChipmunkV[m]   ] [ Gambit       ] [ lobby      ] [ nore        ] [ TC01     ] [ Yaiyan   ] 
20170928 03:41:58[ crimson_penguin] [ Greg_Boggs[m]] [ loonycyborg] [ nurupo      ] [ TC02     ] 
20170928 03:41:58[ DDR            ] [ grzywacz     ] [ madmax28   ] [ oldlaptop   ] [ TheJJ    ] 
20170928 03:41:58-!- Irssi: #wesnoth-dev: Total of 52 nicks [0 ops, 0 halfops, 0 voices, 52 normal]
20170928 03:42:05-!- Channel #wesnoth-dev created Tue Jan 27 05:28:41 2009
20170928 03:42:52-!- Soliton [~Soliton@wesnoth/developer/soliton] has joined #wesnoth-dev
20170928 03:42:55-!- Irssi: Join to #wesnoth-dev was synced in 68 secs
20170928 05:05:29-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has joined #wesnoth-dev
20170928 05:43:04-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20170928 05:48:29-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has quit [Quit: Going to work]
20170928 05:57:53-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170928 06:29:30-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has joined #wesnoth-dev
20170928 08:11:06-!- vultraz [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20170928 08:52:43-!- vn9711 [~vasya@94.158.103.15] has quit [Quit: Leaving.]
20170928 08:54:43-!- Bhoren [~Bhoren___@82.64.3.181] has joined #wesnoth-dev
20170928 08:59:22-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20170928 09:24:19-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has quit [Quit: .]
20170928 09:47:06-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170928 09:59:58-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20170928 10:13:18-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170928 10:38:46-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has quit [Quit: [endlevel]]
20170928 10:53:42-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has joined #wesnoth-dev
20170928 11:09:24-!- vn971 [~vasya@94.158.103.15] has joined #wesnoth-dev
20170928 11:11:31< vn971> Hi guys. I observe the following sequence pretty often: * people start playing * one player decides his side lost * he quits. In a good case he writes gg or "I surrender". Often, unfortunately, there's just a silent quit.
20170928 11:12:01< vn971> So I wonder, can we do something about those silent quits. Maybe there's just no good GUI for quitting? Maybe we should improve _that_?
20170928 11:12:47< vn971> Like, have a button/function/whatever that means "I lost and I'm quitting". Or even just "I lost".
20170928 11:13:33< DeFender1031> A simple "forfeit" or "surrender" button could work, perhaps.
20170928 11:15:27< zookeeper> yeah, that seems like a pretty common request. i don't recall any particular objections to that. not sure how past suggestions have dealt with one player in a team surrendering, though.
20170928 11:16:01 * zookeeper isn't really here
20170928 11:17:08< DeFender1031> some interface to select which of the remaining team members control of that side gets transferred to? (Or just transfer to the sole team member if there's only one.)
20170928 12:08:51-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 12:28:56-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20170928 12:56:12-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20170928 12:56:18-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20170928 13:18:04-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 13:27:59< zookeeper> DeFender1031, well, sure, for the host you could pop up the control transfer dialog if one player surrenders, and when the last player of the team surrenders, that triggers victory/defeat.
20170928 13:28:14< DeFender1031> Exactly.
20170928 13:28:46< zookeeper> if you surrender, then presumably you'd still stay in the game as an observer. unless there were another "surrender and quit" option.
20170928 13:29:54< JyrkiVesterinen> I think it would make more sense to automatically quit when surrendering.
20170928 13:30:43< Necrosporus> Existing interface allows it all, but users probably have troubles using console instead of GUI
20170928 13:31:37< Necrosporus> I have troubles with Wesnoth console too, since I do not have any idea what many commands are for such as layers
20170928 13:32:09< Necrosporus> And there is no autocompletion for user names when you type :control 2 JyrkiVesterinen or something
20170928 13:32:11< Necrosporus> or is it?
