--- Log opened Mon Oct 02 00:00:18 2017
20171002 00:18:48-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20171002 00:19:21-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 00:20:24-!- vultraz [uid24821@wesnoth/developer/vultraz] has joined #wesnoth-dev
20171002 01:12:22-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20171002 01:12:55-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 01:21:29< mattsc> celticminstrel: typo here: https://github.com/wesnoth/wesnoth/blob/master/data/ai/micro_ais/cas/ca_fast_move.lua#L109
20171002 01:21:35< mattsc> That should be a capital M
20171002 01:23:32< celticminstrel> :(
20171002 01:27:01< mattsc> Not a biggie  :)
20171002 01:27:36< mattsc> Btw, found it because of this: https://forums.wesnoth.org/viewtopic.php?f=8&t=32384&start=6435#p617542
20171002 01:33:55-!- irker247 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20171002 01:33:55< irker247> wesnoth: doofus-01 wesnoth:master 8ab0024ddbf1 / data/core/music/return_to_wesnoth.ogg: placing return_to_wesnoth.ogg file in music folder for BfW 1.14, not wired in an https://github.com/wesnoth/wesnoth/commit/8ab0024ddbf142748287e8e99976eecbf4084720
20171002 01:53:31< shadowm> mattsc: So, under what circumstances will the AI actually take the risk and attack a player unit that's 1 XP short of leveling up?
20171002 01:54:23< shadowm> If you know off the top of your head, otherwise don't worry, it's just a fun question I decided to ask.
20171002 01:54:56< shadowm> Mostly because the AI has left my Enchantress completely untouched for several turns despite me using her as a meat shield.
20171002 01:57:01< shadowm> I'm pretty sure I already knew this was a thing but it's been a while since I last played Wesnoth, so I can't help but find this situation tremendously amusing.
20171002 01:58:42< mattsc> shadowm: I tried to explain the logic here:  https://github.com/wesnoth/wesnoth/blob/master/data/ai/lua/ca_high_xp_attack.lua#L6
20171002 02:00:46-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171002 02:01:52< shadowm> This is 1.12 without any Lua AIs, though.
20171002 02:02:47< shadowm> "unit value (gold)" Wait a second.
20171002 02:03:09< shadowm> Does this mean that the AI looks at the unit cost for unrecruitable units to do calculations?
20171002 02:06:20< mattsc> Oh, in 1.12.  Only if the AI has a chance to kill the unit then, either directly or by throwing enough units at it after the leveling.
20171002 02:06:40< shadowm> Yeah, it decided to throw a Chaos Overlord at her and kill her in one move.
20171002 02:06:47< mattsc> As for the cost question, yeah, I guess it does.  I can see how that might be a problem ...
20171002 02:07:22< celticminstrel> Some of the unrecruitable units have had their costs tweaked recently.
20171002 02:07:26< shadowm> I know it's just a stupid program, but I'd like to think the leader just decided "ENOUGH OF THIS BULLCRAP I'LL HANDLE HER".
20171002 02:07:42< celticminstrel> And the fact that the AI relies on it is IMO a good reason to continue to tweak them if needed.
20171002 02:07:53< shadowm> Also, UnitTypeWML in the wiki says nothing about the recruitment cost value having an ulterior significance.
20171002 02:08:28< shadowm> This poses some interesting questions, though.
20171002 02:08:54< shadowm> For years, I've always used arbitrary numbers for unrecruitable units, so e.g. Elynia's unit type in IftU has a recruitment cost of 114.
20171002 02:09:04< shadowm> She's a L4.
20171002 02:09:17< shadowm> The final boss has a recruitment cost of 900 and he's a L6 unit.
20171002 02:10:15< shadowm> A couple of spoiler L6 units in AtS have costs of 820 and 910.
20171002 02:11:44< mattsc> Well, frankly, the high-xp-attack CA was added by me not that long ago and I simply did not think of this issue of non-recruitable units
20171002 02:13:02< mattsc> There might, in fact, be something along those lines in one or two of the Micro AIs as well.
