--- Log opened Fri Sep 23 00:00:33 2016
20160923 00:28:10-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 265 seconds]
20160923 00:41:59-!- irker884 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160923 00:46:55-!- celmin [~celticmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160923 00:58:41-!- TC02 [~quassel@venus.arosser.com] has quit [Remote host closed the connection]
20160923 00:58:53-!- TC02 [~quassel@venus.arosser.com] has joined #wesnoth-dev
20160923 01:09:17-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 265 seconds]
20160923 01:09:22-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 01:14:21-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has quit [Read error: Connection reset by peer]
20160923 01:15:02-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 01:18:09-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 01:22:55-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds]
20160923 01:22:55-!- gfgtdf [~chatzilla@x4e368599.dyn.telefonica.de] has quit [Ping timeout: 260 seconds]
20160923 01:23:37-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 01:31:09-!- enchi [enchilado@defocus/yummy/enchilado] has quit [Ping timeout: 276 seconds]
20160923 01:33:32-!- enchi [enchilado@defocus/yummy/enchilado] has joined #wesnoth-dev
20160923 01:36:16-!- gfgtdf [~chatzilla@x4e36304c.dyn.telefonica.de] has joined #wesnoth-dev
20160923 01:38:49-!- stikonas_ is now known as stikonas
20160923 02:09:06-!- gfgtdf [~chatzilla@x4e36304c.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 48.0.2/20160823121617]]
20160923 02:12:29-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has joined #wesnoth-dev
20160923 02:21:00-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160923 02:24:47-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 02:25:35-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 03:01:25-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 265 seconds]
20160923 03:04:43-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 03:16:09-!- celmin [~celticmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The minstrel departs, to spread the music to the masses!]
20160923 03:22:55< EliDupree> I made a test scenario that automatically casts 3 random spells every turn, and it helped me catch some bugs :-)
20160923 03:23:12< celticminstrel> Yay!
20160923 03:24:31< EliDupree> It doesn't have great code coverage, though, because it doesn't visit any of the UI for choosing/targeting the spells, and only about half of the spells are ones that it makes sense that cast in random situations
20160923 03:29:28< EliDupree> For instance, all of the attack spells require there to be a specific wizard unit who is the one attacking. Although I can probably make special cases in my code for attacks with no source.
20160923 03:31:12-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Ping timeout: 240 seconds]
20160923 03:31:57-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160923 03:33:24< EliDupree> Or even just make AI wizard units and force them through the entire normal process of casting a spell...
20160923 03:33:45< EliDupree> Aside from showing the dialogs of course
20160923 03:55:50< tad_> Sheesh. Testing this is so slow. Who decided that starting a level with over 75 enemy units for 8 enemy leaders was a good idea? And passing them gold to create another 12 or so when you get close to the leaders?
20160923 03:56:25< tad_> I'm thinking the 'best' fix to balance THoT S12 is to delete it and put S11 back as the final scene.
20160923 04:08:16-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has quit [Ping timeout: 255 seconds]
20160923 04:08:36-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has joined #wesnoth-dev
20160923 04:25:30-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity]
20160923 04:26:06-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160923 04:32:34-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has quit [Ping timeout: 255 seconds]
20160923 04:32:54-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has joined #wesnoth-dev
20160923 04:56:23-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 05:00:38-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 244 seconds]
20160923 05:13:54-!- vultraz [~chatzilla@124.109.10.167] has joined #wesnoth-dev
20160923 05:13:54-!- vultraz [~chatzilla@124.109.10.167] has quit [Changing host]
20160923 05:13:54-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160923 05:14:22-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20160923 05:14:22-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20160923 05:16:41< vultraz> power been out most of the day
20160923 05:16:41< vultraz> :|
20160923 05:17:32< vultraz> now, let us see what my build did earlier..
20160923 05:17:34< vultraz> g++.exe: error: libgomp.spec: No such file or directory
20160923 05:17:35< vultraz> hmm
20160923 05:17:43-!- Kwandulin [~Miranda@p200300760F2C717FAC6B034096777966.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160923 05:17:45< vultraz> ok, so it looks like this version must have omp enabled :/
20160923 05:17:47< vultraz> t's not good
20160923 05:20:18< vultraz> oh wait
20160923 05:20:24< vultraz> other way around
20160923 05:20:26< vultraz> durr
20160923 05:21:34< vultraz> tdm had an option for omp, dunno if plain gcc does..
20160923 05:24:18< vultraz> shadowm: do you build with omp?
20160923 05:30:24< vultraz> guess the simplest solution for me is to not do so for now
20160923 05:37:25< vultraz> I'll disable omp in my cb build for now
20160923 05:37:33< vultraz> onto building boost 
20160923 06:16:55-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 260 seconds]
20160923 06:24:28-!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Ping timeout: 265 seconds]
20160923 06:31:24-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160923 06:33:36< vultraz> ok, 1.61 build, -fopenmp removed... let's see if this works
20160923 06:40:50-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 260 seconds]
20160923 06:44:38-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 06:44:46-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160923 06:45:41-!- vultraz [~chatzilla@124.109.10.167] has joined #wesnoth-dev
20160923 06:45:41-!- vultraz [~chatzilla@124.109.10.167] has quit [Changing host]
20160923 06:45:41-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160923 06:48:34-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 240 seconds]
20160923 06:48:34-!- wedge010 is now known as wedge009
20160923 06:48:48-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 244 seconds]
20160923 06:55:11-!- Kwandulin [~Miranda@p200300760F2C717FAC6B034096777966.dip0.t-ipconnect.de] has quit [Ping timeout: 250 seconds]
20160923 06:55:15-!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #wesnoth-dev
20160923 06:57:18< vultraz> oh come on :|
20160923 06:58:24< vultraz> how am i getting undefined references to omo stuff when I'm not building with omp :|
20160923 06:58:31< vultraz> s/omo/omp
20160923 06:59:03< vultraz> er
20160923 06:59:04< vultraz> hm..
20160923 06:59:18< vultraz> but that's not stuff from inside wesnoth
20160923 07:01:03< vultraz> not.. sure what to do here :|
20160923 07:02:33< vultraz> I mean, the solution to this would be to build WITH -fopenmp... but I can't do that since this build of gcc doesn't have omp :|
20160923 07:05:23< vultraz> :\
20160923 07:05:38< Sirp> we have openmp support in Wesnoth?
20160923 07:05:42< Sirp> what do we do with it?
20160923 07:06:04< vultraz> not much
20160923 07:10:35-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 260 seconds]
20160923 07:11:37< celticminstrel> That is a very good question.
20160923 07:12:38< Sirp> it seems a strange choice
20160923 07:14:38< celticminstrel> The XCode build does not use it. The MSVC build does. Not sure about scons and CMake.
20160923 07:15:40< celticminstrel> I really don't know. Maybe someone thought it improved performance or something.
20160923 07:15:50-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160923 07:16:16-!- Kwandulin [~Miranda@p200300760F2C7109AC6B034096777966.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160923 07:17:28< vultraz> Sirp: it's mostly used for a not-really efficient performance boost in the display class in lieu of proper hardware acceleration (wesnoth still uses software rendering)
20160923 07:19:53-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has joined #wesnoth-dev
20160923 07:20:15< Sirp> vultraz: so sure but openmp is just a library for parallelism. I'd have imagined we'd just use the same concurrency libs as the rest of Wesnoth uses. (which I think is still the SDL threading primitives?P)
20160923 07:21:51< celticminstrel> I don't know, does anything use SDL threads? The loading screen is using Boost threads.
20160923 07:22:15< celticminstrel> There is SDL timer use (no idea if that's related to threads).
20160923 07:22:17< Sirp> oh sure Boost threads would be cool
20160923 07:22:20< Sirp> no it's not
20160923 07:22:37< Sirp> Boost threads or SDL threads or whatnot are fairly 'common' to use in games.
20160923 07:22:47< celticminstrel> Pretty sure there's not much parallelism in general in Wesnoth...
20160923 07:22:50-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 260 seconds]
20160923 07:22:51< Sirp> openmp is if you want to program a simulation of a nuclear explosion.
20160923 07:23:00< celticminstrel> Haha, what?
20160923 07:24:59< Sirp> Well, some people like to program simulations of nuclear explosions. I mean, they probably don't *like* to do it. But it's their job. And usually if you want to do something like that, parallel processing helps a lot. OpenMP was created for that kind of purpose -- massively parallel numerical computation.
20160923 07:25:19< celticminstrel> Right, that makes sense.
20160923 07:25:23< Sirp> It can be used in a game like Wesnoth, I suppose, it just seems like a slightly odd choice.
20160923 07:49:13-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20160923 07:52:08-!- boucman_work [~boucman@209.57.66.86.rev.sfr.net] has joined #wesnoth-dev
20160923 07:56:29-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!]
20160923 07:59:20< zookeeper> Sirp, do boost/SDL threads allow easy parallelization of loops? that's what seemed like the most convenient part of OMP to me.
20160923 08:02:18< zookeeper> i used a lot of that when trying to speed up some terrain graphics -related stuff, although none of that ever ended up in master.
20160923 08:19:05< JyrkiVesterinen> zookeeper: Boost threads indeed don't offer the easy parallelization that OpenMP has.
20160923 08:19:49< JyrkiVesterinen> Although the convenience of OpenMP has cost us with the issues OpenMP has caused. Such as vultraz having linking problems now.
20160923 08:20:22< JyrkiVesterinen> Or the fact that Wesnoth needs to restart itself on launch to prevent constant 100% CPU usage.
20160923 08:21:07< JyrkiVesterinen> https://github.com/wesnoth/wesnoth/blob/d1053abe7195e717ca6eaebd88a5356819ee7575/src/wesnoth.cpp#L963-L988
20160923 08:24:04-!- Kwandulin [~Miranda@p200300760F2C7109AC6B034096777966.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160923 08:24:06< zookeeper> neat
20160923 08:32:54-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 08:37:20-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 265 seconds]
20160923 08:58:37-!- Kwandulin [~Miranda@p200300760F2C710964B56865253C13C0.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160923 09:03:19< wedge009> I just noticed the nice background image for the MP create screen.
20160923 09:04:04-!- shadowm [~ignacio@wesnoth/developer/shadowm] has quit []
20160923 09:09:25-!- shadowm [~ignacio@wesnoth/developer/shadowm] has joined #wesnoth-dev
20160923 09:10:26-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has quit [Quit: .]
20160923 09:13:18< shadowm> No, I don't normally use OpenMP.
20160923 09:21:07< shadowm> Soliton: Neither dash nor zsh call the exit trap in SIGTERM as far as I can tell.