20170928 13:34:18< Ravana_> I believe many commands are just not useful to keep around
20170928 13:36:19-!- sevu [~Shiki@141.57.59.86] has joined #wesnoth-dev
20170928 13:38:05< Necrosporus> Actually there is :give_control instead of auto-completion
20170928 13:38:31 * Necrosporus decided to re-read command help in ingame help
20170928 13:51:51-!- Bhoren [~Bhoren___@82.64.3.181] has quit [Remote host closed the connection]
20170928 13:53:32< Necrosporus> layers command is not in Wesnoth help
20170928 13:58:39< sevu> it's not in https://wiki.wesnoth.org/COMMANDMODE either
20170928 14:00:04-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170928 14:00:37-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 14:02:13< Necrosporus> Do we need a bug report?
20170928 14:12:04< Necrosporus> Will non-ascii characters in addon filenames cause problems?
20170928 14:12:36< JyrkiVesterinen> Do you mean extended Latin alphabet, like Ä?
20170928 14:12:47< Necrosporus> I know Wesnoth handling of pathnames on Windows has been improved by boos filesystem
20170928 14:13:13< Necrosporus> JyrkiVesterinen, yes, and greek, cyrillic, georgian, japaneses...
20170928 14:13:22< Necrosporus> anything that has code >127
20170928 14:13:29< JyrkiVesterinen> Those characters don't cause problems. Any modern file system can handle such characters, and Wesnoth is also capable of loading files with such characters in their names.
20170928 14:13:53< Necrosporus> Problem is that filesystems hanlde them differently
20170928 14:14:14< Necrosporus> Also extended latin characters could be written at least in two ways
20170928 14:14:54< Necrosporus> for example á can be latin a followed by combining character or single codepoint
20170928 14:15:30< JyrkiVesterinen> That's not a problem as long as the character is encoded in the same way both in the actual file name and in the reference to it.
20170928 14:15:31< Necrosporus> Linux filesystem would allow both names in same directory since ext* filesystem are byte-oriented
20170928 14:15:47< Necrosporus> They keep name as byte array not caring about unicode or whater
20170928 14:15:52< JyrkiVesterinen> And if the encodings are somehow different, then what the hell is the UMC author doing?
20170928 14:16:17< Necrosporus> Problem will arise when addon is transfered through network
20170928 14:16:48< Necrosporus> AFAIR Windows and MacOS X will normalize filenames
20170928 14:17:04< Necrosporus> While linux will not
20170928 14:17:23< JyrkiVesterinen> Well, if they normalize file names when saving, then I'd guess that they normalize them while loading as well, right?
20170928 14:17:45< Necrosporus> I don't know. I do not have mac or windows to experiment
20170928 14:18:22< Necrosporus> Also Windows filesystem doens't support characters outside of BMP (basic multiligual plane)
20170928 14:18:30< Necrosporus> Since it uses 2 bytes per character
20170928 14:19:17< JyrkiVesterinen> Fortunately, supplementary planes barely contain anything useful.
20170928 14:19:30< Necrosporus> Windows itself could emulate it using UTF-16 way, which is using surrogate pairs, but different windows version and application may or may not handle surrogates
20170928 14:19:46< JyrkiVesterinen> IIRC, some rare CJK characters are the only things in supplementary planes someone might attempt to use in filenames.
20170928 14:20:16< Necrosporus> I am not sure that Wesnoth will handle them properly
20170928 14:20:27< Necrosporus> Does Wesnoth require .cfg to be in UTF-8?
20170928 14:20:39< Necrosporus> Does it handle BOM?
20170928 14:20:56< Necrosporus> Some text editors insert BOM in UTF-8, others do not
20170928 14:22:33< Necrosporus> Some characters might be problematic if utf-8 to utf-16 conversion is used
20170928 14:23:12< Necrosporus> Perhaps Battle for Wesnoth + boost is smart enough to handle, I didn't test
20170928 14:23:13< sevu> If you have a BOM you get lot's of warnings in stderr (which probably nobody reads :p), but it works
20170928 14:23:43< Necrosporus> But firefox for example won't display UTF-8 text without BOM
20170928 14:23:46< Necrosporus> Which is rather lame
20170928 14:25:39-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170928 14:26:13-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 14:29:09-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has quit [Quit: Going home]
20170928 14:37:28< Necrosporus> I think that today all browser getting text/plain should default to UTF-8 unless document content doesn't contradict this assumption
20170928 14:37:44< Necrosporus> Firefox does not default to UTF-8 unless there is BOM
20170928 14:38:17< Necrosporus> I don't know about other browsers
20170928 14:49:07-!- vultraz [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20170928 14:52:14-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20170928 14:52:47-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20170928 15:06:24< AI0867> actually, I believe there is some standard that says browsers should assume ISO-8859-1 unless told otherwise
20170928 15:08:34< Necrosporus> even if that's so it should be reviewed
20170928 15:08:47< Necrosporus> A random text file is much more likely to be UTF-8
20170928 15:09:06< Necrosporus> Especially if uses non-latin script
20170928 15:09:19< Necrosporus> Then it's guaranteed to be not ISO-8859-1
20170928 15:10:01< Necrosporus> I'd say standard or not, but UTF-8 is much saner default. Standard should be updated
20170928 15:20:50< vultraz> then write to the ISO :P
20170928 15:25:08-!- Oebele [~quassel@143.177.58.202] has joined #wesnoth-dev
20170928 15:25:46-!- sevu [~Shiki@141.57.59.86] has quit [Quit: Verlassend]
20170928 15:27:47< AI0867> Necrosporus: I agree, but that's what's actually specified
20170928 15:28:16< Necrosporus> AI0867, where?