20171002 02:16:12< mattsc> … and the main reason for that, I guess, is because most of my AI development lately has been on this small MP map AI, where gold value is a very relevant quantity.
20171002 02:16:48< mattsc> In any case, if somebody has a better suggestion, that’s fine by me.
20171002 02:24:26< shadowm> I can't count past 2, so...
20171002 02:32:06< shadowm> I love that thing AI leaders are prone to doing where they move one hex away from their keep to let my leader (who just so happens to be nearby) recruit on it and deny them the possibility of recruiting more units to throw at me.
20171002 02:32:34< shadowm> It's pretty much always to perform an ineffective attack on a decoy unit.
20171002 02:34:59-!- sigurdfd [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has joined #wesnoth-dev
20171002 02:40:02< mattsc> I assume you know that there are options to turn many of these things on and off.  The problem with them is that they’re either always on or always off, neither of which is always the best thing to do.
20171002 02:41:11< mattsc> I’m trying to build “smarter” behavior into the AI on which I am currently working, but I am doubtful that that will ever reach the level of general usability.
20171002 04:22:20-!- sigurdfd [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has quit []
20171002 05:06:40-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171002 05:41:24-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20171002 05:50:20-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!]
20171002 06:18:14-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has joined #wesnoth-dev
20171002 06:21:49< JyrkiVesterinen> 20171001 18:47:38< vultraz> And updating MSVC isn't that hard
20171002 06:22:39< JyrkiVesterinen> I was mostly referring to 1) some people may have a commercial version (Professional or Enterprise) of 2015 but not 2017, and 2) Visual Studio is a huge download, and some people may not want to spend the necessary Internet bandwidth.
20171002 06:23:39< JyrkiVesterinen> I wouldn't limit support to *only* the latest version of Visual Studio unless it's really necessary.
20171002 06:24:08< Necrosporus> Is there anything wrong in supporting also 2013?
20171002 06:24:35< JyrkiVesterinen> 2013 is too old. It's definitely the compiler that most holds us back.
20171002 06:25:16< Necrosporus> Does 2015 or 2017 support at least C99 or better C11?
20171002 06:25:26< Necrosporus> or 2013?
20171002 06:27:21< JyrkiVesterinen> "Visual Studio 2015 fully implements the C99 Standard Library, with the exception of any library features that depend on compiler features not yet supported by the Visual C++ compiler (for example, <tgmath.h> is not implemented)."
20171002 06:27:42< JyrkiVesterinen> The C99 language isn't supported by VS, its compiler is a pure C++ compiler.
20171002 06:29:11< Necrosporus> Maybe it's possible to replace VS compiler with gcc
20171002 06:29:24< Necrosporus> It should probably produce better code and so on
20171002 06:29:27< JyrkiVesterinen> Absolutely not.
20171002 06:29:49< aeth> GNU treats Windows as a second-class platform.
20171002 06:29:52-!- atarocch [~atarocch@93.56.164.28] has quit [Quit: Leaving]
20171002 06:29:59< Necrosporus> That's correct
20171002 06:30:07< JyrkiVesterinen> It's better to compile the code with as many compilers as possible, because doing so alerts us if we accidentally rely on compiler-specific behavior.
20171002 06:30:18< Necrosporus> But nevertheless GNU GCC works on windows and it's probably better than all native Windows compilers
20171002 06:30:34< shadowm> It's really not.
20171002 06:31:07< Necrosporus> clang does not even work on windows without include files from mingw
20171002 06:31:40< Necrosporus> JyrkiVesterinen, tcc then?
20171002 06:32:05< Necrosporus> Though it doesn't support C++
20171002 06:32:41< JyrkiVesterinen> Therefore it's definitely not an option.
20171002 06:33:28< Necrosporus> C++ is too complex language, it's supported only by few
20171002 06:36:13< JyrkiVesterinen> Uh.. if you find C++ "too complex", then I have news for you. By that metric almost every language on Earth is "too complex".
20171002 06:36:19< JyrkiVesterinen> C#. Java. Lua.
