--- Log opened Sun May 18 00:00:05 2014
--- Day changed Sun May 18 2014
20140518 00:00:05< iceiceice> that reminds me its time for dinner
20140518 00:00:21-!- iceiceice [~chris@cpe-66-108-20-80.nyc.res.rr.com] has quit [Quit: Leaving]
20140518 00:01:08< Coffee_irc> a fun aside: all java bytecodes are prefixed with "cafebabe" in binary
20140518 00:01:54-!- mjs-de [~mjs-de@f048215213.adsl.alicedsl.de] has quit [Ping timeout: 240 seconds]
20140518 00:18:20-!- RiftWalker [~nathan@ip24-252-126-205.no.no.cox.net] has quit [Remote host closed the connection]
20140518 00:19:12-!- RiftWalker [~nathan@ip24-252-126-205.no.no.cox.net] has joined #wesnoth-dev
20140518 00:21:06-!- RiftWalker [~nathan@ip24-252-126-205.no.no.cox.net] has quit [Client Quit]
20140518 00:29:29-!- ancestral [~ancestral@12.23.74.29] has joined #wesnoth-dev
20140518 00:32:02< gfgtdf> mattsc: online ?
20140518 00:45:42-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has quit [Ping timeout: 258 seconds]
20140518 00:50:43-!- Appleman1234 [~Appleman1@pool-173-74-87-52.dllstx.fios.verizon.net] has joined #wesnoth-dev
20140518 01:14:41< irker252> wesnoth: David Mikos wesnoth:master 6475cd1ad2d7 / / (6 files in 3 dirs): Fix bug 22020: correct help pointer for base units with variations http://git.io/ZYoTxA
20140518 01:14:49< irker252> wesnoth: David Mikos wesnoth:1.12 ced6d9e15f1c / / (6 files in 3 dirs): Fix bug 22020: correct help pointer for base units with variations http://git.io/9WKCyw
20140518 01:20:40-!- Necrosporus_ [~Necrospor@unaffiliated/necrosporus] has joined #wesnoth-dev
20140518 01:23:31-!- Necrosporus [~Necrospor@unaffiliated/necrosporus] has quit [Ping timeout: 240 seconds]
20140518 01:24:53-!- Necrosporus_ is now known as Necrosporus
20140518 01:35:20-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20140518 01:38:39-!- tomreyn [~tomreyn@megaglest/team/tomreyn] has quit [Quit: tomreyn]
20140518 01:42:29-!- travis-ci [~travis-ci@ec2-54-87-68-13.compute-1.amazonaws.com] has joined #wesnoth-dev
20140518 01:42:29< travis-ci> [travis-ci] wesnoth/wesnoth#2688 (1.12 - ced6d9e : David Mikos): The build has errored.
20140518 01:42:29< travis-ci> [travis-ci] Build details : http://travis-ci.org/wesnoth/wesnoth/builds/25429363
20140518 01:42:29-!- travis-ci [~travis-ci@ec2-54-87-68-13.compute-1.amazonaws.com] has left #wesnoth-dev []
20140518 01:45:25-!- gfgtdf [~chatzilla@e177122105.adsl.alicedsl.de] has quit [Quit: ChatZilla 0.9.90.1 [Firefox 29.0.1/20140506152807]]
20140518 01:56:00-!- sachith500 [~kvirc@112.134.100.145] has joined #wesnoth-dev
20140518 01:56:42-!- iceiceice [~chris@cpe-66-108-20-80.nyc.res.rr.com] has joined #wesnoth-dev
20140518 02:12:38< irker252> wesnoth: Chris Beck wesnoth:master 43166e03c56f / src/log.cpp: small refactor of strict mode logging http://git.io/STLmpw
20140518 02:12:40< irker252> wesnoth: Chris Beck wesnoth:master 8fafdbecd42d / src/ (10 files in 2 dirs): rename builder.cpp -> terrain_builder.cpp http://git.io/ujbs-Q
20140518 02:12:42< irker252> wesnoth: Chris Beck wesnoth:master 49dc82b5d8f2 / src/lexical_cast.hpp: bad_lexical_cast derives from std::exception http://git.io/BiTU1g
20140518 02:12:44< irker252> wesnoth: Chris Beck wesnoth:master afb3fec155fc / src/ (5 files in 3 dirs): catch bad lexical casts (from parsing map locations from configs) http://git.io/YxIcPQ
20140518 02:12:46< irker252> wesnoth: Chris Beck wesnoth:master b242735c7e19 / src/ (CMakeLists.txt SConscript tests/test_map_location.cpp): add map_location characterization tests http://git.io/pqhIMw
20140518 02:12:48< irker252> wesnoth: Chris Beck wesnoth:master 39f2b669ae99 / src/synced_commands.cpp: refactor synced_commands [move] to use read_locations, not reimpl. http://git.io/Ddm8zA
20140518 02:12:50< irker252> wesnoth: Chris Beck wesnoth:master 3e6c6d1f168e / src/tests/test_map_location.cpp: fixup test_map_location.cpp http://git.io/cK46KQ
20140518 02:12:52< irker252> wesnoth: Chris Beck wesnoth:master 9e600ac0f657 / src/ (map_location.cpp synced_commands.cpp): Fix map_locations.cpp read_locations function http://git.io/7yfN6Q
20140518 02:12:54< irker252> wesnoth: Chris Beck wesnoth:master 96ba999d49e9 / src/map_location.cpp: improve the performance of map_location.cpp http://git.io/RZa-aQ
20140518 02:12:56< irker252> wesnoth: Chris Beck wesnoth:master d4df8f165ab8 / src/ (20 files in 5 dirs): Merge branch 'map_location' http://git.io/JIdJoA
20140518 02:15:17-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 264 seconds]
20140518 02:16:40< irker252> wesnoth: Chris Beck wesnoth:1.12 7247cb4e2ef6 / src/lexical_cast.hpp: bad_lexical_cast derives from std::exception http://git.io/0MQg7A
20140518 02:16:42< irker252> wesnoth: Chris Beck wesnoth:1.12 a7b81ea2e27d / src/ (5 files in 3 dirs): catch bad lexical casts (from parsing map locations from configs) http://git.io/Ztz_rQ
20140518 02:16:44< irker252> wesnoth: Chris Beck wesnoth:1.12 2ed0908eb595 / src/ (CMakeLists.txt SConscript tests/test_map_location.cpp): add map_location characterization tests http://git.io/Km7rIQ
20140518 02:16:46< irker252> wesnoth: Chris Beck wesnoth:1.12 42ce06bd22a5 / src/synced_commands.cpp: refactor synced_commands [move] to use read_locations, not reimpl. http://git.io/zMdR1g
20140518 02:16:48< irker252> wesnoth: Chris Beck wesnoth:1.12 d0c7e9b6f6af / src/tests/test_map_location.cpp: fixup test_map_location.cpp http://git.io/e0hybQ
20140518 02:16:50< irker252> wesnoth: Chris Beck wesnoth:1.12 7370daa0da47 / src/ (map_location.cpp synced_commands.cpp): Fix map_locations.cpp read_locations function http://git.io/zctHMw
20140518 02:16:52< irker252> wesnoth: Chris Beck wesnoth:1.12 8f03e00e81c7 / / (6 files in 3 dirs): Merge branch '1.12' of git://github.com/wesnoth/wesnoth into 1.12 http://git.io/4kxCag
20140518 02:19:52-!- vultraz [~chatzilla@124.109.10.167] has joined #wesnoth-dev
20140518 02:26:43-!- vultraz [~chatzilla@124.109.10.167] has quit [Ping timeout: 240 seconds]
20140518 02:37:00-!- vultraz [~chatzilla@124.109.10.167] has joined #wesnoth-dev
20140518 02:39:04-!- Ivanovic_ [~ivanovic@frnk-5f751def.pool.mediaWays.net] has joined #wesnoth-dev
20140518 02:42:14-!- Ivanovic [~ivanovic@wesnoth/developer/ivanovic] has quit [Ping timeout: 240 seconds]
20140518 02:42:58-!- Ivanovic_ is now known as Ivanovic
20140518 02:55:07-!- happygrue [~happygrue@wesnoth/developer/wintermute] has quit [Ping timeout: 240 seconds]
20140518 02:56:09-!- bumbadadabum [~bumbadada@d155109.upc-d.chello.nl] has quit [Quit: Ik ga weg]
20140518 03:19:07< mattsc> iceiceice, shadowm: the first thing that comes to my mind when I see ice^3 is the neutrino experiment at the South Pole.