20160923 09:21:20< shadowm> *on
20160923 09:25:10-!- irker051 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160923 09:25:10< irker051> wesnoth: Ignacio R. Morelle wesnoth:master cfd4a0438181 / utils/wesnoth-defaults: Make wesnoth-defaults handle arguments and paths with whitespace correctly https://github.com/wesnoth/wesnoth/commit/cfd4a04381813610ac0f398c6b914e68b8029780
20160923 09:26:11< shadowm> Oh, right, my Mingw-w64 cross-build uses OpenMP, actually.
20160923 09:26:56< shadowm> I'm not the right person to ask if you have problems building with it, anyway.
20160923 09:27:30< shadowm> vultraz? I don't remember who was asking. Don't have time to check the logs. Bye.
20160923 09:34:39-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 250 seconds]
20160923 09:35:28< Soliton> shadowm: indeed looks like that behaviour is not consistent across shells. i guess POSIX is not perfectly clear on how the EXIT trap should work... nevermind then.
20160923 09:36:02-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160923 09:42:11< vultraz> just gonna assume here that i can't copy the libgomp.spec file from the gcc 5.1.0 folder
20160923 09:44:45< vultraz> so the question is, why is the compiler looking for omp if I disabled the option in cb :/
20160923 09:45:07< vultraz> ohhh
20160923 09:45:09< vultraz> wait
20160923 09:45:14< vultraz> stupid of me
20160923 09:45:19< vultraz> I removed it from the *linker* settings
20160923 09:45:26< vultraz> but not the *compiler settings*
20160923 09:52:14< irker051> wesnoth: Matthias Krüger wesnoth:master c9a3c0a76795 / changelog data/tips.cfg players_changelog: tips: add tip to recall units instead of recruiting them if costs > 20 https://github.com/wesnoth/wesnoth/commit/c9a3c0a7679525449315cf7af8cc8fbd80472b7b
20160923 09:52:16< irker051> wesnoth: Charles Dang wesnoth:master d7833f196c78 / changelog data/tips.cfg players_changelog: Merge pull request #795 from matthiaskrgr/tip https://github.com/wesnoth/wesnoth/commit/d7833f196c78022460b1db1781c6ef1338999854
20160923 09:56:07-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has quit [Ping timeout: 255 seconds]
20160923 09:56:27-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has joined #wesnoth-dev
20160923 10:00:21-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev
20160923 10:01:00-!- Appleman1234 [~Appleman1@KD119104043017.au-net.ne.jp] has quit [Ping timeout: 265 seconds]
20160923 10:10:05-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 10:11:31-!- Appleman1234 [~Appleman1@KD119104042061.au-net.ne.jp] has joined #wesnoth-dev
20160923 10:14:16< vultraz> ok, wesnoth seems to build with boost 1.61 and my gcc 6.2.1 drop-in
20160923 10:14:21< vultraz> but wesnothd still failing :/
20160923 10:14:33-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 248 seconds]
20160923 10:16:12< vultraz> this time with even more errors :|
20160923 10:18:55< vultraz> hmm
20160923 10:20:09< vultraz> ahhh ok
20160923 10:20:20< vultraz> I forgot to add server_base.*pp to cb :|
20160923 10:21:08-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 10:21:25< vultraz> I hope all this wasn't caused by a missing file :|
20160923 10:21:47< vultraz> well, still, I finally figured out why I'd been getting all those errors with previous boost versions
20160923 10:21:55< vultraz> except wesnoth doesn't start now :)
20160923 10:21:57< vultraz> fun!
20160923 10:22:23< vultraz> #1 0x11f699d	_GLOBAL__sub_I__ZN5boost6detail16thread_data_baseD2Ev () (??:??)
20160923 10:22:24< vultraz> #2 0x401560	__gcc_register_frame () (??:??)
20160923 10:22:26< vultraz> #3 0x77c962c4	KERNEL32!BaseThreadInitThunk() (C:\WINDOWS\SysWoW64\kernel32.dll:??)
20160923 10:22:31< vultraz> well this doesn't look good P_P
20160923 10:25:42-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 264 seconds]
20160923 10:27:24< vultraz> gonna pull up a debug build
20160923 10:27:33< vultraz> even though i doubt it'll tell me anything
20160923 10:35:35-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev
20160923 10:53:47< vultraz> yeah that gives me absolutely nothing :|
20160923 10:53:52< vultraz> anyone have any thoughts?
20160923 11:01:15-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160923 11:02:29-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 244 seconds]
20160923 11:02:59-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev
20160923 11:25:24< vultraz> maybe it' something screwed up with the compiler :/
20160923 11:26:13< vultraz> google isn't helping much
20160923 11:28:41< wedge009> Huh, celmin's on holidays right?
20160923 11:28:44< wedge009> Or otherwise away.
20160923 11:29:36< wedge009> Prods celmin to have a look at https://gna.org/bugs/?25093 when available. No rush, just pointing it out.
20160923 11:36:02< vultraz> could try reverting to 5.1.0 and rebuilding boost AGAIN but I'd rather not :/
20160923 11:36:20< vultraz> would rather... not
20160923 11:39:59< vultraz> let's try
20160923 11:40:27< vultraz> ok, let's try separating from the tdm install entirely 
20160923 11:41:22< vultraz> new folder, using the MinGW-w64 download
20160923 11:41:43< vultraz> (which seems to be separate from mingw..)
20160923 12:04:34-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 240 seconds]
20160923 12:07:09< vultraz> doesn't help
20160923 12:09:26-!- mkdroid [~null@unaffiliated/matthiaskrgr] has joined #wesnoth-dev
20160923 12:09:26-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 12:11:33-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev
20160923 12:13:49< loonycyborg> vultraz: what was your issue again?
20160923 12:14:14< vultraz> uhhh
20160923 12:14:33-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 272 seconds]
20160923 12:15:01< vultraz> well the original was undefined references during linking while building wesnothd, which I now think might have simply been caused by a missing fiel
20160923 12:15:03< vultraz> file*
20160923 12:15:33< vultraz> in the project
20160923 12:16:03< vultraz> anyway, I've been attempting to update to gcc6 on top of a tdm install and though it builds, wesnoth doesn't launch
20160923 12:16:58< vultraz> thinking i'll just have to revert to tdm gcc 5.1.0 and rebuild boost AGAIN :|
20160923 12:17:17< loonycyborg> don't you have your old boost still?
20160923 12:18:29< vultraz> eh, yeah
20160923 12:18:53< vultraz> but now that I know why I'd been having those issues with 1.60+ I might as well upgrade
20160923 12:20:09-!- boucman_work [~boucman@209.57.66.86.rev.sfr.net] has quit [Ping timeout: 265 seconds]
20160923 12:21:16< vultraz> hm.. the tdm installer does have an option for a 32-and-64 bit install based on mingw-w64...
20160923 12:22:08< vultraz> eh, nah
20160923 12:22:15< vultraz> pretty sure i've had problems before :/
20160923 12:23:16< vultraz> would be nice to use gcc6 but it seems it's not worth the bloody effort
20160923 12:24:56< vultraz> building boost isn't that difficult, thankfully
20160923 12:34:58-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 244 seconds]
20160923 12:42:09< loonycyborg> lolwtf, boost.asio has two builtin ways to handle sigpipe issue, but not one of them applies to linux platforms due to conditional compilation
20160923 12:44:46< loonycyborg> one of them is for defined(__sun) || defined(__QNX__) || defined(__hpux) || defined(_AIX) \
20160923 12:44:46< loonycyborg>   || defined(__osf__)
20160923 12:45:08< loonycyborg> other is for defined(__MACH__) && defined(__APPLE__) || defined(__FreeBSD__)
20160923 12:48:02< vultraz> fun!
20160923 13:07:08< loonycyborg> oh wait
20160923 13:07:21< loonycyborg> there's a linux specific workaround too
20160923 13:07:26< loonycyborg> so why it doesn't work? :P
20160923 13:08:42-!- mkdroid [~null@unaffiliated/matthiaskrgr] has quit [Quit: I'll be back!]
20160923 13:14:27-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160923 13:25:11< vultraz> blah, more of this placeholder stuff
20160923 13:25:19< loonycyborg> oh I got it, boost's workaround is more targeted so it doesn't apply to my custom sendfile call
20160923 13:26:49< matthiaskrgr> I also had a look at wesnoth using the thread sanitizer now, but I'm not really familiar with it. here's some output when starting the game, in case it makes sense to anybody:  http://pastebin.com/uD5ynurd  
20160923 13:27:50< vultraz> data race?
20160923 13:27:53< vultraz> doesn't sound good
20160923 13:28:24< matthiaskrgr> (this is my first time looking at something via TSAN, basically)
20160923 13:28:43< vultraz> matthiaskrgr: btw, weren't you able to get that boost placeholders error to go away without editing the boost source (which is what some of us have been doing)?
20160923 13:29:01< matthiaskrgr> you mean the static assert? 
20160923 13:29:10< matthiaskrgr> I compiled with gcc instead of clang and that "fixed" it 
20160923 13:29:11< vultraz> yeah
20160923 13:29:14< vultraz> ah
20160923 13:29:17< vultraz> meh :/
20160923 13:29:24< matthiaskrgr> exactlay :( 
20160923 13:29:26< vultraz> edit the sources it is! :P
20160923 13:30:03< vultraz> for the record, it seems to trigger on bind_void calls... which might indicate bind_void should not be used
20160923 13:30:13< vultraz> but celticminstrel insist it's necessary for some compiler...
20160923 13:30:17< vultraz> compilers
20160923 13:30:24< matthiaskrgr> mmh
20160923 13:30:55< matthiaskrgr> I could upgrade to fedora25 beta at some point and hope it ships boost 1.61 with fixed headers 
20160923 13:32:25< vultraz> im building with boost 1.61 right now
20160923 13:32:29< vultraz> fresh boost build
20160923 13:32:35< vultraz> which is why i was getting this again
20160923 13:32:52< matthiaskrgr> oh?
20160923 13:32:59< matthiaskrgr> so it's still a problem with up to date boost?
20160923 13:33:01< vultraz> yeah
20160923 13:33:12< matthiaskrgr> :(
20160923 13:33:18< vultraz> i don't think it's boost, I think it's us 
20160923 13:33:50< matthiaskrgr> lol
20160923 13:34:00< matthiaskrgr> TSAN reports a potential deadlock when  quitting the game
20160923 13:34:05< vultraz> o_O
20160923 13:34:36< matthiaskrgr> http://pastebin.com/raw/F6YPwhCP 
20160923 13:34:39< matthiaskrgr> or whatever that means :D
20160923 13:35:14< vultraz> fun fun fun
20160923 13:35:40< matthiaskrgr> ^__^
20160923 13:37:43< vultraz> yay, I have a working wesnoth again!