20171002 06:37:10< Necrosporus> Lua is not, C# and Java are
20171002 06:37:31< JyrkiVesterinen> Lua has only two notable implementations.
20171002 06:37:46< Necrosporus> Not because it's too complex though
20171002 06:38:44< Necrosporus> But Scheme is definitily less complex, given there are perhaps hundreds of implementations
20171002 06:39:10< Necrosporus> Anyway, I have looked through Wikipedia and it seems that Clang and GCC are only free software compilers for C++
20171002 06:39:26< Necrosporus> With support for C++11 and later
20171002 06:39:35< shadowm> Today I learned that the number of implementations of a language is an objective measurement of its complexity.
20171002 06:41:11< JyrkiVesterinen> Necrosporus: so, I assume that you're trying to state that the number of FOSS compilers itself is a problem.
20171002 06:41:23< JyrkiVesterinen> I strongly disagree with that.
20171002 06:41:34< Necrosporus> I didn't say that
20171002 06:41:43< Necrosporus> Perhaps 1 or two but quality one is enough
20171002 06:45:38< Necrosporus> What I said is that: C++ is very complex. It's a reason why there are so few implementations despite it being one of very popular languages.
20171002 06:46:28< JyrkiVesterinen> Right. And my counterargument is that even for popular languages it's completely normal to have only one or two implementations.
20171002 06:47:20< JyrkiVesterinen> C++ is relatively complex by modern standards, but other languages aren't that much easier, either.
20171002 06:47:51< JyrkiVesterinen> For example, web browser developers have spent millions of dollars writing as fast JavaScript engines as they can.
20171002 06:48:58< Necrosporus> Nevertheless it's not necessary for language to be complex in order to be powerful
20171002 06:49:07< Necrosporus> Scheme is a nice example
20171002 06:50:06< JyrkiVesterinen> I can't really comment on that. I have never programmed in Scheme.
20171002 06:52:58< Necrosporus> Another example of simple but very powerful language is Tcl
20171002 06:53:32< Necrosporus> Lua is not as simple, but probably count too
20171002 06:54:07< JyrkiVesterinen> Please shut up. Continuing this discussion is pointless.
20171002 06:59:34-!- NicolasF[m] [nicolasfma@gateway/shell/matrix.org/x-jogdneoyhjnomprl] has joined #wesnoth-dev
20171002 07:07:13< Necrosporus> Sorry, I got sidetracked from main point. It is, there are no more FOSS compilers with C++11 support. Though there is Solaris Studio and Borland C++. Do you think it is useful to add their support?
20171002 07:07:56< JyrkiVesterinen> Not really, there are more important things to do.
20171002 08:00:59-!- mjs-de [~mjs-de@185.22.140.184] has joined #wesnoth-dev
20171002 08:10:35< aeth> I've programmed in Scheme. "Small language" just means that every Scheme adds its own, incompatible extensions. If any language can be too small, it's Scheme. You can't easily write meaningful, portable Scheme so each implementation is its own ecosystem.
20171002 08:10:42-!- atarocch [atarocch@nat/redhat/x-vphieengogdssvkd] has joined #wesnoth-dev
20171002 08:15:13< aeth> Hundreds of incompatible, mostly-slow implementations where portability is hard is not an advantage Scheme has over C++.
20171002 08:17:06< Necrosporus> Git stuff seems somewhat complex. For example if you fork a repository there is no button to update it to origin in github. It seems that you are supposed to first pull changes to your local machine and then to remote machine
20171002 08:17:18< Necrosporus> At least on github
20171002 08:17:40< AI0867> Necrosporus: that is because git does not gloss over the complexity of what "update to origin" means
20171002 08:18:03< AI0867> It's simple if the repositories are in the same state, or one of the two has updated
20171002 08:18:04< JyrkiVesterinen> Git is a complex version control system. Overkill for nearly all projects out there.
20171002 08:18:08< AI0867> but what if both have updated?
20171002 08:18:35< AI0867> the svn solution is to simply dump the remote changes into your working tree and let you handle the conflicts
20171002 08:18:35< JyrkiVesterinen> Don't forget that Git was originally created for the purpose of developing the Linux kernel.