20170928 15:28:38< Necrosporus> Anyway browser does not have to actually do that. If it's required then it should be preference option
20170928 15:28:59< Necrosporus> I think then it will pass standard conformance if that's necessary
20170928 15:29:57< Necrosporus> Also I think that ISO-8859-1 and UTF-8 could be easily deduced in most cases. Test several first non-ascii sequences. If you find invalid UTF-8, default to ISO
20170928 15:30:55< Necrosporus> I don't think it will take long, since iso for extended latin is mostly used in languages which use mostly regular latin with few special letters, such as German
20170928 15:31:25< Necrosporus> then the first non-ascii letter will be different between unicode and iso
20170928 15:31:26< AI0867> https://www.w3.org/International/articles/http-charset/index
20170928 15:31:50< Necrosporus> Since they approved EME I don't think it makes sence to follow their standards
20170928 15:33:20< Necrosporus> I mean strictly follow
20170928 15:33:33< Necrosporus> "Follow when and only when it makes sense" is better
20170928 15:33:40< shadowm> lmao
20170928 15:33:55< Necrosporus> What does whatwg say on this matter?
20170928 15:34:16< shadowm> Necrosporus: Are you developing a web browser?
20170928 15:35:03< Necrosporus> No. But I use one and I notice that experience degrades every year or so
20170928 15:35:27< shadowm> Well, unfortunately this is not web browser devs central and you already know Wesnoth aggressively assumes UTF-8 for everything.
20170928 15:35:58< Necrosporus> I agree with Wesnoth on that
20170928 15:36:50< Necrosporus> Except that I think it's best not to use non-ascii in WML at all and put it into po files instead. But if you really can't understand English, then fine
20170928 15:36:52< shadowm> Any news on 1.13.10? I'm still waiting for my PR to be reviewed, which I assume will be after that is out of the way.
20170928 15:37:02< shadowm> vultraz: ?
20170928 15:37:24< vultraz> I've been hella busy
20170928 15:37:27< shadowm> I swear to god, I should've released 1.15.0 instead of 1.13.2.
20170928 15:37:28< vultraz> I'm sorry
20170928 15:37:32< shadowm> It must be the 13 number.
20170928 15:37:55< shadowm> I actually seriously considered changing the version scheme altogether to get rid of the 13.
20170928 15:38:18< Necrosporus> Numbers have no problems by themselves
20170928 15:38:31< Necrosporus> People give them meaning, it could cause problems though
20170928 15:38:32< shadowm> Says you.
20170928 15:40:40-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has joined #wesnoth-dev
20170928 15:41:17< Necrosporus> I personally consider right-click bug quite minor. I didn't notice it until I was told about it
20170928 15:42:04< Necrosporus> Also it could be considered a feature
20170928 15:42:36< Necrosporus> For example, you have several units which can spawn other units on locations which they can reach
20170928 15:42:44< shadowm> I think the devs know what they are doing.
20170928 15:42:58< shadowm> If they decided it's a thing that needs to be fixed, I trust they have good reasons to do so.
20170928 15:42:59< Necrosporus> Well, I'm not saying otherwise
20170928 15:43:02< shadowm> I don't even play 1.13.x.