20160923 13:38:56< vultraz> obviously I didn't have the right setup with that gcc 6 dropin
20160923 13:39:54< irker051> wesnoth: Charles Dang wesnoth:master 86e3161b29d5 / projectfiles/CodeBlocks/ (wesnoth.cbp wesnothd.cbp): Updated CB projfile (missing wesnothd source and Boost 1.61 linker target) https://github.com/wesnoth/wesnoth/commit/86e3161b29d5663836c9e6da9f8f3b95b115e6b7
20160923 13:40:09< vultraz> will need to get a new SDK package together 
20160923 13:55:11-!- boucman_work [~boucman@209.57.66.86.rev.sfr.net] has joined #wesnoth-dev
20160923 13:59:31< irker051> wesnoth: loonycyborg wesnoth:master 22fc02e9cbea / src/campaign_server/campaign_server.cpp: Fix bug #25102 https://github.com/wesnoth/wesnoth/commit/22fc02e9cbeaed322d719269b860fd9f55c4df3a
20160923 14:04:05< vultraz> I'm curious about the REGISTER_WIDGET3 macro..
20160923 14:05:00< vultraz> or REGISTER_WINDOW
20160923 14:05:53< vultraz> if I'm reading it correctly, they define structs with constructors and then create a static object instance of that struct in order to execute the constructor...
20160923 14:06:55< vultraz> oh, I guess that's needed since the substitution locations are out of function scope
20160923 14:07:01< vultraz> meaning you can't call the functions directly
20160923 14:07:32< vultraz> curious, I never really thought about that...
20160923 14:09:25< vultraz> actions are only inside functions 
20160923 14:09:47< loonycyborg> shadowm: fix for blocker is ready 
20160923 14:19:09-!- Kwandulin [~Miranda@p200300760F2C710964B56865253C13C0.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160923 14:21:00-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 14:42:04-!- boucman_work [~boucman@209.57.66.86.rev.sfr.net] has quit [Ping timeout: 240 seconds]
20160923 14:47:43-!- DeFender [~DeFender1@89-138-245-35.bb.netvision.net.il] has quit [Quit: I'm not back now.]
20160923 14:53:53-!- Kwandulin [~Miranda@p200300760F2C710905F9EDCF87D0796C.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160923 14:54:38< irker051> wesnoth: mattsc wesnoth:master d3a3825a26bb / data/ai/lua/ca_high_xp_attack.lua: High XP attack CA: don't attack hidden enemies https://github.com/wesnoth/wesnoth/commit/d3a3825a26bbebd49902bda91a82b196f9785040
20160923 14:55:33-!- Kwandulin_2 [~Miranda@p200300760F2C710940652F33CA7C58A8.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160923 14:58:23-!- Kwandulin [~Miranda@p200300760F2C710905F9EDCF87D0796C.dip0.t-ipconnect.de] has quit [Ping timeout: 250 seconds]
20160923 15:26:10-!- Ravana_ [~Ravana@unaffiliated/ravana/x-2327071] has quit [Read error: Connection reset by peer]
20160923 15:27:11-!- Ravana_ [~Ravana@unaffiliated/ravana/x-2327071] has joined #wesnoth-dev
20160923 15:32:52< irker051> wesnoth: mattsc wesnoth:master 20a8a2f45664 / data/ai/lua/ca_high_xp_attack.lua: High XP attack CA: fix rare case of AI choosing wrong weapon https://github.com/wesnoth/wesnoth/commit/20a8a2f456642cd7c351976e1981ff098d228353
20160923 15:33:51-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20160923 15:35:03-!- boucman_work [~boucman@209.57.66.86.rev.sfr.net] has joined #wesnoth-dev
20160923 15:40:41-!- boucman_work [~boucman@209.57.66.86.rev.sfr.net] has quit [Ping timeout: 240 seconds]
20160923 15:42:24-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160923 15:42:56-!- mkdroid [~null@unaffiliated/matthiaskrgr] has joined #wesnoth-dev
20160923 15:43:10-!- mkdroid [~null@unaffiliated/matthiaskrgr] has quit [Client Quit]
20160923 15:47:51-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 15:55:54< irker051> wesnoth: Charles Dang wesnoth:master 7a915f90357e / projectfiles/CodeBlocks/README.md: Slight readme update https://github.com/wesnoth/wesnoth/commit/7a915f90357eb9ffcbcd4fd597e335e2f6305f2d
20160923 15:56:00< vultraz> shadowm, loonycyborg: could one of you upload this to sourceforge? https://www.dropbox.com/s/np8ezg8h44j908o/CodeBlocksWinSDK_24_9_2016.7z?dl=0
20160923 15:59:11-!- gfgtdf [~chatzilla@x4e36304c.dyn.telefonica.de] has joined #wesnoth-dev
20160923 16:03:19-!- Bonobo [~Bonobo@2001:44b8:254:3200:5496:1151:bc7:bf7c] has quit [Ping timeout: 255 seconds]
20160923 16:04:36-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160923 16:05:05< vultraz> tad_: so, I've got my build working again
20160923 16:05:07< vultraz> :D
20160923 16:05:29< vultraz> turns out i was missing some files in the project, but I was able to update to boost 1.61, so not a total loss
20160923 16:05:42< tad_> Ah. Good.
20160923 16:05:59< tad_> And you're on gcc6 now, too?
20160923 16:06:02< vultraz> no
20160923 16:06:09< vultraz> the binary wouldn't start :|
20160923 16:06:17< vultraz> wesnoth binary*
20160923 16:07:02< vultraz> [21:22:20]	vultraz	#1 0x11f699d	_GLOBAL__sub_I__ZN5boost6detail16thread_data_baseD2Ev () (??:??)
20160923 16:07:03< vultraz> [21:22:20]	vultraz	#2 0x401560	__gcc_register_frame () (??:??)
20160923 16:07:05< vultraz> [21:22:20]	vultraz	#3 0x77c962c4	KERNEL32!BaseThreadInitThunk() (C:\WINDOWS\SysWoW64\kernel32.dll:??)
20160923 16:07:08< vultraz> figured it's not worth it
20160923 16:07:14< vultraz> so i reverted to tdm gcc 5.1
20160923 16:07:19< vultraz> and rebuild boost again
20160923 16:08:05< tad_> I suppose I should put together a HOWTO for Windows 10 AE on building a Windows version using the free command-line MS C compiler and another using WSL and gcc.
20160923 16:09:19< loonycyborg> vultraz: Can't you upload it youself?
20160923 16:09:27< vultraz> loonycyborg: no, no access
20160923 16:09:51< loonycyborg> I think shadowm can give you access
20160923 16:10:51< tad_> SysWoW64/kernel32.dll .. "System, Windows-on-Windows, 64-bit Kernel for 32-bit applications"
20160923 16:14:40< tad_> zookeeper: So I'm looking at THoT S12 (Underlevels) and reading the comment at the top of the file when it occurs to me: this is an add-on not in the original design, done by someone else.
20160923 16:15:47< tad_> zookeeper: Then I could the number of pre-placed units (70 or so) and the ones recruited by the enemies (another 80 or so) and a little light goes on 'No wonder it's so slow right from the start'
20160923 16:16:26< tad_> zookeeper: So .. question .. would there by any serious opposition to 'fixing' the scene using the 'rm' command?
20160923 16:16:39-!- Kwandulin_2 [~Miranda@p200300760F2C710940652F33CA7C58A8.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160923 16:16:53< vultraz> :P
20160923 16:17:59< vultraz> i could redesign it
20160923 16:18:50< tad_> vultraz: Well, I was thinking of removing the triggers for gold, forcing the enemies to work with what they have.
20160923 16:19:05< vultraz> boring with a map so big
20160923 16:19:19< tad_> vultraz: And I was thinking of shackling all those in the cages at the start.
20160923 16:20:12< tad_> vultraz: I usually use left-hand-rule to explore and on this map it means I have to expore 100% of the caverns to get to the win.
20160923 16:20:45< vultraz> why don't you just leave it
20160923 16:20:51< vultraz> I'll do a redesign sometime
20160923 16:21:19-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20160923 16:21:33< vultraz> what's next on your list?
20160923 16:21:38< tad_> vultraz: It **IS** winnable and, actually, not that hard once you realize you do NOT have a turn limit and the best advance is 1-forward 2-back kill all 2-forward, rinse, repeat
20160923 16:21:48< vultraz> yes
20160923 16:21:49< vultraz> I've won it
20160923 16:21:52< vultraz> in like 200 turns?
20160923 16:21:58< vultraz> ok maybe not that much
20160923 16:22:04< tad_> Sounds about right, though.
20160923 16:22:48< vultraz> I'd significantly shorten it
20160923 16:23:03< tad_> The first few dozens runs at it you feel like it's designed to be unwinnable and I imagine most people give up. And that's not good for the game's over-all reputation.
20160923 16:23:16< vultraz> I've played a lot of really good dungeon scenarios 
20160923 16:23:43< vultraz> so, what's next on the list?
20160923 16:23:52< vultraz> TRoW/
20160923 16:23:55< tad_> The common complaint there (another today on the forums) is "twisty passages, all different" (hahaha showing my age)
20160923 16:23:55< vultraz> ?
20160923 16:25:06< tad_> "Adventure" or "Cave" originally in FORTRAN. The game which started all this.
20160923 16:25:38< vultraz> .......wha??
20160923 16:26:14< tad_> Sorry. Anyway, the common complaint I see about cave scenes is the narrow passages.
20160923 16:26:26< shadowm> vultraz: What is your SF.net username?
20160923 16:26:29< vultraz> yeah, that can be tedious 
20160923 16:26:40< vultraz> shadowm: uhh... hang on...
20160923 16:28:00< tad_> vultraz: Actually, going down the list, next is Descent into Darkness.
20160923 16:28:21< vultraz> shadowm: oh, it's just 'vultraz'.
20160923 16:28:42< vultraz> tad_: ah, another campaign i cleaned up
20160923 16:28:46< vultraz> expect clean code
20160923 16:29:04< shadowm> vultraz: You are sure, right?
20160923 16:29:04< tad_> vultraz: Don't be surprised if I find bugs ... :P
20160923 16:29:38< shadowm> Yeah, looks like you all right.
20160923 16:29:59< vultraz> i just logged in so I should be
20160923 16:30:26< shadowm> You are admin now.
20160923 16:30:27< vultraz> tad_: so why did you tackle DM already which is after DiD?