20140518 03:19:17< mattsc> Shows you what kind of geek I am …  :P
20140518 03:20:13< fabi_> mattsc: :-)
20140518 03:20:29< mattsc> hi fabi_ :)
20140518 03:20:32< mattsc> shadowm: do you know what make_neutral_surface is when it comes to colors?
20140518 03:21:19< mattsc> Alarantalara wrote me a message that the color problem with the portraits on OS X is “likely a missing make_neutral_surface somewhere”.
20140518 03:22:03< mattsc> He says that the device RGB order on Macs does not match the neutral RGB order, unlike on most other platforms.
20140518 03:22:27< shadowm> No.
20140518 03:22:56< mattsc> shadowm: ok.  Me neither.
20140518 03:23:29< shadowm> I think it's supposed to give you a surface with a color format that can be safely blitted to the framebuffer.
20140518 03:23:54< shadowm> And that's where my knowledge of low-level SDL crap ends.
20140518 03:24:40-!- sachith500|2 [~kvirc@112.134.47.42] has joined #wesnoth-dev
20140518 03:24:55< shadowm> So yeah, if you think you found a solution, feel free to commit it...
20140518 03:25:46< mattsc> No, not at all.  I’m just passing on information.  I might try the test that gfgtdf suggested, but other than that I’ll wait for mordante to be back before I look into it myself.
20140518 03:26:45< mattsc> I have pretty much zero background knowledge in this, so it wouldn’t be very efficient for me to do this if there are other people who know much more about it.
20140518 03:27:02< mattsc> “other people” = maybe mordante, in this case
20140518 03:27:41< shadowm> Certainly mordante.
20140518 03:27:53-!- sachith500 [~kvirc@112.134.100.145] has quit [Ping timeout: 264 seconds]
20140518 04:22:09-!- iceiceice [~chris@cpe-66-108-20-80.nyc.res.rr.com] has quit [Quit: Leaving]
20140518 04:22:59< irker252> wesnoth: Chris Beck wesnoth:master 259e1f0c8321 / src/map_location.cpp: simplify implementation of read_locations, using STL http://git.io/4gN-nQ
20140518 04:23:14-!- iceiceice_ [~chris@cpe-66-108-20-80.nyc.res.rr.com] has joined #wesnoth-dev
20140518 04:32:28-!- Gallaecio [~quassel@84.120.115.132.dyn.user.ono.com] has quit [Quit: No Ping reply in 180 seconds.]
20140518 04:32:53-!- Gallaecio [~quassel@84.120.115.132.dyn.user.ono.com] has joined #wesnoth-dev
20140518 04:38:38-!- Gambit [~derek@wesnoth/developer/grickit] has quit [Read error: Connection reset by peer]
20140518 04:53:45< iceiceice_> gfgtdf: I actually learned the answer to your question
20140518 04:54:33< iceiceice_> in action_wml, there is a wml message logger function
20140518 04:54:38< iceiceice_> the handler is [wml_message]
20140518 04:54:42< iceiceice_> you can pass it like 
20140518 04:55:21-!- wesbot changed the topic of #wesnoth-dev to: 1.11.14 (1.12 beta 5) is too buggy to be announced | string+feature freeze active on 1.12 | 240 bugs, 344 feature requests, 28 patches | Logs: http://irclogs.wesnoth.org | Alternate logs: http://wesnoth.debian.net | Don't paste on IRC! Use a pastebin: http://pastebin.com | http://imagebin.org
20140518 04:55:23< iceiceice_> logger = "warn", message = "ruh roh!"
20140518 04:55:35< iceiceice_> and it will post things to the corresponding loggers i guess
20140518 04:56:39< iceiceice_> the syntax is here: https://github.com/wesnoth/wesnoth/blob/master/src/game_events/pump.cpp#L404
20140518 05:01:40< iceiceice_> apparently its since v 1.9.4 or smth
20140518 05:06:26< irker252> wesnoth: David Mikos wesnoth:master e92781614dbb / changelog src/unit.cpp: smooth unit movement over terrain with elevation http://git.io/cOpqWQ
20140518 05:07:05< shadowm> 1.5.x, I'm pretty sure.
20140518 05:07:30< shadowm> It's just that I renamed it at some point.
20140518 05:09:07< irker252> wesnoth: Chris Beck wesnoth:master a3377d611382 / projectfiles/ (4 files in 4 dirs): fixup project files after rename builder.?pp -> terrain_builder.?pp http://git.io/RNzrBw
20140518 05:09:09< irker252> wesnoth: Chris Beck wesnoth:master 9fd633e5dad6 / changelog src/unit.cpp: Merge branch 'master' of git://github.com/wesnoth/wesnoth http://git.io/MNJeXA
20140518 05:17:17< iceiceice_> shadowm: ah 
20140518 05:17:56-!- iceiceice_ is now known as iceiceice
20140518 05:19:32-!- [Relic] [~relic@99-58-54-211.lightspeed.milwwi.sbcglobal.net] has quit [Quit: Leaving]
20140518 05:42:08< irker252> wesnoth: David Mikos wesnoth:master 62e527c8b7ad / src/unit.cpp: don't move health bar with unit where terrain elvation differs. http://git.io/AAIM6A
20140518 06:14:05< iceiceice> hmm the plot thickens
20140518 06:14:24< iceiceice> here's a picture of what regions wesnoth thinks are in what directions from a hex: http://i.imgur.com/NbuEfoC.png
20140518 06:14:43< iceiceice> http://i.imgur.com/7NJabWd.png
20140518 06:15:00< iceiceice> when the hex has an odd x coordinate the regions aren't even connected...