20171002 08:18:39< AI0867> yes
20171002 08:19:47< Necrosporus> Anyway it seems easier to just use github web editor
20171002 08:19:55< aeth> JyrkiVesterinen: Git might be overkill, but working with a simple subset of git is easy
20171002 08:22:00< Necrosporus> In case of github it might be easier to simply remove repo and fork again
20171002 08:26:02< irker247> wesnoth: Arcanister wesnoth:master c1ed69ff1218 / .gitignore: Add *.so to .gitignore https://github.com/wesnoth/wesnoth/commit/c1ed69ff1218cb7eb46f61006c184f4ab14ed896
20171002 08:49:42-!- mjs-de [~mjs-de@185.22.140.184] has quit [Remote host closed the connection]
20171002 08:59:13-!- Kwandulin [~Kwandulin@pD9FD50F1.dip0.t-ipconnect.de] has quit [Quit: [endlevel]]
20171002 09:01:09-!- travis-ci [~travis-ci@ec2-54-205-64-34.compute-1.amazonaws.com] has joined #wesnoth-dev
20171002 09:01:10< travis-ci> AI0867/wesnoth#21 (whitelist-filename-chars - 786c419 : Alexander van Gessel): The build has errored.
20171002 09:01:10< travis-ci> Build details : https://travis-ci.org/AI0867/wesnoth/builds/282141661
20171002 09:01:10-!- travis-ci [~travis-ci@ec2-54-205-64-34.compute-1.amazonaws.com] has left #wesnoth-dev []
20171002 09:44:08-!- travis-ci [~travis-ci@ec2-54-198-24-151.compute-1.amazonaws.com] has joined #wesnoth-dev
20171002 09:44:09< travis-ci> AI0867/wesnoth#21 (whitelist-filename-chars - 786c419 : Alexander van Gessel): The build passed.
20171002 09:44:10< travis-ci> Build details : https://travis-ci.org/AI0867/wesnoth/builds/282141661
20171002 09:44:10-!- travis-ci [~travis-ci@ec2-54-198-24-151.compute-1.amazonaws.com] has left #wesnoth-dev []
20171002 09:44:48-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has quit [Quit: .]
20171002 09:50:16< Necrosporus> What version of libreadline is included in Windows 1.13.8 release?
20171002 10:09:43-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 10:21:58< Necrosporus> Is it possible to update libraries coming with Wesnoth for Windows? Should be easier than for Visual Studio since they are compiled with MinGW
20171002 10:23:04< Necrosporus> In particular libreadline to fix @1708
20171002 10:23:20< Necrosporus> Which is called history5.dll for some reason
20171002 10:32:23< irker247> wesnoth: Nils Kneuper wesnoth:master 4b5acb720f03 / po/wesnoth-sota/en_GB.po: updated British English translation https://github.com/wesnoth/wesnoth/commit/4b5acb720f03eb4b1436dc32a1250c893408292b
20171002 10:37:42-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has joined #wesnoth-dev
20171002 10:39:27-!- mkdroid [~null@unaffiliated/matthiaskrgr] has joined #wesnoth-dev
20171002 10:40:02< JyrkiVesterinen> loonycyborg: You're responsible for Windows builds, can you deal with Necrosporus's messages above?
20171002 10:41:17-!- mkdroid [~null@unaffiliated/matthiaskrgr] has quit [Client Quit]
20171002 10:46:01< Necrosporus> It seems AI0867 has taken my idea from earlier seriously. Do you think campaignd should also reject filenames ending with dots?
20171002 10:46:43< Necrosporus> In windows they silently overwrite files with same name but without dot currently, as Jyrki indicated
20171002 10:50:44-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20171002 10:51:17-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 11:00:58< loonycyborg> Necrosporus: I've used a prebuilt binary of that lib from readline
20171002 11:01:12< loonycyborg> I can update it
20171002 11:01:16< Necrosporus> From gnuwin32.sourceforge.net ?