20160923 16:31:06< vultraz> shadowm: thanks
20160923 16:32:10< shadowm> loonycyborg: Could you close the bug report? It doesn't apply to any releases.
20160923 16:32:25< tad_> vultraz: Because when I started I had questions which HttT and DM had the answers to. But I'm not doing a UMC mainline-themed campaign because the existing mainline needs me ...
20160923 16:32:38< loonycyborg> shadowm: I want you to watch the fixed server a bit first
20160923 16:33:04< shadowm> Meh.
20160923 16:33:13< shadowm> Well, just finished building it.
20160923 16:33:47< shadowm> loonycyborg: Okay, deployed.
20160923 16:33:54< loonycyborg> cool
20160923 16:34:29< shadowm> shikadibot: commit HEAD
20160923 16:34:33< shikadibot> shadowm: Revision 7a915f90357e (Charles Dang) on Fri Sep 23 15:54:30 2016:
20160923 16:34:36< shikadibot> shadowm: Slight readme update
20160923 16:34:38< shikadibot> shadowm: Web interface URL: https://github.com/wesnoth/wesnoth/commit/7a915f90357e
20160923 16:34:57< shadowm> Don't mind me, just making sure I built the right commit.
20160923 16:36:45-!- noy_ [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 16:39:19-!- noy [~Noy@wesnoth/developer/noy] has quit [Ping timeout: 255 seconds]
20160923 16:39:19-!- noy_ is now known as noy
20160923 16:41:10-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has joined #wesnoth-dev
20160923 16:44:37-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 16:44:53-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 16:51:05-!- Bonobo [~Bonobo@2001:44b8:254:3200:4179:8a2c:fc46:a1b4] has joined #wesnoth-dev
20160923 16:51:49-!- Bonobo [~Bonobo@2001:44b8:254:3200:4179:8a2c:fc46:a1b4] has quit [Client Quit]
20160923 16:59:35-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 17:05:23< zookeeper> tad_, well that's why i've suggested that someone should redesign it
20160923 17:06:14-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20160923 17:06:16< tad_> zookeeper: vultraz just volunteered to do that ... 
20160923 17:06:48< aeth> tad_: I have experimented with a lot of different cave designs on my own map, since it's about 1/3 cave and it's a 400x200 RPG map
20160923 17:07:31< aeth> What I've found is to use as little of the cave terrain as possible, mostly using medium/dark dirt, cave path, ford, gray water, etc. This also naturally leads to wider caves since the edges against the walls are usually cave (especially rockbound)
20160923 17:07:43< zookeeper> i have no objections to there being massive dungeon crawls in general, _but_ it's not good when one appears all of a sudden as the final scenario.
20160923 17:09:13< aeth> I've tried almost every combination of cave design possible, including some that are quite unexpected, in an attempt to fill my map. Except, of course, lots of cave and rockbound cave grouped together with no flat. I did this in a really old map that was meant for dwarves vs trolls but it doesn't really work with a general unit set.
20160923 17:09:45< zookeeper> such a scenario requires much more commitment and patience than what the player has adjusted to up to that point, so unless they really enjoy that kind of scenarios, it might be a rather unpleasant surprise
20160923 17:10:13< zookeeper> roughly equivalent to a scenario that's suddenly a lot harder than the ones before it
20160923 17:10:15< aeth> scenario 22 in DM?
20160923 17:10:20< aeth> it's... all flat
20160923 17:10:30-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 17:10:31< tad_> For THoT 'rather unpleasant' is a huge understatement.
20160923 17:11:38< aeth> THoT scenario 12, hmm
20160923 17:11:44< zookeeper> tad_, if you decide to take DiD next then it'd be great if you'd be receptive to doing some extra work for the final scenario :J
20160923 17:12:29< aeth> oh wow, it looks like I have designed better undergound and indoor areas than mainline campaigns
20160923 17:13:08-!- Kwandulin [~Miranda@p200300760F2C7109DD268D91F1142483.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160923 17:14:17< tad_> aeth: get with zookeeper if you want to take a run at fixing THoT S12
20160923 17:14:18< vultraz> aeth: No Shit, Sherlock :P
20160923 17:14:33< vultraz> aeth: mainline maps are very basic and need updates 
20160923 17:14:39< vultraz> especially to use newer terrains
20160923 17:14:48< vultraz> a lot of them were drawn when stuff in wesnoth was simpler overall
20160923 17:14:54< vultraz> and the quality was lower
20160923 17:15:19< vultraz> compare the maps of the first 9 NR scenarios between 1.12 (old) and 1.13 (my redesigns)
20160923 17:15:42< aeth> vultraz: to be fair, most of my maps were started a long time ago too, I just keep them updated with new terrains :-p
20160923 17:16:02< vultraz> see, a problem wesnoth has is that the quality of mainline is half great, half sub-par
20160923 17:16:16< aeth> I think 1.8 or so was around where mainline *could* look good, though. Mainline has a ton of flat, and you can only make that look good with a ton of flat.
20160923 17:16:39< vultraz> It has some amazingly well done campaigns like UtBS and TRoW
20160923 17:16:42< aeth> At some point 3 grass types were introduced, and unfortunately in the bit of mainline I've played since, the grass was only paintcan replaced most of the time, making not much of a difference
20160923 17:16:48< vultraz> (TRoW is, IMO, the highest quality mainline campaign)
20160923 17:16:48< aeth> Also, there are three dirts now
20160923 17:17:01< vultraz> but then you have stuff like HttT
20160923 17:17:03< vultraz> which are just..
20160923 17:17:04< zookeeper> i'm inclined to agree
20160923 17:17:06< vultraz> kinda... dull :|
20160923 17:17:11< aeth> You can mix three dirts and three grasses and three cobblestones to get some variety in the flat (although lots of flat is kinda a bad design anyway)
20160923 17:17:32< aeth> For large forests (e.g. the first scenario in HttT) there are tons of different forest types you can mix
20160923 17:17:43< vultraz> yeah
20160923 17:17:44< aeth> Doesn't make much of a biological sense but they're elves, they cultivate it or something
20160923 17:18:14< tad_> zookeeper: Yes, DiD is where I'm headed next
20160923 17:18:14< vultraz> but really, take a look at my changes to NR's maps in 1.13
20160923 17:20:05< aeth> vultraz: have you seen my caves in my map?
20160923 17:20:12< vultraz> no
20160923 17:20:17< vultraz> I've seen nothing of yours
20160923 17:20:44< aeth> it's (mostly) 1.12 because it's MP... the map itself has to be 1.12 for now.
20160923 17:20:55< zookeeper> tad_, basically the final scenario should have a bit more variety, like heroes coming in from different directions, and maybe the map getting slowly more decrepit, and especially for the final dialogue to actually differ based on the hero (doesn't really fit orcs at all for instance).
20160923 17:21:14< aeth> vultraz: https://github.com/Aethaeryn/wesnoth-umc/tree/master/Aethaeryns_Mod
20160923 17:21:21< vultraz> zookeeper: i also kinda dislike the fact that you have to lose to win :|
20160923 17:21:42< tad_> zookeeper: I will keep that in mind.
20160923 17:21:50< zookeeper> vultraz, well that part is great
20160923 17:22:03< aeth> vultraz: that's almost like those achievements in games where one is like "die, lol, isn't our game hard" and the real achievement is never getting the achievement
20160923 17:22:17< aeth> games should never reward losing/death/etc. imo
20160923 17:23:03< tad_> OK. On my THoT PR: I have a couple minor fixes from Celtic Minstrel's comments, and I'm dropping further work on S12 in favor of Aeth or Vultraz doing a rework.
20160923 17:23:13< vultraz> tad_: ok
20160923 17:23:29< zookeeper> arguably the scenario could be just a never-ending survival where you just have to last as long as possible, but that'd require quite a bit more planning.
20160923 17:24:14< aeth> vultraz: https://raw.githubusercontent.com/Aethaeryn/wesnoth-umc/master/Aethaeryns_Mod/maps/big-woods.map
20160923 17:24:20< tad_> aeth: 'never' I disagree with .. but it should be obviously what's needed and obviously to the playter's advantage
20160923 17:24:28< aeth> I basically merged all of my RPG maps into a woods map
20160923 17:24:29< vultraz> holy jesus fuck P_P
20160923 17:24:45< vultraz> this I have to see
20160923 17:24:49< aeth> there were 4 or so cave areas before I merged it
20160923 17:24:53< aeth> and several outdoor ones
20160923 17:25:13< aeth> Most of the work is  the cave and indoor areas because they're quite hard to do in the Wesnoth engine
20160923 17:25:20< aeth> I've done little pieces at a time for the past few years
20160923 17:25:31< aeth> I've almost filled the cave, finally.
20160923 17:25:43< aeth> Then I'll detail the less detailed cave areas.
20160923 17:25:56< vultraz> holy mother of map
20160923 17:25:58-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160923 17:26:18< aeth> The indoor parts are by far the hardest to do with the engine, followed by cave. Rivers are surprisingly nice because of the hexagons and transitions
20160923 17:27:04< vultraz> tbh, though
20160923 17:27:06< aeth> I probably have more than 10 different cave/indoor styles and I can then copy the rules that apply to each area into mainline SP campaign maps. (Obviously MP generally needs to be balanced and thus boring)
20160923 17:27:10< vultraz> I don't exactly like all of it :P
20160923 17:27:19< vultraz> then again
20160923 17:27:27< vultraz> I prefer maps that look like a landscape
20160923 17:27:28< aeth> vultraz: if there are any plain parts, those parts aren't done yet and will be returned to
20160923 17:27:46< aeth> the outdoors definitely look like a landscape, zoom out :-p
20160923 17:27:47< vultraz> also, that's not the cave design i favor
20160923 17:28:00< vultraz> but my overall reaction is *wow*
20160923 17:28:03< aeth> which one?
20160923 17:28:07-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160923 17:28:11-!- JyrkiVesterinen [~JyrkiVest@87-92-62-189.bb.dnainternet.fi] has joined #wesnoth-dev
20160923 17:28:27< vultraz> any of them
20160923 17:28:30< aeth> I've tried just about every cave design somewhere in the map
20160923 17:28:52< tad_> aeth: Well, for THoT I'd suggest a size more appropriate to that of the remainder of the campaign. And the effort should be about the same, too. The problem with THoT S12 is it's way larger and vastly harder so it promoted quitting instead of promoting finishing while having fun.