20140518 06:15:22< iceiceice> and it doesnt even get the hexes at distance one right
20140518 06:23:09< vultraz> iceiceice: wtf are you running so many instances of e_e
20140518 06:23:35< iceiceice> whats e_e
20140518 06:24:00< vultraz> look at the taskbar
20140518 06:24:03< Necrosporus> iceiceice, some prog with e-like icon
20140518 06:24:23< iceiceice> oh i always do that :)
20140518 06:24:31< iceiceice> uhh
20140518 06:24:35< iceiceice> i guess its 6 evince's with pdfs
20140518 06:24:39< iceiceice> 3 chromes
20140518 06:24:44< iceiceice> 3 gedits
20140518 06:31:05< Coffee_irc> iceiceice: what's the issue exactly?
20140518 06:31:14< Coffee_irc> just looks like a screenshot to me
20140518 06:31:20< iceiceice> so what i did was
20140518 06:31:27< iceiceice> i took the output of map_location relative dir
20140518 06:31:30-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20140518 06:31:33< iceiceice> from the central spot thats labeled void
20140518 06:31:40< Coffee_irc> ok
20140518 06:31:43< iceiceice> and i mapped each direction to a terrain
20140518 06:31:56< iceiceice> so you can see what directions wesnoth thinks are "N, S, NE, NW, SE, SW"
20140518 06:32:04< iceiceice> whats wierd is
20140518 06:32:18< iceiceice> you would think that all the hexes touching the central one
20140518 06:32:20< iceiceice> should each be different
20140518 06:32:27< iceiceice> so the one just north should be one terrain type
20140518 06:32:31< iceiceice> the one just NE should be a different one
20140518 06:32:33< iceiceice> etc.
20140518 06:32:35< iceiceice> because,
20140518 06:32:38< Coffee_irc> oh, I see
20140518 06:32:44< iceiceice> if there was a unit at the void,
20140518 06:32:50< iceiceice> that's how wesnoth would select which way its facing
20140518 06:32:58< iceiceice> if it attacked / moved to one of those hexes
20140518 06:33:08< c74d> e_e is an emoticon vultraz uses that supposedly represents a squinting face.
20140518 06:34:31< Coffee_irc> iceiceice: yeah, the alrogithm is all wrong
20140518 06:34:59< Coffee_irc> it will only work for even vertical hex lines
20140518 06:35:41< Coffee_irc> I wrote an algorithm to calculate this once
20140518 06:35:48< iceiceice> really?
20140518 06:35:55< Coffee_irc> so that the default unit direction would be set for units
20140518 06:36:00< Coffee_irc> as a patch
20140518 06:36:06< Coffee_irc> that didn't get anywhere
20140518 06:36:20< Coffee_irc> kind of forgot about it
20140518 06:36:46< iceiceice> was it going to go into get_relative_dir or somewhere else?
20140518 06:37:08< Coffee_irc> I tried using relative_dir and realized it didn't work properly in all cases
20140518 06:37:19< Coffee_irc> it was to set the default facing direction for units
20140518 06:37:21< iceiceice> y i see
20140518 06:37:30< iceiceice> i think we should just patch relative_dir maybe
20140518 06:37:53< iceiceice> esp. if you know what a better algorithm is
20140518 06:38:07< Coffee_irc> iceiceice: http://pastebin.com/iSbhTxPc for reference
20140518 06:38:45< Coffee_irc> slightly different thing though
20140518 06:39:02< Coffee_irc> I wrote a couple of different ones
20140518 06:40:38< Coffee_irc> interesting that the legacy_sum_assign function used for this in map_location.cpp seems to deal with odd columns
20140518 06:40:52< iceiceice> y i dont really get it
20140518 06:40:55< iceiceice> its strange though,
20140518 06:41:10< iceiceice> i foudn that the picture for x = 7, y =8 is the same whether i use legacy difference or regualr difference
20140518 06:41:12< Coffee_irc> when you move one hex east or west it moves up/down one
20140518 06:41:59< Coffee_irc> so the north east is either one across, or one across and one up
20140518 06:48:43< Coffee_irc> so, if going from y to y2, north should be y<y2, south y>y2
20140518 06:49:59< Coffee_irc> it's hard to split whether a direction is more north or north east
20140518 06:50:30< Coffee_irc> I'm going to try to draw it in the editor
20140518 06:51:17< iceiceice> y i'm not sure
20140518 06:51:24< iceiceice> i mean i think we have a "rotate" function that makes sense
20140518 06:51:25< Coffee_irc> fabi_'s scenario editor allows labels to be drawn
20140518 06:51:43< iceiceice> i would think that it would be good if the regions rotate onto one another
20140518 06:52:07< iceiceice> idk how easy it is to write a little function that will do that though
20140518 06:52:51< Coffee_irc> yeah, we get hexes that could be either north or northeast
20140518 06:57:34-!- iceiceice [~chris@cpe-66-108-20-80.nyc.res.rr.com] has quit [Ping timeout: 240 seconds]
20140518 06:57:55-!- Sulfur [~Miranda@p5B327CD8.dip0.t-ipconnect.de] has joined #wesnoth-dev
20140518 07:08:08-!- iceiceice [~chris@cpe-66-108-20-80.nyc.res.rr.com] has joined #wesnoth-dev
20140518 07:08:12< Coffee_irc> basically we want to minimize the number of hexes from the lines that make N,NE,NW,S,SE,SW
20140518 07:08:19< iceiceice> yeah
20140518 07:08:31< Coffee_irc> so, my http://i.imgur.com/DBUkzdb.jpg is not quite right
20140518 07:08:50< iceiceice> so this i what i think might make sense
20140518 07:08:59-!- cib_ [~cib@p5DC75E1A.dip0.t-ipconnect.de] has joined #wesnoth-dev
20140518 07:09:17< Coffee_irc> iceiceice: it is easy to calculate the distance in hexes from a line
20140518 07:09:21< iceiceice> err ok im going to propose something that makes sense to me mathematically
20140518 07:09:26< iceiceice> idk if its okay visually
20140518 07:09:39< iceiceice> but anyways what i would think you want to do is...