20170928 15:43:06< shadowm> I don't even play 1.12.x.
20170928 15:43:12< shadowm> I started playing again last night.
20170928 15:43:59< Necrosporus> But consider my idea anyway. What if you need to have several units which all can do some actions with adjacent location. 
20170928 15:44:17< Necrosporus> Like summon a zombie in exchange of their XP points
20170928 15:44:39< JyrkiVesterinen> I myself found the right-click bug very annoying when I playtested the pseudo-1.13.10 build.
20170928 15:44:53< Necrosporus> If you can select a unit and then spawn right-click menu and you have two units with such ability adjacent to location where you want to spawn a zombie...
20170928 15:45:54< Necrosporus> Then if selection is retained while right-click menu is displayed then you can write an event which will determine which unit you wanted to take XP from 
20170928 15:46:28< Necrosporus> Imagine the bug is fixed and behavior brought to same as in 1.12 
20170928 15:46:49< Necrosporus> How do you summon a zombie in location which is adjacent to two units which can do it?
20170928 15:47:17< Necrosporus> While preserving ability to select which one of two spends XP (or HP or whatever)
20170928 15:48:30< Necrosporus> Though I agree that few if any at all of addon use anything like it
20170928 15:49:43< Necrosporus> Also... even if behavior is reverted to 1.12 a WML-accessible variable can keep last-selected unit even if it's deselected
20170928 15:51:36< shadowm> What.
20170928 15:52:19< zookeeper> don't feed.
20170928 15:52:40< zookeeper> just exhale, look the other way and think of something nice.
20170928 15:52:56< shadowm> I don't know about you, but it's mildly entertaining from my point of view.
20170928 15:53:49< zookeeper> exactly opposite qualifiers for me
20170928 15:54:10< Necrosporus> Sorry
20170928 15:58:13< vultraz> zookeeper: think of coffee
20170928 16:00:39< zookeeper> is it nice?
20170928 16:02:36-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170928 16:04:43< vultraz> you're finnish, you should know.
20170928 16:04:53< vultraz> finnish people drink 3x as much coffee as us americans
20170928 16:06:05< zookeeper> finnish people except me, apparently
20170928 16:06:40< JyrkiVesterinen> I don't drink coffee either.
20170928 16:07:05< zookeeper> :o
20170928 16:07:10< zookeeper> rare!
20170928 16:08:37< vultraz> !!
20170928 16:15:08-!- gfgtdf [~chatzilla@x4e368c53.dyn.telefonica.de] has joined #wesnoth-dev
20170928 16:19:23< gfgtdf> i wonder whether we should add abug cteagory 'important' liek the odl tracker had, which isn't really a blocker but still more impotant thean others.
20170928 16:21:59< Necrosporus> gfgtdf, do you know about SDL internals?
20170928 16:22:53< gfgtdf> Necrosporus: not really i investigates the utf8 bug a little back then but that infromation is probably outdated aswell.
20170928 16:23:13-!- irker300 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20170928 16:23:13< irker300> wesnoth: gfgtdf wesnoth:master 74e4ea04e70c / src/generators/default_map_generator_job.cpp: don't generate names if [generator] has no [naming] tag  https://github.com/wesnoth/wesnoth/commit/74e4ea04e70c7183adc5413eb30529ad3fe3207b
20170928 16:36:54< vultraz> gfgtdf: that block doesn't make sense. don't you already copy the contents of the [naming] tag when using config::child?
20170928 16:37:09< vultraz> why do you need to then append the attributed 
20170928 16:37:11< vultraz> attributes 
20170928 16:38:02< gfgtdf> i want to copy them, not sure what you question is.
20170928 16:38:23< vultraz> oh wait i misread he code
20170928 16:38:26< vultraz> neverind
20170928 16:53:03-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20170928 16:57:03< Necrosporus> I have found another FullScreen bug
20170928 17:53:08-!- mjs-de [~mjs-de@x4db67954.dyn.telefonica.de] has joined #wesnoth-dev
20170928 17:56:09-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has joined #wesnoth-dev
20170928 18:09:09-!- travis-ci [~travis-ci@ec2-54-197-102-206.compute-1.amazonaws.com] has joined #wesnoth-dev
20170928 18:09:10< travis-ci> wesnoth/wesnoth#15113 (master - 74e4ea0 : gfgtdf): The build has errored.