20160923 17:28:55< aeth> I even have a mostly standard cave around (268, 33)
20160923 17:29:24< vultraz> aeth: https://raw.githubusercontent.com/Vultraz/Shadows_of_Deception/master/episode1/maps/06_Awakening_part_1.map
20160923 17:29:25< aeth> tad_: well, yes, I'm not proposing expanding it into a 400x200 map
20160923 17:29:55< vultraz> tad_: working on a quick draft of the new thot map as we speak
20160923 17:30:01-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 17:30:42< aeth> illegal map tile
20160923 17:30:50< vultraz> ah
20160923 17:30:56< vultraz> you'll need to clone my repo, the
20160923 17:30:58< vultraz> n
20160923 17:31:10< tad_> Well, I'm bowing out of THoT then. I'll finish up my PR and someone else can update S12 ...
20160923 17:31:12< vultraz> (it's for 1.12, btw)
20160923 17:31:52-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 17:32:18< vultraz> honestly, shadow m makes the absolutely best cave and dungeon maps ever
20160923 17:32:26< vultraz> absolute*
20160923 17:32:31< aeth> vultraz: yeah I don't like your style
20160923 17:32:34< shadowm> Who is shadow space m?
20160923 17:32:41< vultraz> aeth: heh :P
20160923 17:32:43< aeth> vultraz: too flat
20160923 17:32:47< vultraz> xD
20160923 17:32:55< vultraz> I say *your* style is too flat
20160923 17:33:12< vultraz> shadowm: your evil twin
20160923 17:33:14< aeth> vultraz: every indoor area of the cave that is mostly flat is merely unfinished
20160923 17:33:20< aeth> in my map, at least
20160923 17:33:32< aeth> except perhaps the starting area in the bottom right corner, where no combat is expected
20160923 17:33:39< vultraz> well, different tastes, I guess
20160923 17:33:54< aeth> I also don't like the cobble path on the dark flagstone
20160923 17:34:00< aeth> perhaps it's because of the transition between the two
20160923 17:34:09< aeth> a custom terrain with a better transition might work better than the dark flagstone
20160923 17:34:19< aeth> tbh, though, idk why the two terrains have such a bad transition
20160923 17:34:28< aeth> I would have mixed them in more places on my map if they had a better one
20160923 17:35:26< vultraz> ask doofus to draw us a better one
20160923 17:35:32< aeth> vultraz: why isn't gate a mainline terrain? (is it in 1.13?)
20160923 17:35:37< vultraz> 'it is
20160923 17:35:40< vultraz> in 1.13
20160923 17:35:43< aeth> ah
20160923 17:35:48< vultraz> along with brazier
20160923 17:35:51< vultraz> and campfire
20160923 17:35:59< aeth> when it's not a terrain it doesn't transition well with the indoor wall
20160923 17:36:18< aeth> I am aware the brazier is mainlined. All of the campfires will be regexp replaced with braziers when I drop support for 1.12
20160923 17:36:18< vultraz> I use custom terraingraphics wml
20160923 17:36:24< aeth> Perhaps that affects your opinion of my map a bit
20160923 17:37:02-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.2 Aria]
20160923 17:37:13< aeth> I don't have the gate. I guess I need to update my git for 1.13
20160923 17:37:29< vultraz> hm
20160923 17:37:31< vultraz> wait
20160923 17:37:37< vultraz> perhaps i did not mainline the gate
20160923 17:37:44< vultraz> hm
20160923 17:37:49< vultraz> actually, on second thought
20160923 17:37:54< vultraz> i did not mainline the gate 
20160923 17:38:02< vultraz> remind me to do this sometime
20160923 17:38:19< aeth> I wouldn't mind the gate just as an image, *but* it doesn't transition properly just as an image. You either see the whole gate placed on top of the wall as if there's no overlap/transition or you place the gate, but it gets cut off at the hex boundary on the right side
20160923 17:38:31< aeth> the gate as a terrain in your map actually looks properly done
20160923 17:38:50< aeth> Although, to be fair, there are quite a few transitions that are bad with indoor stuff.
20160923 17:39:11< vultraz> thank shadowm 
20160923 17:39:14< vultraz> he did the wml
20160923 17:39:40< vultraz> as I said, he does the absolute best cave and dungeon maps
20160923 17:40:14< aeth> Naturally, I'm going to be biased toward my own.
20160923 17:40:25< tad_> Of course.
20160923 17:40:27< aeth> Too bad I expanded the map to be huge, because some parts of it are quite finished and some aren't
20160923 17:41:05< aeth> I had to double MP on my map, too. I didn't want to and it's perfectly playable without it but... some people (not naming names) take forever to end turn in MP RPGs
20160923 17:41:31< aeth> so no one would ever leave the first area without double moves. in SP campaigns, moves isn't an issue.
20160923 17:41:48< aeth> also apparently "move points" clash with "multiplayer" in acronyms
20160923 17:42:12< aeth> some people is actually kind of half of the players. I'd have to have a timer on or something.
20160923 17:42:41< tad_> aeth: Or reduce the amount of thinking required ...
20160923 17:43:31< aeth> the only thinking is really that my map is open-ended so there's a lot of possible ways to play it
20160923 17:43:53< aeth> which actually is kind of a problem, now that you bring it up... the moving parts that *should* be fast are now probably the only part that people need to think hard about
20160923 17:44:30< tad_> Personally, there are times I love a huge map. I found it improved my gameplay some to do a few max-size random islands and use a new map instead of reloading
20160923 17:45:03< aeth> I hate Wesnoth's map generator, though. Probably the weakest part of Wesnoth's engine.
20160923 17:45:16< aeth> I did play a few games just because but I wish there was something considerably better.
20160923 17:45:58< vultraz> patches welcome :)
20160923 17:46:31< tad_> aeth: A fractal based generator would be good but it'd be a lot of work. 'Random' and 'Realistic' are not synonyms.
20160923 17:46:43< aeth> too busy working on height map generation, which is completely useless to Wesnoth's engine unless I e.g. say mountains = 4, hills = 3, everything else = 2, water = 1 or something
20160923 17:46:52< aeth> and 4's a lot less than 256 :-p
20160923 17:47:21-!- prkc_ [~prkc@46.166.188.236] has quit [Ping timeout: 250 seconds]
20160923 17:47:22< aeth> the problem is, good looking proc gen (and arguably good looking maps in general) are hostile to Wesnoth's concept of default MP balance.
20160923 17:47:37< aeth> I'm amazed that Wesnoth has so many maps that look decent in MP given how hard it is to make a balanced map
20160923 17:47:59< aeth> And somehow those maps tend to be better than SP campaign maps even though those are easier to make (well, yeah, I know, it's because they were all made 6+ years ago)
20160923 17:48:05< celticminstrel> aeth: You could write a map generator in Lua. :P
20160923 17:48:32< aeth> celticminstrel: the only thing that really lends itself to proc gen in Wesnoth is probably caves
20160923 17:49:00< aeth> dungeons would also work too if there were more to true indoors than rubble, dwarven castles, and flat
20160923 17:49:12< aeth> (that's why almost all of my indoors areas are severely ruined and mixed with cave)
20160923 17:50:09< aeth> you could probably proc gen the Wesnoth way of microbalancing MP maps, but that would have to be done by hand rather than use existing algorithms, and the maps would probably be ugly since a lot of tricks have to be used to avoid this in MP maps
20160923 17:50:30< aeth> the current map gen only really works with players vs AI
20160923 17:51:14< celticminstrel> You can totally play random maps against players.
20160923 17:55:25< aeth> but the only *balanced* maps are the non-random ones
20160923 17:55:37< aeth> most strategy games have PvP random maps
20160923 17:56:13< vultraz> celticminstrel: just confirmed unit placement works in master :/
20160923 17:56:14< vultraz> in the editor
20160923 17:56:33< vultraz> oh wait
20160923 17:56:35< vultraz> now it breaks
20160923 17:56:36< celticminstrel> The balanced maps tend to be a bizarre patchwork. :/
20160923 17:56:38< aeth> actually, it's gotten better. it's still not balanced but it's not horrible
20160923 17:56:39< vultraz> but it initally work
20160923 17:57:03< gfgtdf> i dont reall think the defalt map generator code has changed recently
20160923 17:57:11< vultraz> abilities.cpp:312
20160923 17:57:15< gfgtdf> but i wonder what is/was bad about it?
20160923 17:57:26< celticminstrel> gfgtdf: It has changed a lot, but the logic of the generation shouldn't have changed much.
20160923 17:57:44< celticminstrel> vultraz: Yeah, wants the game_board.
20160923 17:58:18< gfgtdf> celticminstrel: well yes i did multipel refactor to is mainyl to make random seeds work in 1.13. But i mean the result map not the internal code.
20160923 17:58:39< celticminstrel> You weren't the only one who did refactors. :P
20160923 17:59:43< tad_> Personal opinion: if 'balanced maps' look like artificial patchworks, the problem isn't the map it's the units and eras
20160923 18:00:38< tad_> I'd suggest the basic problem is "I want this era/unit to use tactical thermonuclear weapons .. deal with it."
20160923 18:00:57< aeth> tad_: actually, it's a problem with the terrain. You *can* hide balance behind terrain diversity, e.g. rubble is technically hills. There's just not enough of it
20160923 18:01:44< aeth> The more terrain there is that has the exact same stats as some other terrain that looks very different, the more you can hide patchwork balance.
20160923 18:02:02< aeth> e.g. muddy quagmire goes well with different areas than swamp does even though they're both swamp
20160923 18:02:55< aeth> One problem is that there's not enough of this for e.g. indoor dungeons. There's literally just the pile of rubble that are really hills and the dwarvish castle (where it even transistions properly)
20160923 18:03:25< aeth> There would also need to be e.g. villages, forests, etc. Probably not mountains, though.
20160923 18:03:38< tad_> Another way to put it is the movement effects are too 'course' and the scale of the terrain too 'variable'.
20160923 18:04:30< aeth> Wesnoth's abstract terrain scale is basically the main feature of Wesnoth map design. Yes, it has issues when you want to push the engine and do something a bit more concrete like an RPG or something
20160923 18:04:36< tad_> It's hard to balance when you have a 'small cave of 10 tiles' which is one 'room' and an entire village is 1 tile. cf HttT Siege of Elensefar
20160923 18:05:08< aeth> tad_: generally, "villages" are buildings on such maps, which is still a variable scale but not as bad
20160923 18:05:40< aeth> although, yes, one map can have a city where a village is a house, a wilderness where a village is a village, and a cave or building where it has a room larger than one hex
20160923 18:06:09< aeth> (although, I think in general even in cities, the "village" represents several buildings so the scale isn't quite that bad)
20160923 18:08:05< tad_> aeth: So I've changed my THoT PR to say you or vultraz will fix S12 .. please note the PR uses 1.13.5+dev features and I have a couple changes for S12 in the PR.