20140518 07:09:45< iceiceice> just sort of take the six lines going out
20140518 07:09:52< iceiceice> in each of the directions
20140518 07:09:59< iceiceice> and see whih region you end up in
20140518 07:10:02< iceiceice> and if you are on a line,
20140518 07:10:05< iceiceice> then bias counter clockwise
20140518 07:10:08< iceiceice> or smth like that
20140518 07:10:29< iceiceice> so i guess if i wrote it might do something like,
20140518 07:10:33< iceiceice> if diff.x == 0
20140518 07:10:41< iceiceice> then, if above its north, if below its south
20140518 07:10:44< iceiceice> otherwise is x > 0,
20140518 07:10:52< iceiceice> take the negative of diff... and assume henceforth x < 0
20140518 07:11:04< iceiceice> then...
20140518 07:11:07< iceiceice> i have to think for am inute,,,
20140518 07:11:56< iceiceice> so i guess i want to think about the two lines going NW and SW from the origin
20140518 07:12:17< Necrosporus> I wonder, is WML simple and easy to use or opposite. Many things seem quite hard to implement
20140518 07:12:26< Coffee_irc> iceiceice: I've drawn a picture of what it should be http://imgur.com/8rGQ6BX
20140518 07:12:27< iceiceice> i'm not sure off the top of my head what the equations are in hex grids... but anyways for some x there is a unique NWy
20140518 07:12:30< iceiceice> and a unique SWy
20140518 07:12:39< iceiceice> and i would compare my y to those lines
20140518 07:12:58< Coffee_irc> the gaps could be either of what is next to it and it would be fine
20140518 07:13:36< Coffee_irc> iceiceice: we want to minimize the distance in hexes to any of the lines that intersect the starting hex
20140518 07:13:39< iceiceice> ok i see 
20140518 07:13:42< iceiceice> my thing is wrong,
20140518 07:13:57< Coffee_irc> I think it is easier than we think
20140518 07:14:00< iceiceice> mine would be correct if we wanted directions W E, NW, NE, SW , SE or something
20140518 07:14:50< Coffee_irc> the distance to north line would be (for y diff >0) just abs(x diff)
20140518 07:14:55-!- mordante [~mordante@roadie.xs4all.nl] has joined #wesnoth-dev
20140518 07:14:55-!- mordante [~mordante@roadie.xs4all.nl] has quit [Changing host]
20140518 07:14:55-!- mordante [~mordante@wesnoth/developer/mordante] has joined #wesnoth-dev
20140518 07:14:56< Coffee_irc> similarly for the south line
20140518 07:15:10< mordante> servus
20140518 07:15:16< Coffee_irc> hi mordante
20140518 07:15:30< mordante> hi Coffee_irc 
20140518 07:15:40< iceiceice> y so what's the border between SW and S in your picture
20140518 07:15:50< iceiceice> its just like, down one, over one, down one, over one?
20140518 07:16:00< iceiceice> hi mordante
20140518 07:16:08< Coffee_irc> iceiceice: ?
20140518 07:16:57< Coffee_irc> each label has to be there in that picture I think because it minimizes the distance to the nearest directional line
20140518 07:17:21< Coffee_irc> the SW are at most one hex away from the line
20140518 07:17:29< Coffee_irc> and similarly for all the other directions
20140518 07:17:55< Coffee_irc> the gaps could be either of what is around them
20140518 07:18:01< mordante> mattsc, shadowm the difference between SDL_Blit and blit_surface is how they handle transparency
20140518 07:19:14< Coffee_irc> iceiceice: the diagonal differences are just abs(dy - dx) and abs(dx - dy) for dy>0 and dy<0
20140518 07:19:59< Coffee_irc> remembering to add or subtract 1 from the y coordinate if even/odd rows
20140518 07:20:27< Coffee_irc> which is kind of similar to what we have already :P
20140518 07:20:45< Coffee_irc> just that it doesn't work, probably because of attention to detail
20140518 07:21:13< iceiceice> hmm so you think the NE line is dy = dx? or something like this?
20140518 07:21:19< iceiceice> err NW
20140518 07:21:35< mordante> mattsc, shadowm I assume https://gna.org/bugs/index.php?22045 is a bug in the implementation of blit_surface
20140518 07:21:47< Coffee_irc> iceiceice: yeah, we want the deviation from that
20140518 07:22:01< Coffee_irc> so minimize abs(dy - dx) for dy>0
20140518 07:23:21< iceiceice> y its pretty simple when you say it that way :p
20140518 07:23:26< iceiceice> uhm ok
20140518 07:23:32< iceiceice> so this is what i think to do
20140518 07:23:42< iceiceice> if x = 0, its easy
20140518 07:23:57< iceiceice> if x > 0, then negate the diff vector, see what answer you get, and give opposite_dir of that
20140518 07:24:03< iceiceice> so you restrict attention totally to x < 0
20140518 07:24:04< iceiceice> then,
20140518 07:24:17< iceiceice> check if dy <= -dx, then it is south
20140518 07:24:33< iceiceice> else, if dy <= 0 then it is south east
20140518 07:24:46< iceiceice> *south west
20140518 07:24:54< iceiceice> else, if dy <= dx, then it is north west
20140518 07:24:58< iceiceice> else, it is north
20140518 07:25:19< iceiceice> except that, maybe the dy <= 0 needs to be some like 0 or 1 depending on parity of x or smth
20140518 07:25:28< iceiceice> not totally sure :/
20140518 07:25:28< Coffee_irc> iceiceice: eh, you will get like what we have
20140518 07:25:44< iceiceice> hmm so heres what i hope to improve
20140518 07:25:45< Coffee_irc> you won't minimize the distance to the line without calculating the distances and taking the smallest one
20140518 07:25:53< Coffee_irc> http://pastebin.com/Mvhf8jaP
20140518 07:25:57< mordante> iceiceice, did travis survive your force push?
20140518 07:26:06< iceiceice> mordante: i have no idea
20140518 07:26:13< iceiceice> i think it built something coffee pushed later
20140518 07:26:21< iceiceice> but i really have no idea how to even check that
20140518 07:26:45< Coffee_irc> iceiceice: the above pastebin will calculate the minimum distances to all the lines
20140518 07:27:25< Coffee_irc> I wrote that to set units to face eachother
20140518 07:27:32< Coffee_irc> but it is the same principal
20140518 07:27:39< Coffee_irc> *principle
20140518 07:27:44< iceiceice> ok 
20140518 07:27:54< iceiceice> so i dont know what we consider to be the goal post
20140518 07:27:57< iceiceice> to me,
20140518 07:28:08< iceiceice> i think it would be good if a.get_dir(b) = opposite of b.get_dir(a)
20140518 07:28:22< Coffee_irc> iceiceice: I'm happy to rewrite the get_relative_dir function
20140518 07:28:36< iceiceice> and also i thinkt he things at distance 1 should especially satsify that
20140518 07:28:49< iceiceice> like, the point that is north east of your hex... should have relative dir NE
20140518 07:28:51< Coffee_irc> and let you know when I'm done to test with that map algorithm
20140518 07:28:58< iceiceice> ok
20140518 07:29:10< iceiceice> hmm so i can also write a second unit test
20140518 07:29:22< iceiceice> that will check the distance one points
20140518 07:29:44< iceiceice> i was going to write one to check that adjacent hexes are exactly the ones at distance one according to our distance thing...