20171002 11:01:22< loonycyborg> iirc yes
20171002 11:03:21< Necrosporus> It's still 5.0 there 
20171002 11:04:00< loonycyborg> is there some place with newer version?
20171002 11:04:06< loonycyborg> or I'd need to build my own?
20171002 11:06:32< Necrosporus> I think the best solution would be to create a set of scripts to build wesnoth on windows from scratch including all libs
20171002 11:08:29< JyrkiVesterinen> That would be quite difficult. GTK+ stack in particular (Pango, etc.) is very hard to build.
20171002 11:08:39< loonycyborg> I'm planning to eventually move to http://www.msys2.org/
20171002 11:09:12< loonycyborg> but I had some issues with it, I'll look into it more thoroughly after 1.14 is released
20171002 11:09:28< loonycyborg> for now I'll maintain my own wesnoth-deps sdk
20171002 11:09:32< Necrosporus> It is some fork not vanilla mingw
20171002 11:09:49< Necrosporus> MinGW itself has package repo, right?
20171002 11:09:54< loonycyborg> no
20171002 11:10:00< loonycyborg> at least
20171002 11:10:07< loonycyborg> it doesn't have a package manager
20171002 11:10:26< loonycyborg> and msys2 uses a windows port of arch's package manager
20171002 11:15:16-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171002 11:58:53-!- vultraz [uid24821@wesnoth/developer/vultraz] has quit [Quit: Connection closed for inactivity]
20171002 11:59:49-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 12:23:41-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171002 12:59:16-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20171002 13:32:34-!- irker247 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20171002 14:14:34-!- mjs-de [~mjs-de@b9168e98.cgn.dg-w.de] has joined #wesnoth-dev
20171002 14:15:09-!- JyrkiVesterinen [~JyrkiVest@195-192-251-124.s1networks.fi] has quit [Quit: .]
20171002 14:16:48-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 14:31:12-!- mjs-de [~mjs-de@b9168e98.cgn.dg-w.de] has quit [Quit: On the road again]
20171002 14:34:12-!- Oebele [~quassel@143.177.58.202] has joined #wesnoth-dev
20171002 14:59:49< Necrosporus> loonycyborg, no, I cannot find any other source to install Readline 6+ on Windows, except cygwin and msys2
20171002 15:00:07-!- sevu [~Shiki@p54854DEF.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171002 15:00:33< loonycyborg> seems I managed to build libhistory with crocc-compiler
20171002 15:00:37< loonycyborg> *cross
20171002 15:01:06< loonycyborg> but readline itself doesn't build
20171002 15:01:13< loonycyborg> and it's not needed for wesnoth anyway
20171002 15:01:34< loonycyborg> readline-7.0 that is
20171002 15:01:58< loonycyborg> I tried readline-6.3.0 but it's configure script doesn't work for cross-compilers
20171002 15:09:47< Necrosporus> are libhistory and readline different?
20171002 15:11:00< Necrosporus> loonycyborg, what's wrong with msys2 anyway?
20171002 15:11:40< loonycyborg> yes, those are two libs that are part of readline package
20171002 15:13:07< loonycyborg> it's seen by scons as different platform than win32 thus breaking a lot of platform specific logic
20171002 15:17:52-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20171002 15:30:28-!- atarocch [atarocch@nat/redhat/x-vphieengogdssvkd] has quit [Remote host closed the connection]
20171002 15:31:48-!- atarocch [atarocch@nat/redhat/x-qnksgyvzcmnvmmir] has joined #wesnoth-dev
20171002 15:58:22< Necrosporus> I wonder why doesn't gfgtdf appear in IRC
20171002 16:01:15-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has joined #wesnoth-dev
20171002 16:02:32-!- atarocch [atarocch@nat/redhat/x-qnksgyvzcmnvmmir] has quit [Ping timeout: 260 seconds]
20171002 16:19:47-!- mkdroid [~null@unaffiliated/matthiaskrgr] has joined #wesnoth-dev
20171002 16:20:19< Necrosporus> Why is not lua console like normal console? I mean you have separate input field instead of entering query in same console field where output is displaeyed after prompt character
20171002 16:35:27< sevu> why not?