20160923 18:08:11< celticminstrel> I'm pretty sure there are cave villages.
20160923 18:08:18< celticminstrel> No cave forests though.
20160923 18:08:28< celticminstrel> Or frozen cave.
20160923 18:08:39-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has quit [Quit: ChipmunkV]
20160923 18:08:49< aeth> celticminstrel: there are two types of cave villages, which really isn't enough because it's cave monster vs. dwarves or something
20160923 18:09:04< celticminstrel> Two types? Not one?
20160923 18:09:05< aeth> that leaves out e.g. orcs/goblins and undead (well undead that aren't e.g. zombie dwarves)
20160923 18:09:19< aeth> celticminstrel: dwarvish village is a cave village
20160923 18:09:26< aeth> outside, I think the hill village makes more sense for dwarves
20160923 18:09:34< tad_> celticminstrel: Well, what we need is a graphic arts department as active as the engine coders. 'not happening' is a good estimate on that, though
20160923 18:09:37< aeth> (even though the hill village is generic)
20160923 18:09:41-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has joined #wesnoth-dev
20160923 18:10:07< aeth> dwarvish village is, in turn, the only one that goes with the dark flagstone + dungeon wall and so the only one that can work "indoors" (assuming these are very large scale buildings)
20160923 18:10:12< celticminstrel> Well, we do have doofus-01 working on terrains.
20160923 18:10:33< aeth> so there are two villages that can work for caves, and one of those can also work for indoor dungeons
20160923 18:11:06< aeth> I have gotten some other villages in caves to some effect, e.g. swamp village, ruined city village, drake village (by lava obviously), etc. But they're nowhere near as fitting
20160923 18:11:32< aeth> orc village sort of works, too
20160923 18:11:42< aeth> (probably more than those other ones I listed just now)
20160923 18:11:53< tad_> Seems to me those working on graphics are more on the forums and those working on code are more on here and so communication between them is an issue
20160923 18:13:42< aeth> What would be interesting is a village that transitions *into* the indoor wall, as part of the wall. Also a castle that does the same (and/or fix some of the broken transitions of the dwarvish castle... but there needs to be another castle type that works there anyway)
20160923 18:13:46< celticminstrel> Forums are probably a bit more convenient for art.
20160923 18:14:14< aeth> Forums are definitely better for art. On IRC all you have are imgur links and imgur does ruin the quality of higher quality images.
20160923 18:14:23< tad_> Not saying they're wrong on either side .. just that the tools used are different and that causes some issues.
20160923 18:14:34< aeth> In fact, imgur is becoming increasingly unsuitable for IRC. I hope a popular replacement shows up.
20160923 18:14:43< celticminstrel> No, imgur is not all you have. There many other image hosting sites.
20160923 18:15:09< aeth> celticminstrel: every image hosting site before imgur was universally horrible for image sharing with the exception of that one that was for photos (flickr?)
20160923 18:15:18< aeth> e.g. photobucket, which is surprisingly even worse now
20160923 18:15:22 * tad_ wonders about merging IRC with Torent filesharing for images.
20160923 18:15:33< zookeeper> well there's not that much coder<->artist communication that seems to be needed
20160923 18:15:41< celticminstrel> photobucket's not great, no. There was also imageshack which I haven't seen linked recently, and there's tinypic too.
20160923 18:15:42< aeth> tad_: IRC does have its own file sharing, it's just that no one uses it, at least on Freenode
20160923 18:15:45< aeth> DCC?
20160923 18:16:21< aeth> celticminstrel: photobucket and imageshack were both horrible. tinypic was/is okay afaik
20160923 18:16:24< tad_> GAK. Of course nobody uses it. Even if it's been fixed in the past 20 years the bad rep it got was earned
20160923 18:16:41< aeth> celticminstrel: what I like about imgur is that I don't need to register an account, which makes it great for throwaway images
20160923 18:17:01< aeth> I have long since forgotten my username/password for photobucket/imageshack
20160923 18:17:21< tad_> btw, the imgur discussion you're having mirrors the talk on reddit about it
20160923 18:17:21< aeth> and they were terrible to share anyway. the image was/is zoomed out to like 1/20th of an add-ridden screen
20160923 18:19:48< tad_> zookeeper: hand-waving about DM PR .. and THoT has only a couple changes to do for celmin's comments .. I have to do some system work here, quick fix THoT PR and I'm off to DiD ...
20160923 18:19:58-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160923 18:23:17< EliDupree> Great, my lua AI code crashed wesnoth
20160923 18:25:30-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 260 seconds]
20160923 18:33:19-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160923 18:33:51< celticminstrel> Note, DCC is not specifically filesharing.
20160923 18:34:11< celticminstrel> (Actually, DCC is used a lot to create IRC file-sending bots, I believe.)
20160923 18:34:42< celticminstrel> EliDupree: Do go one.
20160923 18:34:51< EliDupree> ?
20160923 18:35:06< celticminstrel> What did you do? How did it crash?
20160923 18:35:53< mattsc> celticminstrel: I’m only here for a few minutes, so quickly: the weapon number is changed by +/-1 by the API as needed when doing attacks, or when calling simulate_combat etc.
20160923 18:36:19< EliDupree> It's not in my output history anymore, but it was something like "PANIC: unchecked error in lua API call (attempt to index a function value)"
20160923 18:36:19< mattsc> But apparently when the weapon information table is returned, it gives you the values as they are in the C++ config
20160923 18:36:33-!- vultraz [~chatzilla@124.109.10.167] has joined #wesnoth-dev
20160923 18:36:33-!- vultraz [~chatzilla@124.109.10.167] has quit [Changing host]
20160923 18:36:33-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160923 18:36:37< celticminstrel> Weapon information table?
20160923 18:36:52< EliDupree> https://github.com/elidupree/wesnoth-ai
20160923 18:37:09< mattsc> https://wiki.wesnoth.org/LuaWML/Units#wesnoth.simulate_combat
20160923 18:37:12< EliDupree> commit 99ee03cbc38b52b4df084691f7ed258ec37db2c9
20160923 18:37:26< mattsc> Return values #3 and 4.
20160923 18:37:37< EliDupree> ./wesnoth -d --debug-lua -m --multiplayer-repeat 1000 --scenario Tiny_Close_Relation_LAI --controller 1:ai --controller 2:ai
20160923 18:37:48< EliDupree> That should be enough information to reproduce it
20160923 18:39:20< celticminstrel> So the bug is in wesnoth.simulate_combat?
20160923 18:39:31< celticminstrel> Ack, was scrolled up somehow.
20160923 18:39:58-!- Kwandulin [~Miranda@p200300760F2C7109DD268D91F1142483.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160923 18:40:27 * celticminstrel saves EliDupree's stuff for later perusal.
20160923 18:42:25-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 260 seconds]
20160923 18:42:27< EliDupree> celticminstrel: Neat. Back in the day, whenever I reported a wesnoth bug, the old wesnoth devs would say "why would you do the thing that triggered the bug?" instead of trying to fix it
20160923 18:42:55< celticminstrel> gfgtdf seems to have a tendency to ask that sort of question.
20160923 18:43:15< celticminstrel> No idea if he's the only one.
20160923 18:43:27< mattsc> EliDupree: I’ll have to check out what you’re doing with the AI there later.  Looks interesting.
20160923 18:44:07< celticminstrel> So basically, attack_num is incorrect in the third and fourth return values.
20160923 18:44:17< EliDupree> mattsc: I'm in the middle of trying out using recurrent neural networks to make an AI in lua. I have NO idea whether it will be any good or not
20160923 18:44:36< celticminstrel> Do you happen to know to what extent MicroAIs or addons correct for it?
20160923 18:44:44-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160923 18:45:07< mattsc> EliDupree: well, it’s interesting to me whether it will turn out to be good or not.  :)
20160923 18:45:11< mattsc> celticminstrel: no
20160923 18:45:12< EliDupree> When I get it to actually play the game, rather than – for instance – looping so hard that I have to kill -9 wesnoth, I'll say more :-)
20160923 18:45:22< celticminstrel> The most compatible fix would be replacing attack_num with a different key holding the correct value (then attack_num would be deprecated and removed after the 1.14 release).
20160923 18:45:43< mattsc> celticminstrel: sure
20160923 18:45:46< EliDupree> (How's that for another feature request: do not completely freeze while lua is looping)
20160923 18:46:01< mattsc> I’m off again though right now.  TTYL.
20160923 18:46:23-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20160923 18:47:02< celticminstrel> EliDupree: I think that requires editing the Lua base source, though it's possible the config file might provide a way to call user code at defined intervals.
20160923 18:47:37< EliDupree> What about running the lua in a separate thread?
20160923 18:47:44< celticminstrel> To make Wesnoth not freeze, you would have to call controller_base::play_slice periodically, I believe.
20160923 18:47:48< EliDupree> I mean, besides all the difficulties of multithreading in C++
20160923 18:48:11< celticminstrel> Yeah, multithreading is not a panacea. :P
20160923 18:48:27-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 250 seconds]
20160923 18:48:27-!- wedge010 is now known as wedge009
20160923 18:48:48< EliDupree> Really, the main thing I want is to be able to kill wesnoth with the window close button or ctrl-C
20160923 18:49:41< shadowm> For the record, I've always said that any time Lua or WML can cause Wesnoth to misbehave or crash that's a bug that we, rather than the content creator, ought to fix.
20160923 18:49:59< celticminstrel> Absolutely.
20160923 18:50:26< EliDupree> :-)
20160923 18:54:50-!- JyrkiVesterinen [~JyrkiVest@87-92-62-189.bb.dnainternet.fi] has quit [Quit: .]
20160923 18:55:17-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160923 19:01:49< EliDupree> Weird: I'm getting E_NOT_REACHED_DESTINATION errors, even though I only do moves that I got directly from wesnoth.find_reach (unit)
20160923 19:04:15< tad_> I get those sometimes, too. And a few 'results not checked' as well. I've never tracked them down and assume they're some AI feature being overly pedantic.
20160923 19:05:01< tad_> My favorite is move from (12,13) to (12,14) not reached in a single move when I check and the unit most certainly did move from 12,13 to 12,14
20160923 19:05:17< EliDupree> !!!
20160923 19:05:48< EliDupree> This is a big problem, because it's causing my freezes – the AI keeps trying to do the same move that wesnoth is rejecting for some reason
20160923 19:09:15< EliDupree> Interesting… My code uses AI.move (x0,y0,x1,y1), but the wiki only says that you can do AI.move (unit,x1,y1)
20160923 19:10:44-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160923 19:12:07< tad_> Well, the wiki is sorta laggy, and that's when it's correct.