20170928 18:09:10< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/280943853
20170928 18:09:10-!- travis-ci [~travis-ci@ec2-54-197-102-206.compute-1.amazonaws.com] has left #wesnoth-dev []
20170928 18:15:33< Necrosporus> Why skeletons can use weapons such as bows and axe, zombies can't?
20170928 18:25:43-!- fabi_ [~fabi@2a02:810c:c840:2e65:c4b9:f4dd:fbe0:452b] has joined #wesnoth-dev
20170928 18:28:35-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20170928 18:39:07-!- sevu [~Shiki@141.57.59.86] has joined #wesnoth-dev
20170928 18:57:04< sevu> I thought all the time nobody manages anything without coffee in IT jobs
20170928 18:58:42< elias> Amazon mobile-ads devs clearly manage on just vodka
20170928 19:06:20< fabi_> sevu: I prefer black tea :-)
20170928 19:29:51-!- irker300 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20170928 19:36:09-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20170928 19:58:41-!- mjs-de [~mjs-de@x4db67954.dyn.telefonica.de] has quit [Remote host closed the connection]
20170928 19:59:11< sevu> I assume vodka and black tea have a *slightly* different effect onto the result
20170928 20:02:05-!- sevu [~Shiki@141.57.59.86] has quit [Remote host closed the connection]
20170928 20:06:11< EliDupree> So, after making a naive AI, I made another AI that takes each possible move and does 100 playouts assuming both players are played by the naive AI. This AI actually performs WORSE than the naive AI. 
20170928 20:33:58-!- mkdroid [~null@unaffiliated/matthiaskrgr] has joined #wesnoth-dev
20170928 20:45:08-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20170928 20:52:51-!- mkdroid [~null@unaffiliated/matthiaskrgr] has quit [Quit: I'll be back!]
20170928 20:55:43-!- JyrkiVesterinen [~JyrkiVest@85-23-197-3.bb.dnainternet.fi] has quit [Quit: .]
20170928 21:03:30< EliDupree> Haha, I taught a naive AI a few more things and now it wins more than half the time (on this tiny map) because the default AI is very careless with its leader
20170928 21:04:32< EliDupree> It won 17/24 games
20170928 21:06:13< DeFender1031> There's always the question though when pitting AIs against each other whether it's truly indicative of how each would do agains a human though, since humans think and act differently.
20170928 21:08:17-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 248 seconds]
20170928 21:11:28< EliDupree> Well, if I can't defeat the default AI, I clearly won't be able to defeat humans, so it seems fine to use as a first goal
20170928 21:11:55< EliDupree> (Good humans, that is)
20170928 21:15:23< EliDupree> Also, it would be fun to have an AI that ruthlessly exploits the default AI's weaknesses, to run in campaigns and such
20170928 21:15:23< DeFender1031> Not necessarily, but probably.
20170928 21:29:33-!- gfgtdf_ [~chatzilla@x4e368c53.dyn.telefonica.de] has joined #wesnoth-dev
20170928 21:30:16-!- mkdroid [~null@unaffiliated/matthiaskrgr] has joined #wesnoth-dev
20170928 21:30:38-!- gfgtdf [~chatzilla@x4e368c53.dyn.telefonica.de] has quit [Ping timeout: 258 seconds]
20170928 21:30:43-!- gfgtdf_ is now known as gfgtdf
20170928 21:36:56-!- Oebele [~quassel@143.177.58.202] has quit [Remote host closed the connection]
20170928 21:38:52-!- vultraz [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20170928 22:15:18-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20170928 22:19:51-!- vultraz [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20170928 22:24:00-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20170928 22:27:43-!- mkdroid [~null@unaffiliated/matthiaskrgr] has quit [Quit: I'll be back!]
20170928 22:42:12< mattsc> EliDupree: welcome to AI development.  It’s a fascinating and overwhelming (at least for me, with no background whatsoever) topic.
20170928 22:42:20< EliDupree> yup!
20170928 22:43:13< EliDupree> I don't even know whether the "lookahead is worse than naive" thing means I have a bug, or if it is just a counterintuitive behavior (the lookahead AI decides that it needs to end its turn early so the naive AI doesn't do something that would make it vulnerable to the enemy naive AI)
20170928 22:44:23< EliDupree> I should really find a way to display the simulated playouts, to make sure my simplified wesnoth physics are actually correct
20170928 22:45:06< mattsc> Yeah …  Sometimes weird behavior comes from something that is only tangentially related.