20140518 07:29:54< iceiceice> because it would be pretty bad if that failed
20140518 07:30:55< iceiceice> idk it would be wierd if you have two units attack
20140518 07:31:04< iceiceice> and they aren't facing eachother when they do so
20140518 07:39:07< Coffee_irc> iceiceice: http://pastebin.com/Zsfr16h9
20140518 07:39:28< Coffee_irc> I think this should do it for the get_relative_dir function
20140518 07:39:48< Coffee_irc> either that or it is the wrong way around :P
20140518 07:40:00< Coffee_irc> and will produce the opposite pattern
20140518 07:40:11< Coffee_irc> I couldn't tell which should be relative to which
20140518 07:40:16< iceiceice> ok give me just a sec to make a picture
20140518 07:41:34< Coffee_irc> priority in this if the distance is even should go to N/S
20140518 07:45:34-!- shadowm_desktop [ignacio@wesnoth/developer/shadowmaster] has quit [Ping timeout: 240 seconds]
20140518 07:47:37-!- shadowm_desktop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20140518 07:47:45< iceiceice> sry it took a while, i decided to clean build, i had some wierd stuff happen earlier
20140518 07:48:20< Coffee_irc> fine by me
20140518 07:50:04< iceiceice> hm ok so its not returning NDIRECTIONS when you give it 0,0
20140518 07:50:09< iceiceice> so the pictures are a little funny
20140518 07:50:13< iceiceice> but anyways here they are:
20140518 07:51:07< iceiceice> heres with x = 7: http://i.imgur.com/M4HiiDs.png
20140518 07:51:29< iceiceice> heres with x= 8
20140518 07:51:30< iceiceice> http://i.imgur.com/v9I7g8w.png
20140518 07:53:28< Coffee_irc> iceiceice: should it return NDIRECTIONS when the start and end are the same?
20140518 07:53:56< iceiceice> i guess, thats what its been doing
20140518 07:55:54< Coffee_irc> iceiceice: http://pastebin.com/Aj1zxqzq
20140518 07:56:18< iceiceice> lol here was my attempt
20140518 07:56:21< iceiceice> iceiceice x=7: http://i.imgur.com/fo6ouJJ.png
20140518 07:56:39< iceiceice> iceiceice x=8: http://i.imgur.com/ny2yeI1.png
20140518 07:57:11< Coffee_irc> iceiceice: is that your code?
20140518 07:59:04< mordante> mattsc, around?
20140518 07:59:12< iceiceice> Coffee_irc: yeah...
20140518 07:59:25< Coffee_irc> iceiceice: is the test scenario in WML?
20140518 07:59:28< iceiceice> i had some NORTH / SOUTH mixup on the x=0 line
20140518 07:59:33< iceiceice> no its not WML
20140518 07:59:47< iceiceice> i'll show oyu one sec
20140518 08:00:50< iceiceice> its this branch: https://github.com/cbeck88/wesnoth/commits/test_relative_dir
20140518 08:00:51-!- sachith500|2 [~kvirc@112.134.47.42] has quit [Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/]
20140518 08:00:52< iceiceice> just one commit
20140518 08:00:58< iceiceice> https://github.com/cbeck88/wesnoth/commit/517c8a8e1ee6d336b71a5c824f937932a5738bcd
20140518 08:00:59-!- sachith500 [~kvirc@112.134.47.42] has joined #wesnoth-dev
20140518 08:01:30< iceiceice> you just compile the test executable,
20140518 08:01:42< iceiceice> and copy the stuff in between the *'s into a .map file :)
20140518 08:06:04< Coffee_irc> iceiceice: don't mind if I ask how to run the test?
20140518 08:06:57< iceiceice> hehe
20140518 08:07:03< iceiceice> for me its `scons test`
20140518 08:07:06< iceiceice> then when its done,
20140518 08:07:10< iceiceice> ./test
20140518 08:07:16< iceiceice> in the same place as you would normally go ./wesnoth
20140518 08:07:25< iceiceice> but idk are you on a fancy ide or smth?
20140518 08:07:39< iceiceice> Coffee_irc: here was my version: https://github.com/cbeck88/wesnoth/blob/iceiceice_relative_dir/src/map_location.cpp#L207
20140518 08:07:46< Coffee_irc> iceiceice: I am just using a terminal
20140518 08:07:59< Coffee_irc> I use geany for writing code
20140518 08:08:10< iceiceice> oh ok, yeah then i guess ./test should work
20140518 08:08:20< Coffee_irc> in the past I used eclipse but found it a pain to keep up to date
20140518 08:08:27< iceiceice> i ask because i'm not sure if codeblocks and the others even build the test executable
20140518 08:12:29< iceiceice> hmm y theres something wierd going on...
20140518 08:13:04< iceiceice> i dont really understand why my thing is not symmetric about the origin
20140518 08:13:06< Coffee_irc> iceiceice: characterize_distance is using the old algorithm...
20140518 08:13:25< iceiceice> y 
20140518 08:13:28< iceiceice> if you want to update,
20140518 08:13:36< iceiceice> you have to change the macro flag at the top
20140518 08:14:01< iceiceice> to define MAP_LOCATION_GET_OUTPUT
20140518 08:14:04< iceiceice> then run it once
20140518 08:14:15< iceiceice> take all the (std::make_pair... ) from console
20140518 08:14:22< iceiceice> and paste it in to the .cpp file
20140518 08:15:25< Coffee_irc> iceiceice: are you sure the test is good?
20140518 08:15:40< Coffee_irc> you can choose a border around or not
20140518 08:15:55< Coffee_irc> and this will skew the results by 1 in the vertical direction
20140518 08:16:06< iceiceice> for whcih function
20140518 08:16:15< Coffee_irc> in the generated map
20140518 08:16:28< iceiceice> oh the visualization?
20140518 08:16:33< Coffee_irc> yeah
20140518 08:16:41< iceiceice> idk i didnt realize that would change it
20140518 08:16:48< iceiceice> what do you think makes the most sense?
20140518 08:17:09< iceiceice> i thought we were trying to deprecate all the "format = map  border =1 " stuff
20140518 08:18:25< Coffee_irc> iceiceice: ok
20140518 08:18:27< iceiceice> hmmso if you use border = 1 what will happen though
20140518 08:18:29< Coffee_irc> it makes no difference
20140518 08:18:34< iceiceice> ok
20140518 08:23:08< Coffee_irc> iceiceice: it doesn't appear to be using my function?
20140518 08:23:36< Coffee_irc> I just set it to return NDIRECTIONS all the time and it prints out the maps like you got earlier?