20160923 19:12:22< EliDupree> yeah...
20160923 19:13:27 * tad_ chuckles. I just removed this variable from another campaign, while fixing this exact same code. Now I know where it came from. Deja vu all over again with the fix .. *sigh*
20160923 19:14:29< EliDupree> Also, I can't figure out how to keep wesnoth from falling back to the default AI after my own AI stops making its moves each turn
20160923 19:14:47< EliDupree> I can use stopunit_all for the units, but there isn't a way like that to stop recruiting
20160923 19:15:25< tad_> set gold to zero?
20160923 19:16:42< EliDupree> haha
20160923 19:17:05< celticminstrel> I think the move function actually takes two locations rather than a unit and a target location.
20160923 19:17:19< celticminstrel> But a unit also happens to satisfy the requirements of a location.
20160923 19:17:25< EliDupree> heh
20160923 19:17:52< celticminstrel> The formula version only accepts a location pair, I think.
20160923 19:22:29 * tad_ is searching Google for a good blog post on the difference between 'clean code' and 'well, it works, and doesn't cause errors' ...
20160923 19:22:40< celticminstrel> Why?
20160923 19:22:47< EliDupree> I think find_reach might have been returning locations with units in them
20160923 19:23:11< celticminstrel> No idea if that's expected behaviour.
20160923 19:23:13< EliDupree> I also made a hard limit of 100 move attempts for now, to prevent future loops
20160923 19:23:42< EliDupree> I think it is, which is understandable, just inconvenient for this use case
20160923 19:24:04< tad_> Just looking at a scene in DiD and trying to decide whether it needs the same fixes because it's the source of the problems in THoT I fixed or if it's working well enough, here, I can ignore it and move on.
20160923 19:25:13< tad_> At least, now, I understand why Shaun and Ed were mentioned in THoT ...
20160923 19:27:21< EliDupree> LOL, my "AI" occasionally wins a game even though its behavior is still completely random (because the default AI moves its leader into danger)
20160923 19:28:14< zookeeper> tad_, huh, curiously yes "done_shaun" is a non-existent variable that the code tries to clear, in THoT S05
20160923 19:28:29< zookeeper> i guess someone forgot to clean that up when copypasting from DiD :J
20160923 19:28:46< tad_> zookeeper: One of the minor fixes in the PR.  But now I know where it came from ...
20160923 19:29:42< zookeeper> tad_, yeah. just don't "fix" it in DiD :p
20160923 19:30:19< tad_> zookeeper: Fix it? I'm thinking about adding a Walking Dead easter egg is I see the opportunity ...
20160923 19:31:11-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 19:31:58< zookeeper> eh what the heck is this... id=guard_$x1|_$y1|_$guard_location.n
20160923 19:32:17< zookeeper> no one's documented [random_placement] so i can't even check that for what that .n might be
20160923 19:32:50< zookeeper> according to the rest, it's supposed to evaluate to 1 or 2
20160923 19:33:38< tad_> zookeeper: I'm reading that. And noticing it's a random 0-to-2 but a check for 2 and wondering what's up with that ...
20160923 19:33:49< zookeeper> i guess per each placement, the location is stored in the given variable=, and they'll have the extra member variable n to denote the number of the location
20160923 19:33:57< tad_> zookeeper: So I have some edge conditions to test ...
20160923 19:34:18< zookeeper> tad_, well the easter egg can only trigger when there's two guards, so it checks for when that condition is true
20160923 19:34:39< zookeeper> shouldn't be any problems with that code, it's not that old
20160923 19:35:18-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 19:35:18< zookeeper> hmh. i guess that was a weird thing for me to say, since it's usually new code that has problems, not old code :p
20160923 19:35:37 * tad_ smiles.
20160923 19:36:23< tad_> It needs testing, for sure. There's a smell there and I'm not sure if it's good or bad. So I'm wagging my tail and pointing and am gonna sniff around some more.
20160923 19:37:57-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 19:38:37< zookeeper> what part gives off a smell?
20160923 19:39:34< zookeeper> the only thing that's a problem is that the number of guards is truly random (0..2) every time, so if the player gets very lucky or unlucky, the difficulty can vary massively
20160923 19:40:16< zookeeper> and that could definitely be fixed
20160923 19:40:31< tad_> Well, first check: what if random placement can't find a place .. do we ignore it, error message, find another place?
20160923 19:41:21< tad_> And the way it was used in THoT all those village and unit variables were not really needed and could be eliminated.
20160923 19:42:27< zookeeper> well, presumably allow_less=yes takes care of that
20160923 19:42:30< tad_> In ThoT when I was stress-testing logic like this I ended up with a crash due to an error in Monte Carlo ..
20160923 19:43:37< tad_> So, it's worth a day kicking the tires and slamming the gear-shift back and forth ...
20160923 19:48:25< gfgtdf> zookeeper: [random_placement] gives each generated item a numbewr stored in $varname|.n
20160923 19:48:52< gfgtdf> zookeeper: and [random_placement] is dodumented here https://wiki.wesnoth.org/InternalActionsWML#.5Brandom_placement.5D. although that particular part is not mentioned
20160923 19:49:22< zookeeper> gfgtdf, ah, so it's only missing from the sidebar
20160923 19:50:08< zookeeper> dunno why i didn't think of that possibility
20160923 19:51:15< tad_> gfgtdf: ISTM random_placement is in Lua and similar Lua in THoT had crashes at the edges. Last time I looked at it was HttT Crossroads and the fix there was what to do about a unit randomly appearing atop another. But I didn't really stress-test it there.
20160923 19:52:22< gfgtdf> tad_: [random_placement] has a min_distance=parmeter that also controls whether the same location cna be chosen twice.
20160923 19:53:10< tad_> gfgtdf: This issue in HttT was two executions overlapped the placement areas and randomly caused a problem because of that. So the fix was to remove the overlap
20160923 19:53:40< gfgtdf> tad_: if you want to exclude locations that are already occupied by aunit you must pass [not][filter][/filter][not] to the location_filter
20160923 19:54:29< gfgtdf> tad_: note that [random_placement]isn't oynl used to place units, so it does make sense to use it also onl locations that are already occupied by a unit.
20160923 19:54:35< tad_> gfgtdf: The issue in THoT was when there were NO available tiles at all. But that was not [random_placement] but a local function sort of like it.
20160923 19:56:09< gfgtdf> tad_: [random_placement] ha an option allow_less= that controls what to do if not enough locations were found, currently the only options are 'print error'(default), or 'do nothing'
20160923 19:56:21-!- Appleman1234 [~Appleman1@KD119104042061.au-net.ne.jp] has quit [Read error: Connection reset by peer]
20160923 19:57:56< tad_> gfgtdf: We'll see when I get to it. Right now I'm just thinking about it in terms of the problems THoT had with its local function. Not sure if it's worth the effort here to really stress-test [random_placement] because .. as you pointed out .. it has more uses than it's put to here.
20160923 20:01:32< EliDupree> 20160923 16:01:19 error ai/engine/lua: error while initializing ai:  [string "--ai_rnn.cfg..."]:88: syntax error near '<'
20160923 20:01:33< EliDupree> 20160923 16:01:19 error ai/engine/cpp: side 2 : UNABLE TO CREATE engine[lua]
20160923 20:01:33< EliDupree> PANIC: unprotected error in call to Lua API (attempt to index a function value)
20160923 20:01:47< EliDupree> Looks like wesnoth crashes whenever there is a syntax error in a lua AI engine
20160923 20:04:02< celticminstrel> That was a crash to desktop?
20160923 20:04:18< tad_> You mean crash-to-desktop or crash-to-titlescreen .. I get crashes to desktop pretty easily where it should crash to tittlescreen but haven't looked at em
20160923 20:04:20< EliDupree> yes
20160923 20:04:44< tad_> Mine show the same error twice, then end with a coredump, too.
20160923 20:05:10-!- Appleman1234 [~Appleman1@KD119104042061.au-net.ne.jp] has joined #wesnoth-dev
20160923 20:07:55-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160923 20:09:24-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 20:11:48< tad_> celticminstrel: I jsut tried to re-create a crash-to-desktop from WML and didn't get the coredump but got the error THREE times as it died. Never seen that before.
20160923 20:12:15< celticminstrel> Which error?
20160923 20:13:47< tad_> Invalid tag. Add '[' to the end of any S01 and start the campaign. To get a coredump just change [/scenario] to [scenario] at the end of the same S01, instead.
20160923 20:15:01< tad_> I'm working on DiD S01 so I'm testing there. But I see it all the time when I typo a tag-pair
20160923 20:16:06< tad_> Since it's almost always my fault and easily fixed, I've never tracked it down.
20160923 20:16:12-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160923 20:16:50< tad_> I'm thinking my crash and EliDupree's crash are similar but not related, though.
20160923 20:17:43-!- irker051 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160923 20:27:06-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160923 20:29:54< EliDupree> LOL, my AI and the default AI has reached a stalemate, so the games have stopped
20160923 20:48:06-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev
20160923 21:11:53-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 250 seconds]
20160923 21:14:25-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160923 21:16:09-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 21:16:53-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 21:18:30-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 264 seconds]
20160923 21:20:47-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev
20160923 21:31:17-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 21:31:55-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 21:39:39< EliDupree> LOL, my AI's win rate is going DOWN over time. That's not how genetic algorithms are supposed to work. :-P
20160923 21:48:02< stikonas> well, maybe it's learning to loose...
20160923 21:53:08< stikonas> my classmate actually once made some error (maybe sign error) and his checkers program was trying to loose :D
20160923 21:53:55-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has quit [Ping timeout: 250 seconds]
20160923 21:53:59< EliDupree> Yeah, that's a possible issue. Although the win rate is not actually changing by that much, so it's probably just that the AI is no good at learning.
20160923 21:54:45< EliDupree> What I could really use is a way to visualize what's going on internally, so that I can tell which parts of it aren't working how I think they should be.
20160923 21:56:17< EliDupree> If it was actually trying to lose, I would expect it to be much better at losing
20160923 21:58:21-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160923 22:00:23< celticminstrel> EliDupree: You could possible call wesnoth.show_gamestate_inspector{} or wesnoth.show_lua_console at strategic points in your code?
20160923 22:00:26< celticminstrel> ^possibly
20160923 22:01:03-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160923 22:02:42< EliDupree> I'll try it out, although I think it won't give quite human- readable data
20160923 22:03:07< EliDupree> I did not know about wesnoth.show_lua_console
20160923 22:03:24< celticminstrel> It might be undocumented.