20170928 22:45:32< mattsc> I have a lot of display and visulaitzation code in my AIs just for that purpose.
20170928 22:48:45< EliDupree> The annoying thing is that there's a lot of stats to display to represent the game state, but it's not easy to plug into wesnoth's existing way of displaying the game state
20170928 22:51:46< mattsc> Well, it’s a multi-dimensional problem …
20170928 22:53:34< EliDupree> True. I guess I should be able to throw together some sort of funky OpenGL thing to display exactly the data I want, but that's a bunch of work and my agenda for today is basically "lazily mess around" rather than "do a serious good job"
20170928 22:55:15< mattsc> Well, I am not talking about today (I’ve been at this for 5 years now), but let me grab a couple screenshots for potential ideas.
20170928 22:56:09< mattsc> For exmaple, I am using this to display certain types of information:  https://imgur.com/a/WddyQ
20170928 22:56:27< mattsc> or this:  https://imgur.com/a/RdZI3
20170928 22:57:28< EliDupree> I'm looking for a much better display than either of those. I already have a bunch of debug messages that go to the commandline output
20170928 22:59:30< EliDupree> A thing that comes to mind is "left side of the window is a move tree, mouse over one of the tree nodes to display the corresponding game state on the right side of the window"
20170928 23:03:55< mattsc> Sure, if you can pull this off, that sounds great.
20170928 23:03:57< EliDupree> Within wesnoth, I'm not sure there's even a good way to pause the game in between my AI's moves, which I would want to do to look at the visualizations (I guess maybe setting its controller = human and then back to AI repeatedly…?)
20170928 23:04:50< mattsc> I personally find my biggest problem to be on the conceptual side.  How do I rate a given move, in particular when it’s in combination with other units’ moves.
20170928 23:05:18< EliDupree> well of course
20170928 23:05:31< mattsc> I have some tools that let me call evaluation and/or execution of individual AI moves manually.
20170928 23:05:49< EliDupree> Like on a human player's turn?
20170928 23:05:53< mattsc> They are pretty “basic”, but they do the job.
20170928 23:05:58< mattsc> Yes.
20170928 23:08:27< EliDupree> If I continue work on this, I want to focus on machine learning rather than manually tuning AI evaluations
20170928 23:08:58< EliDupree> But before that I can do that, I should make sure my basic setup is working smoothly
20170928 23:11:25< mattsc> Yeah.  We (not I myself) did some machine learning on recruiting only back in the day.  It never really got to the level that it would have beaten our maunally tuned AI recruiter, but that was in part because the person working on it left.
20170928 23:11:37< EliDupree> heh
20170928 23:11:38< mattsc> And you probably know way more about this than I do anyway.
20170928 23:12:02< mattsc> The code is still in the AI-demos add-on, although it’s disabled for the time being.
20170928 23:13:46< mattsc> When you get somewhere with this, you’re always welcome to a match against Fred.  ;)
20170928 23:13:51< EliDupree> Fred?
20170928 23:14:16< mattsc> Fred also only plays one map (Freelands), and only Northerners and only side 1.
20170928 23:15:00< mattsc> He does pretty well against the default AI for that setup though (if I do say so myself).
20170928 23:15:24< mattsc> Fred:  https://forums.wesnoth.org/viewtopic.php?f=15&t=43048
20170928 23:16:18< EliDupree> heh
20170928 23:16:19< EliDupree> well
20170928 23:16:32< EliDupree> At some unspecified and possibly nonexistent point in the future, Fred is going DOWN!
20170928 23:17:52< mattsc> I’m looking forward to it!
20170928 23:18:47< mattsc> So, since you issued the challenge, here’s (approximately) what you need to do against the default AI then:  https://forums.wesnoth.org/viewtopic.php?f=15&t=43048&start=90#p616259
20170928 23:19:15< mattsc> As was mentioned before, what an AI does against another AI is not necessarily an indication of how it does against a human player.
20170928 23:29:20-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20170928 23:29:26-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
--- Log closed Fri Sep 29 00:00:14 2017