20140518 08:23:38< iceiceice> y i had some issues like that i'm not sure whats up
20140518 08:23:52< iceiceice> i think there might be something about scons test that doesnt always rebuild core or smth?
20140518 08:23:54< iceiceice> i really have no idea
20140518 08:23:57< iceiceice> i was doing scons -c 
20140518 08:23:59< iceiceice> to clean 
20140518 08:24:01< iceiceice> before each time
20140518 08:24:10< iceiceice> its tolerable because i have ccache enabled
20140518 08:24:13< Coffee_irc> ...
20140518 08:24:19< Coffee_irc> this will take a while on my computer
20140518 08:24:33< iceiceice> idk whats the alternative
20140518 08:24:43< iceiceice> i guess could paste the code somwhere in game.cpp
20140518 08:24:50< iceiceice> so it runs at start up when you do wesnoth
20140518 08:25:01< iceiceice> or osmeting like that
20140518 08:25:29< iceiceice> i don't think theres anyway to access get_dir in wml right now
20140518 08:33:00-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20140518 08:36:48< Coffee_irc> iceiceice: that looks like the best way to go
20140518 08:39:50< iceiceice> hmm so this is a bit funny
20140518 08:40:09< iceiceice> since my picture wasn't symmetric i was wondering if something was wrong in the vector_negation code
20140518 08:40:16< iceiceice> but i just made a sanity check of it and it looks good
20140518 08:40:36< iceiceice> like i made just an arbitrary walk from 0,
20140518 08:40:42< iceiceice> and a second walk that did the opposite direction steps
20140518 08:40:47< Coffee_irc> these problems always start out fun and then suck after a while LO
20140518 08:40:54< iceiceice> and each time asserted the two vectors were vector negation...
20140518 08:41:01< iceiceice> y idk...
20140518 08:41:18< Coffee_irc> my code has a minor maths error in it
20140518 08:42:14< iceiceice> hehe
20140518 08:42:20< iceiceice> when i tried to google what the answer should be,
20140518 08:42:24< iceiceice> i ended up here: http://en.wikipedia.org/wiki/Hex_map
20140518 08:42:27< iceiceice> apparently we are the experts :p
20140518 08:45:47< iceiceice> are you pushing to a fork or working locally?
20140518 08:46:15< iceiceice> i'm still thinking that i like my idea of restrciting attention to x < 0, i want to copy your code after and see what happens :p
20140518 08:47:33< iceiceice> hmm i'm looking at your check here: 	int dx = loc.x - x;
20140518 08:47:33< iceiceice> 	int dy = loc.y - y;
20140518 08:47:33< iceiceice> 	if (loc.x%2==0 && x%2==1) dy--;
20140518 08:47:33< iceiceice> 	
20140518 08:47:44< iceiceice> why dont you just use the vector_difference_assign thingie
20140518 08:47:59< Coffee_irc> iceiceice: because it gives the wrong result :P
20140518 08:48:06< iceiceice> really?
20140518 08:48:16< iceiceice> you mean like overall or just in this case 
20140518 08:48:21< Coffee_irc> well, the map doesn't look right does it?
20140518 08:48:48< iceiceice> hehe
20140518 08:49:34< Coffee_irc> it may come down to actually drawing a series of points in a table by hand and making an equation to match :P
20140518 08:50:24< Coffee_irc> I suppose I could use excel
20140518 09:01:18< iceiceice> hmm ok
20140518 09:01:21< iceiceice> heres my idea now
20140518 09:01:32< iceiceice> so i'm only thinking about the case x < 0
20140518 09:01:39< iceiceice> so this sounds really silly but anyways
20140518 09:01:52< iceiceice> how about just walk from zero 
20140518 09:01:58< iceiceice> alternating SW, S, using get_direction
20140518 09:02:05< iceiceice> until you get to the right x coordinate
20140518 09:02:25< iceiceice> hmm ok i'm still not using your distance idea exactly
20140518 09:02:30< iceiceice> but i think it should work out the same maybe
20140518 09:02:38-!- shadowm_desktop [ignacio@wesnoth/developer/shadowmaster] has quit [Ping timeout: 240 seconds]
20140518 09:03:14< iceiceice> but anyways basically just literally follow two rays out that are in the two critical directions
20140518 09:03:25< iceiceice> one alternatiing SW S
20140518 09:03:29< iceiceice> and one altnerating NW N i guess
20140518 09:03:34< iceiceice> and see which you are above / below
20140518 09:03:53-!- sachith500 [~kvirc@112.134.47.42] has quit [Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/]
20140518 09:03:55< iceiceice> and then if that looks right, good
20140518 09:04:00< iceiceice> if we can refactor later to make it an equation, great
20140518 09:15:37-!- mjs-de [~mjs-de@f048215213.adsl.alicedsl.de] has joined #wesnoth-dev
20140518 09:42:02< iceiceice> hmm i wonder if theres a bug in get_direction
20140518 09:42:16< Coffee_irc> iceiceice: I don't know
20140518 09:42:27< Coffee_irc> but the test scenario is suspiscious :P
20140518 09:42:52< iceiceice> hehe maybe theres a bug in the picture code, its true
20140518 09:43:03< Coffee_irc> I've done a manual get_relative_dir for 6,8 vs 7,8 and got a different answer than is on the image
20140518 09:43:35< Coffee_irc> I've now got a nice excel table that matches what the output should be (I think)
20140518 09:43:46< Coffee_irc> and it doesn't work! :P
20140518 09:48:36< Coffee_irc> iceiceice: oh
20140518 09:48:41< Coffee_irc> I think I know :)
20140518 09:48:52< Coffee_irc> the test maybe round the wrong way
20140518 09:49:05< Coffee_irc> testing the relative direction the wrong way around
20140518 09:49:31< Coffee_irc> where the may be multiple answers it may be producing the funny results
20140518 09:50:50< iceiceice> y so i understand that
20140518 09:51:00< iceiceice> i feel like the more fundamental problem is that we seem to get the hexes at distance one wrong
20140518 09:51:19< iceiceice> but i think you are furthr along than me maybe you got that already
20140518 09:51:53-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20140518 09:52:17< Coffee_irc> iceiceice: "scons test" seems to work without a clean build
20140518 09:53:37< Coffee_irc> well, it did once, and now doesn't :P
20140518 09:55:52< iceiceice> yeah tbh thats the only thing that is really disturbing me 
20140518 10:00:01-!- sachith500 [~kvirc@112.134.47.42] has joined #wesnoth-dev
20140518 10:02:00< iceiceice> hmm well everything is passing my tests, i guess i'm the one whose crazy :)
20140518 10:02:18< iceiceice> ok i think i'm just going to make my get_dir algorithm just spam all the debugging info
20140518 10:06:30< Coffee_irc> iceiceice: I've just bypassed the test area and copied that code into the main game loop
20140518 10:07:01< iceiceice> the test area is not so slow for me
20140518 10:07:11< iceiceice> i can do clean build of wesnoth, wesnothd, test, in about 5 minutes
20140518 10:07:13< Coffee_irc> iceiceice: it doesn't save changes for me
20140518 10:07:21< iceiceice> even with scons -c?