20171002 16:37:29-!- sevu [~Shiki@p54854DEF.dip0.t-ipconnect.de] has quit [Quit: Verlassend]
20171002 16:49:10-!- mkdroid [~null@unaffiliated/matthiaskrgr] has quit [Remote host closed the connection]
20171002 16:55:08-!- atarocch [~atarocch@93.56.164.28] has joined #wesnoth-dev
20171002 16:58:34-!- atarocch [~atarocch@93.56.164.28] has quit [Remote host closed the connection]
20171002 17:02:55-!- atarocch [~atarocch@93.56.164.28] has joined #wesnoth-dev
20171002 17:34:14-!- sevu [~Shiki@p54854DEF.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171002 17:35:23-!- mjs-de [~mjs-de@x5ce48ae1.dyn.telefonica.de] has joined #wesnoth-dev
20171002 17:37:34-!- gfgtdf [~chatzilla@x4e3681e3.dyn.telefonica.de] has joined #wesnoth-dev
20171002 17:43:14-!- sevu [~Shiki@p54854DEF.dip0.t-ipconnect.de] has quit [Quit: Verlassend]
20171002 18:10:18-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity]
20171002 18:11:58< AI0867> 12:46 < Necrosporus> In windows they silently overwrite files with same name but without dot currently, as Jyrki indicated ← really? foo.txt == foo.txt. according to windows?
20171002 18:12:09< AI0867> that seem more stupid that usual
20171002 18:12:45< Necrosporus> AI0867, exactly
20171002 18:13:06< Necrosporus> if you have both foo.txt and foo.txt. in addon then you get only foo.txt with foo.txt's content
20171002 18:13:23< Necrosporus> Though it depends on what function was used to open file
20171002 18:13:37< Necrosporus> A windows archiver did create both files, with dot and without
20171002 18:14:04< Necrosporus> * with content from foo.txt.
20171002 18:14:32< Necrosporus> Same goes if you have files with different capitalization. You will get name of one file with content of other file
20171002 18:20:33-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20171002 18:21:01< AI0867> yeah, I almost have a PR for that bit
20171002 18:44:35-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171002 18:44:41-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171002 18:46:48-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 19:17:57< gfgtdf> AI0867: your commit has a typo 'conlict'
20171002 19:18:56< gfgtdf> AI0867: maybe you coudl add the trailing . chack to 2070 ? it's probably as easy as changing name == "." to name.back() == '.'
20171002 19:25:10< AI0867> gfgtdf: I was making that a separate one, but I suppose I could add it
20171002 19:26:52< gfgtdf> AI0867: oh ok i was just about to do it but then i noticed that it'd casue merge conflicts with our pr so i decided to ask you instead
20171002 19:27:55< Necrosporus> though perhaps wesnoth unpacking could be changed to actually create files with dots and so on
20171002 19:33:00-!- irker711 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20171002 19:33:00< irker711> wesnoth: Elvish_Hunter wesnoth:master aa4024fda987 / data/tools/wmllint: wmllint: remove UTF-8 BOM characters from cfg files https://github.com/wesnoth/wesnoth/commit/aa4024fda987094efa66e1b85f08814d3febcfec
20171002 19:33:00< irker711> wesnoth: Elvish_Hunter wesnoth:master f6fbd6351a3b / data/tools/ (15 files in 3 dirs): Added missing Python shebangs to several scripts https://github.com/wesnoth/wesnoth/commit/f6fbd6351a3b25345e62a6ef55518728d823b77f
20171002 19:34:18< gfgtdf> Necrosporus: it coudlbe possible but it's probaly too complicated, our time could be better spent on ither things
20171002 19:40:49-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20171002 19:41:23-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20171002 19:49:53-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection]
20171002 19:50:00-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev
20171002 19:55:11< AI0867> gfgtdf: see #2070
20171002 20:03:23-!- sigurdfd [~SigurdFD@dynamic-acs-72-23-110-196.zoominternet.net] has joined #wesnoth-dev
20171002 20:05:29-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has quit [Remote host closed the connection]
20171002 20:18:06< zookeeper> sigurdfd, wtf, you mean zorlan's portrait has never actually been used in-game before?