20160923 22:04:37< EliDupree> 20160923 18:04:13 error scripting/lua: [string "--ai_rnn.cfg..."]:86: attempt to call field 'show_lua_console' (a nil value)
20160923 22:04:50< EliDupree> Is that in 1.12?
20160923 22:04:57< celticminstrel> Oh, you're on 1.12, sorry.
20160923 22:05:27< EliDupree> I guess I could do development in 1.13...
20160923 22:05:41< EliDupree> How good is the lua console?
20160923 22:05:44< celticminstrel> You do have the inspector in 1.12, which could help.
20160923 22:06:15< EliDupree> All my data is in lua though. I could just use old-fashioned debug messages
20160923 22:06:22< celticminstrel> Uhh... the lua console basically lets you enter any number of Lua commands and see their output.
20160923 22:06:28< celticminstrel> Or errors.
20160923 22:07:04< EliDupree> Yeah, that actually isn't very useful to me. I was hoping for something that could display the current values of all local and global variables, for instance
20160923 22:07:40< EliDupree> Obviously you could do them one at a time to the console, but I mean in a convenient UI
20160923 22:08:07< celticminstrel> I don't think it knows about local variables at all. It would be nice to change that, though, somehow.
20160923 22:09:37< tad_> What he's looking for is a full-on Lua debugger module. Would be sorta nice to be able to set a breakpoint and play until it triggers, too, then step through the function.
20160923 22:09:59< tad_> But, at the point the Lua console comes up, the locals for a Lua stack frame are probably gone.
20160923 22:10:44< EliDupree> Point of order, I prefer to be called by gender-neutral pronouns
20160923 22:11:19< celticminstrel> tad_: I'm not sure about that...
20160923 22:11:39< EliDupree> Ha ha I, I just added a simple debug message and it showed that my variables are messed up in an obvious way
20160923 22:12:02< celticminstrel> Given that show_lua_console is itself on the stack.
20160923 22:12:06< tad_> celticminstrel: In 1.12 I suppose you could open the inspector and then open the console. But you'd have to reach up a stack frame to see the locals.
20160923 22:12:13-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160923 22:12:39< celticminstrel> Apart from the fact that the button for it presumably doesn't exist in 1.12. :P
20160923 22:12:54< celticminstrel> But yes, if you're viewing the console from ;inspect in CommandMode, there are no locals to view.
20160923 22:13:02< celticminstrel> But if you invoked it from Lua code, the situation is different.
20160923 22:13:30-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 22:13:33< EliDupree> Pity it doesn't look like the obvious bug was the cause of my problem
20160923 22:13:54< tad_> EliDupree: Is it ever? 
20160923 22:14:19-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20160923 22:20:16< vultraz> I should see if i can make the addon manager functional
20160923 22:22:03-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 22:22:59-!- bumbadadabum [~bumbadada@wesnoth/developer/bumbadadabum] has joined #wesnoth-dev
20160923 22:23:52< tad_> OK. So how dangerous is it to allow undo on capturing a village? I've not seen it done elsewhere.
20160923 22:24:30< tad_> For MP, I'd guess it's a loss-of-sync issue.
20160923 22:28:09< tad_> Ah. Moot point. The allow_undo is wiped out because a later capture event triggers.
20160923 22:29:03< gfgtdf> tad_: there were OOS related to undo on villages on 1.12.4 but they were fixed afaik
20160923 22:30:30< tad_> gfgtdf: Not an issue after all. Either the allow_undo is being ignored or something is wiping it out. I think I'll just remove the call .. seems wrong there anyway.
20160923 22:31:31< gfgtdf> tad_ this happen in mainline?
20160923 22:32:10< tad_> gfgtdf: DiD S01 on a moveto event to 15,11 for a message to entice player into swamp
20160923 22:32:17< tad_> s/01/02/
20160923 22:32:44-!- noy [~Noy@wesnoth/developer/noy] has quit [Read error: Connection reset by peer]
20160923 22:33:44< tad_> Wait a minute .. isn't Undo supposed to be on the Action menu? I can't undo at all ...
20160923 22:34:04< tad_> Doh. Shroud and fog
20160923 22:34:11< gfgtdf> tad_: if there is nothing to undo then it afaik doesnt appear on the menu
20160923 22:34:39< gfgtdf> tad_: revealing fog or shroud always makes a move non-undoable
20160923 22:35:43< tad_> gfgtdf: But it does make it hard to test undo at this point. So .. poof! .. for the test.
20160923 22:36:20< gfgtdf> tad_: it migth be better to test undo at a scneario without fog and shroud.
20160923 22:37:08< tad_> Yep. commenting out the fog and shroud to see what happens at this village. I think removing the allow_undo is the right thing but want to see it in action, first.
20160923 22:38:15< gfgtdf> tad_: whatzs the reason for removing allow_undo?
20160923 22:38:40< gfgtdf> tad_: maybe you should test with dsu active.
20160923 22:39:22-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has quit [Quit: ChipmunkV]
20160923 22:39:29< tad_> capture triggers unit creation .. seems odd and I want to see if I can undo at all.
20160923 22:41:55< gfgtdf> tad_: he onyl event with [allow_undo] in DiD02 is a moveto event that onyl does a message and no unti creation
20160923 22:42:17< gfgtdf> the only*
20160923 22:42:40< tad_> Wrong. 15,11 is a village and there's a capture event following which creates the unit there.
20160923 22:44:01< gfgtdf> tad_: still its a different event te moveto events content are ompatible with undo, so there is no reason afsint having a [allow_undo] there
20160923 22:45:07< tad_> There are 2 capture events and a moveto all trigger on the same tile. The captures are preventing the undo. I'm deleting the allow_undo as dead code
20160923 22:46:34-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20160923 22:47:16< gfgtdf> tad_:i think we shodul rename the variables 'village_$this_item.x|_$this_item.y|_cleared' to 'villages.$this_item.x|_$this_item.y|_cleared' so that in the victory event we cna just clear them by {CLEAR_VARIABLE villages}
20160923 22:47:19< tad_> Well, interesting test. the allow_undo is dead so it's gone but with all the testing I found another bug .. WHEE!!
20160923 22:47:25< zookeeper> fill me in
20160923 22:47:26-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 22:47:28< zookeeper> short version
20160923 22:47:56< tad_> gfgtdf: Actually, I'm thinking they'll all go away like they did in THoT but I'm not quite ready to think about all that.
20160923 22:48:30< tad_> zookeeper: Minor bug. Comment about a unit which didn't appear because random 0..2 gave 0 this time
20160923 22:48:58< zookeeper> right, no undo in that case
20160923 22:49:06< zookeeper> arguably how it should be
20160923 22:49:15< zookeeper> because otherwise you can just move back and get a different result
20160923 22:49:44< zookeeper> ah, nevermind
20160923 22:49:52< zookeeper> the variable check prevents that from happening again
20160923 22:49:54< tad_> zookeeper: The undo I can't get to ever undo anyway. but it's on a captured village and it does randomplacement on the capture evern
20160923 22:51:05< zookeeper> no idea what you mean. so is there any problem or not?
20160923 22:51:16< tad_> so the allow_undo seems dead for a number of reasons, and a bad idea for others, so it's gone. Now to not comment about a unit never created .. then consider doing away with all those variables which really are not needed.
20160923 22:51:28< gfgtdf> zookeeper: using the rng, like in creatign units or using [set_variable] rand= will always make the action non-undoable, regardless of [allow_undo], just like when a move clears shroud/fog.
20160923 22:52:41< zookeeper> gfgtdf, oh that's still the case? right
20160923 22:53:04< gfgtdf> tad_: in the other campaign you seem to have eplac it with a 'owner_side=' check. wouldnt that break in case an enemy captures that willage?
20160923 22:53:22< gfgtdf> replaced*
20160923 22:54:46< tad_> gfgtdf: I've not even considered it yet. I may change it, I may not. Depends on the specifics.
20160923 22:54:48-!- noy [~Noy@wesnoth/developer/noy] has joined #wesnoth-dev
20160923 22:55:32< zookeeper> well i guess there is no problem after all despite all those words
20160923 22:55:45< vultraz> tad_: what are you doing?
20160923 22:55:58< zookeeper> that's what i tried to find out :p
20160923 22:56:07< vultraz> it sounds suspect
20160923 22:57:59< tad_> vultraz: DiD S02 and code reading I wondered about a stray [allow_undo] probably in copied code in a moveto event where two catpure events also trigger on the same cycle for that tile
20160923 22:59:24< zookeeper> there's nothing wrong with that [allow_undo], even if the capture events end up making it meaningless
20160923 22:59:31< vultraz> or just added erroneously 
20160923 22:59:42< vultraz> if it's useless remove it
20160923 23:00:00< zookeeper> it was likely added at a point where gobbos didn't spawn from that village
20160923 23:00:23< tad_> The only way I can get the allow_undo to actually undo is remove shroud/fog and both catpures 
20160923 23:00:23-!- noy [~Noy@wesnoth/developer/noy] has quit [Quit: noy]
20160923 23:01:38< zookeeper> i could have told you all this right away :p
20160923 23:01:54-!- vultraz changed the topic of #wesnoth-dev to: 1.13.6 planned for late September | Wesnoth Inc. board elections: https://r.wesnoth.org/t44616 (deadline is Monday the 26th) | Wesnoth Developers Channel | >>> Want to help? Go here: http://r.wesnoth.org/t42911 (and thanks!) <<< | Logs: http://irclogs.wesnoth.org | Bug tracker: http://bugs.wesnoth.org
20160923 23:02:17< zookeeper> arguably perhaps gobbos shouldn't spawn from that village
20160923 23:02:25< vultraz> Everyone who hasn't voted, please do over the weekend.
20160923 23:02:43< zookeeper> kind of icky when you get the village guard warning and the swamp message at the same time
20160923 23:02:57< tad_> zookeeper: It's fine. It's the only moveto on a village and teh comment is to guide the player into the swamp.
20160923 23:03:27< zookeeper> uh... how is that relevant?
20160923 23:03:36< zookeeper> you're just describing what it is
20160923 23:05:13< tad_> zookeeper: I'm not worried about capture events, or the message. Just the allow_undo seemed strange and I was told it can cause OOS errors so I wanted check on that.
20160923 23:05:50< zookeeper> i guess you weren't replying to the preceding comments i made
20160923 23:08:25-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 260 seconds]
20160923 23:08:55-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160923 23:09:51< zookeeper> and now i'll never know :p
20160923 23:25:11-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20160923 23:26:01-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160923 23:32:28-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 265 seconds]
20160923 23:41:29-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160923 23:47:49-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
--- Log closed Sat Sep 24 00:00:55 2016