20140518 10:07:29< Coffee_irc> yeah, but that takes a lot of time
20140518 10:07:43< iceiceice> y with ccache its not bad
20140518 10:08:04< iceiceice> but its cause i have an ssd, and i set my ccache cache size to liek 5 gigabytes
20140518 10:08:07< Coffee_irc> iceiceice: I have a feeling that I've almost got it :P
20140518 10:08:38< Coffee_irc> yep
20140518 10:08:49< Coffee_irc> I think I can produce the required maps now ;)
20140518 10:09:18< Coffee_irc> http://pastebin.com/hJwMw9j0
20140518 10:09:23< Coffee_irc> iceiceice: ^
20140518 10:09:43< Coffee_irc> bloody hex coordinates :P
20140518 10:10:08< Coffee_irc> at least a heptagon wasn't chosen
20140518 10:11:30< iceiceice> lol
20140518 10:11:39< iceiceice> so you just tweaked the constants in your original idea?
20140518 10:11:47< Coffee_irc> iceiceice: yeah
20140518 10:12:01< Coffee_irc> I was off by a small bit in one equation
20140518 10:13:55< iceiceice> i have to say its going to be really wierd to me if dx, dy is not actually the same as the vector_difference of those two
20140518 10:16:57 * Coffee_irc *sniff*
20140518 10:17:06< Coffee_irc> it still isn't quite right :P
20140518 10:17:11< Coffee_irc> but much better than before
20140518 10:17:13< iceiceice> its really really close
20140518 10:17:33< Coffee_irc> it's just the north south that is off at the moment
20140518 10:18:51< iceiceice> the other thing is its not radially symmetric
20140518 10:18:53< iceiceice> like,
20140518 10:18:58< iceiceice> this one still doesnt have the thing i was saying
20140518 10:19:16< iceiceice> a.relative_dir(b) = opposite(b.relative_dir(a))
20140518 10:19:29< iceiceice> you can see on the horizontal line through the void,
20140518 10:19:53< Coffee_irc> the north and south are different, yes
20140518 10:19:58< Coffee_irc> but that isn't the problem
20140518 10:19:59< iceiceice> if 8,8 is void and 6,8 is swamp, 10,8 should be desert not forest imo
20140518 10:20:32< Coffee_irc> iceiceice: I see
20140518 10:20:39< Coffee_irc> that can be done
20140518 10:21:00< Coffee_irc> but they both have the same distance from the diagonals
20140518 10:21:05< Coffee_irc> so it works with forst or sand
20140518 10:21:48< Coffee_irc> it's the south direction that's stuffed for some reason unknownst to me
20140518 10:22:02< Coffee_irc> everything else is good
20140518 10:31:42< iceiceice> ok
20140518 10:32:53< iceiceice> heres my version, i'm trying to make apicture now...
20140518 10:32:54< iceiceice> https://github.com/cbeck88/wesnoth/blob/ice_coffee_relative_dir/src/map_location.cpp#L235
20140518 10:34:53< iceiceice> whoa thats messed up somehow
20140518 10:34:54< iceiceice> lol
20140518 10:34:58< iceiceice> i can already tell from the text
20140518 10:42:32-!- ancestral [~ancestral@12.23.74.29] has quit [Quit: i go nstuf kthxbai]
20140518 10:43:07-!- bumbadadabum [~bumbadada@d155109.upc-d.chello.nl] has joined #wesnoth-dev
20140518 10:46:27< iceiceice> ok i tried again,
20140518 10:46:29< iceiceice> this looks pretty good
20140518 10:46:34< iceiceice> heres current pcitures: http://i.imgur.com/Ink4rYe.png
20140518 10:47:00< iceiceice> http://i.imgur.com/2NRgBUl.png
20140518 10:47:28< iceiceice> hmm uploaded wrong one fhte first time
20140518 10:48:34< iceiceice> hmm i thik i just have browser caching or something, i guess the link is fine
20140518 10:49:21< Coffee_irc> iceiceice: it's also very close and also has a circular symmetry
20140518 10:49:32< iceiceice> y theres some problems at radius 2 or something but
20140518 10:50:00< iceiceice> all i did was add this to your code: 
20140518 10:50:22< iceiceice> http://pastebin.com/VdBZWavs
20140518 10:51:54< Coffee_irc> I didn't expect this to be so hard :P
20140518 10:52:01< iceiceice> hmmm
20140518 10:52:03< iceiceice> i think maybe
20140518 10:52:22< iceiceice> the line dist_diag_SW_NE < dist
20140518 10:52:24< iceiceice> shoudl become <= 
20140518 10:52:25< iceiceice> or something
20140518 10:52:41< iceiceice> or perhaps the other
20140518 10:52:53< iceiceice> idk maybe you already tried stuff like that
20140518 10:56:18-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 240 seconds]
20140518 10:56:27-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20140518 11:10:19-!- Duthlet [~Duthlet@wesnoth/mp-mod/Duthlet] has joined #wesnoth-dev
20140518 11:11:27-!- Gambit [~derek@wesnoth/developer/grickit] has joined #wesnoth-dev
20140518 11:12:32-!- tomreyn [~tomreyn@megaglest/team/tomreyn] has joined #wesnoth-dev
20140518 11:13:40-!- sachith500 [~kvirc@112.134.47.42] has quit [Quit: KVIrc 4.2.0 Equilibrium http://www.kvirc.net/]
20140518 11:21:25< Coffee_irc> iceiceice: finally I got it :)
20140518 11:21:32< iceiceice> lol
20140518 11:21:39< iceiceice> what did you do?
20140518 11:24:01< Coffee_irc> http://pastebin.com/0X6xtdgi
20140518 11:24:06< Coffee_irc> iceiceice: ^
20140518 11:24:34< Coffee_irc> I don't even care anymore how it works :P
20140518 11:25:05< iceiceice> lol wtf
20140518 11:25:22< iceiceice> ok i have to see this
20140518 11:25:36< vultraz> What's that?
20140518 11:26:03< Coffee_irc> vultraz: it is a working version of the get_relative_dir in map_location.cpp
20140518 11:26:20< Coffee_irc> and iceiceice has found a way to turn it into a nice picture
20140518 11:27:45< vultraz> It seems...inelegant
20140518 11:27:58 * Coffee_irc pokes vultraz in the eye :P
20140518 11:28:18< Coffee_irc> we've been working on this for hours :P
20140518 11:28:49< Coffee_irc> actually it is a much better implementation than before
20140518 11:28:57< Coffee_irc> hence the extra complexity
20140518 11:29:05< vultraz> It returns a direction from a hex?