20171002 20:18:33< sigurdfd> apparently, yes
20171002 20:19:50< zookeeper> oh, it has been. but beetlenaut broke it for some reason in 8dc964fa9254 in 2011, and it's been unused since then -.-
20171002 20:20:32-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20171002 20:20:39< sigurdfd> that must be prior to 1.10.7, I just checked that and 1.12
20171002 20:23:54< mattsc> zookeeper: I just commented on #1955 (Liberty Hide and Seek bug).  I cannot reproduce this by using your method of :cl-ing into the scenario and reloading (either right away or after a couple turns).
20171002 20:25:15-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20171002 20:26:20< zookeeper> mattsc, humm.
20171002 20:26:45< zookeeper> on master, i presume?
20171002 20:27:01< mattsc> Yes. compiled sometime yesterday.
20171002 20:30:11 * zookeeper tries again
20171002 20:31:12-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has joined #wesnoth-dev
20171002 20:31:16-!- sevu [~Shiki@p54854DEF.dip0.t-ipconnect.de] has joined #wesnoth-dev
20171002 20:31:35< zookeeper> hmh. doesn't happen for me either anymore, and my build is a little bit over a week old.
20171002 20:32:30< mattsc> So who gets the credit for fixing it?  ;)
20171002 20:34:03< zookeeper> oh, duh. i have the local WML fix to get it working in place. maybe i'll try _without_ that...
20171002 20:34:15< mattsc> hehe
20171002 20:34:39< mattsc> For the record, I just tried with 1.13.8 and it does happen for me there.  So it doesn’t seem something strange about my system.
20171002 20:35:26< EliDupree> Whee, I made a simple GUI to display game states from in my simplified AI simulations
20171002 20:35:26< zookeeper> still reproduces. i'll recompile and check again
20171002 20:35:35-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has quit [Ping timeout: 240 seconds]
20171002 20:35:46< mattsc> Hmm.
20171002 20:39:38< mattsc> zookeeper: okay, this is bizarre
20171002 20:39:55< mattsc> If I start Wesnoth in debug mode (-d CL option), I do not get this error.
20171002 20:40:01< zookeeper> :o
20171002 20:40:13< mattsc> If I start without and then activate it in the first scenario, I do get it.
20171002 20:40:45< mattsc> don’t ask me how I came up with the idea of testing that :P
20171002 20:41:13< zookeeper> impeccable intuition? :p
20171002 20:41:39< mattsc> Yeah!  I’ll claim that one!
20171002 20:41:53< gfgtdf> zookeeper: does startign it debug mode fix it for you too ?
20171002 20:44:46< mattsc> So, I’ve tried back and forth 3 more times, that does indeed seem to be what is making the difference.
20171002 20:45:06< gfgtdf> hmm ok i can confimr what mattsc said
20171002 20:45:11-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has joined #wesnoth-dev
20171002 20:45:28< zookeeper> yes. if i start with -d, the problem disappears
20171002 20:45:38< zookeeper> pretty cool, i'd say
20171002 20:46:21< mattsc> Yeah, that’s almost a Frankenbug.  If you watch it (by activating debug know-all mode), it disappears.
20171002 20:46:30< mattsc> Ugh.
20171002 20:46:32< mattsc> Heisenbug.
20171002 20:46:45< zookeeper> heisenstein
20171002 20:46:53< mattsc> something like that
20171002 20:46:57-!- travis-ci [~travis-ci@ec2-54-198-24-151.compute-1.amazonaws.com] has joined #wesnoth-dev
20171002 20:46:58< travis-ci> AI0867/wesnoth#22 (case-insensitive-duplicates - d6fc541 : Alexander van Gessel): The build passed.
20171002 20:46:58< travis-ci> Build details : https://travis-ci.org/AI0867/wesnoth/builds/282392174
20171002 20:46:58-!- travis-ci [~travis-ci@ec2-54-198-24-151.compute-1.amazonaws.com] has left #wesnoth-dev []
20171002 20:48:05< zookeeper> i don't know the AI code, but that sounds like something that ought to tremendously help in finding places to investigate
20171002 20:58:39< mattsc> Well, I found the difference between the two (in the saves, not in the code).
20171002 20:59:07< mattsc> In the one that works, there is an empty (except for the name) [stage] tag inside the [ai] for the side.
20171002 20:59:38< mattsc> In the one displayng the bug, there is no [stage] tag whatsoever, which means upon loading it will be replaced by the default stage.
20171002 21:00:28< mattsc> So as far as I can tell, the problem happens either before or during loading.
20171002 21:00:51< mattsc> s/ saving/loading
20171002 21:01:04< gfgtdf> mattsc: iirc the problem was that in 1.12 the 'add default stage' onyl applies when creatign new games not when loading savefiles, but in 1.13. this bahviour was changed
20171002 21:01:08< mattsc> or the other way around, brain not working today
20171002 21:02:41-!- mjs-de [~mjs-de@x5ce48ae1.dyn.telefonica.de] has quit [Remote host closed the connection]
20171002 21:03:32< mattsc> gfgtdf: yeah. might be; but there’s also another difference:
20171002 21:04:05< mattsc> When I start the game in debug mode, the side’s [ai] tag contains an empty [stage] tag
20171002 21:04:27< mattsc> When I start not in debug mode, there’s no [stage] tag
20171002 21:04:59< mattsc> In the latter case, the default stage gets inserted when loading, in the former it does not.
20171002 21:05:56< mattsc> So the problem is that, in “normal” mode, the idle AI does not actually get inserted into the side’s [ai] tag.
20171002 21:06:28< mattsc> Sounds like celticminstrel should know about this.
20171002 21:07:36< gfgtdf> but [stage] not beeing inserted in 'normal' mode is iirc also the bahviour of 1.12. and it works there
20171002 21:08:11< mattsc> Well, yes (or maybe), because one bug or feature cancels out another.
20171002 21:10:41-!- Oebele [~quassel@143.177.58.202] has quit [Remote host closed the connection]
20171002 21:11:38< gfgtdf> yes i dont know the code to see here the actual bug is, i can just analyze in whihc part the behvuiour differs from 1.12
20171002 21:11:44< gfgtdf> where*
20171002 21:15:24< mattsc> Me neither.  Also, it seems that any individual part of this behavior could be considered intentional rather than a bug, but there seem to be at least two inconsistencies:
20171002 21:15:46< mattsc> 1. How the idle AI is inserted into the config is different between debug and no-debug mode
20171002 21:16:17< mattsc> 2. When there’s no [stage], the default AI gets inserted when loading, but not at the beginning of the scenario.
20171002 21:16:36< mattsc> The second might be a result of the first rather than a separate thing though.
20171002 21:17:37< mattsc> Anyway, I feel like I have done what I can here, somebody else will have to look into how to fix this.
20171002 21:39:06-!- gfgtdf [~chatzilla@x4e3681e3.dyn.telefonica.de] has quit [Ping timeout: 240 seconds]
20171002 21:58:45< mattsc> celticminstrel: also let me know whether you want me to fix the m vs. M bug in the MAI found yesterday
20171002 22:01:08-!- APic [apic@apic.name] has quit [Quit: Scheduled Downtime. See You later, Gals and/or Pals.  ☺]
20171002 22:13:33-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20171002 22:14:17-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20171002 22:34:03-!- irker711 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20171002 22:34:06-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 240 seconds]
20171002 22:55:55-!- ACTPic [apic@apic.name] has joined #wesnoth-dev
20171002 22:59:55-!- atarocch [~atarocch@93.56.164.28] has quit [Ping timeout: 258 seconds]
20171002 23:02:57-!- ACTPic [apic@apic.name] has quit [Quit: leaving]
20171002 23:08:56-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20171002 23:13:07-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Read error: Connection reset by peer]
20171002 23:13:28-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
--- Log closed Tue Oct 03 00:00:19 2017