20140518 11:29:30< Coffee_irc> vultraz: yes
20140518 11:29:37< Coffee_irc> and always returns the correct answer
20140518 11:29:44< Coffee_irc> unlike the previous implementation
20140518 11:30:11< vultraz> Didn't mattsc do that in lua
20140518 11:30:17< vultraz> At one point, for shadowm
20140518 11:30:43< Coffee_irc> eh hrmm... maybe
20140518 11:30:52< vultraz> I think it was uh
20140518 11:30:58< Coffee_irc> but I bet it wasn't as good ;)
20140518 11:31:05< vultraz> I think shadowm used it
20140518 11:31:07< vultraz> 1 + math.floor((hex_angle(a, b) / math.pi * n/2 + d_east) % n)
20140518 11:31:41< Coffee_irc> vultraz: ah, but not pis needed here :)
20140518 11:31:41< vultraz> see https://github.com/shikadilord/After_the_Storm/blob/master/lua/hex.lua
20140518 11:32:42< Coffee_irc> well...
20140518 11:32:49< Coffee_irc> that uses floating point maths
20140518 11:33:00< Coffee_irc> this one is more advanced and uses only discrete maths :)
20140518 11:33:03< iceiceice> y this is probably much better
20140518 11:33:08< iceiceice> in temrs of speed
20140518 11:33:16< iceiceice> maybe we should write a proper lua hook for it
20140518 11:33:47< iceiceice> hehe did mattsc make a picture with his?
20140518 11:34:03< vultraz> Not that I know of
20140518 11:34:10< iceiceice> its probably slightly buggy than
20140518 11:34:18< iceiceice> even if its floating point 
20140518 11:34:45< vultraz> I have no idea
20140518 11:34:53< vultraz> Don't ask me about le complex maths
20140518 11:36:11< Coffee_irc> well, iceiceice, do you think I should commit the new (C++) version?
20140518 11:38:10< Coffee_irc> we can remove the legacy code then also
20140518 11:39:39< iceiceice> yeah that would be good :)
20140518 11:39:47< iceiceice> i still think it should have the thing to make it radially symmetric
20140518 11:39:58< iceiceice> unless something is messed up in my build of the final
20140518 11:40:13< Coffee_irc> iceiceice: it is symmetric up and down and left to right
20140518 11:40:28< Coffee_irc> in groups of 3 from the centre it is pretty steady also
20140518 11:40:58< iceiceice> it still doesnt make my earlier assertion okay about a.dir(b) = opposite(b.dir(a))
20140518 11:41:31< iceiceice> but i guess can live with it, esp if the purpose is units facing
20140518 11:41:34< Coffee_irc> iceiceice: why should it be radially symmetric?
20140518 11:41:38< iceiceice> idk
20140518 11:41:42< iceiceice> it just made sense
20140518 11:42:03< Coffee_irc> you would be introducing a bias clockwise or anticlockwise
20140518 11:42:08< iceiceice> sure
20140518 11:42:20-!- Sulfur [~Miranda@p5B327CD8.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20140518 11:42:23< iceiceice> right now there is a bias south i guess
20140518 11:43:13< iceiceice> maybe the game looks better if more units are facing south?
20140518 11:43:25< Coffee_irc> I'm happy to go with that :)
20140518 11:44:06< iceiceice> alright :)
20140518 11:44:31< iceiceice> i think i might basically commit the picture into the unit test like i did with the other ones
20140518 11:44:40< Coffee_irc> ok
20140518 11:45:00< iceiceice> hmm wait so why is this okay to delete the legacy stuff now also?
20140518 11:45:14< Coffee_irc> iceiceice: well...
20140518 11:45:29< Coffee_irc> I was thinking that there was a reason for it to be called legacy
20140518 11:45:35< Coffee_irc> like it didn't work
20140518 11:45:46< Coffee_irc> but sort of worked, hence the legacy tag
20140518 11:47:29< irker252> wesnoth: David Mikos wesnoth:master 829b74c2beaa / src/map_location.cpp: Fix implementation of map_location::get_relative_dir http://git.io/fYeR9A
20140518 11:49:15< Coffee_irc> iceiceice: the hpp file says that the code is to be avoided :P
20140518 11:49:15< iceiceice> idk i guess no one wrote back to my email yet
20140518 11:49:23< iceiceice> i wasnt sure if the reason is,
20140518 11:49:34< iceiceice> changing it out will force people to rewrite a bunch of map generator code or smth
20140518 11:49:52< iceiceice> like, WML in some mainline scnearios
20140518 11:50:00< iceiceice> but anyways if it really is now only being kept around for the terrain_builder,
20140518 11:50:08< iceiceice> then maybe we should kick the legacy code out of map_location,
20140518 11:50:15< iceiceice> and make it private to terrain_builder
20140518 11:50:33< Coffee_irc> I'm compiling aafter deleting it on the offchance it isn't used anywhere
20140518 11:50:46< iceiceice> hehe ok
20140518 11:51:11< Coffee_irc> although it has been somewhat cursed so far, I'm not expecting much
20140518 11:53:50< Coffee_irc> ok, the legacy code stays :P
20140518 11:54:35< Coffee_irc> to remove it from the terrain builder would take another afternoon
20140518 11:57:04< iceiceice> i think maybe i will try to make it private to the terrain builder
20140518 11:57:27< Coffee_irc> iceiceice: that's a good idea
20140518 11:57:32< iceiceice> idk is it the only dependency now?
20140518 11:57:42< Coffee_irc> looks like it
20140518 12:07:52< irker252> wesnoth: David Mikos wesnoth:master ad8b0da26538 / src/map_location.cpp: Fix code comment with last commit http://git.io/cjDDhg
20140518 12:08:16-!- Sulfur [~Miranda@p5B327CD8.dip0.t-ipconnect.de] has joined #wesnoth-dev
20140518 12:18:51< irker252> wesnoth: Chris Beck wesnoth:master 517c8a8e1ee6 / src/tests/test_map_location.cpp: visualize relative dir routine http://git.io/6DfrOw
20140518 12:18:53< irker252> wesnoth: Chris Beck wesnoth:master fdcd63c534eb / src/tests/test_map_location.cpp: add sanity check for vector_negation, based on random walk. http://git.io/V0r-8w
20140518 12:18:55< irker252> wesnoth: Chris Beck wesnoth:master 741051b0b34b / src/tests/test_map_location.cpp: add sanity check for get_direction http://git.io/qGzmaQ
20140518 12:18:57< irker252> wesnoth: Chris Beck wesnoth:master 0e1963559ace / src/tests/test_map_location.cpp: update characterization test for get_relative_dir http://git.io/VXMtfQ
20140518 12:18:59< irker252> wesnoth: Chris Beck wesnoth:master 70f4ccb54511 / src/tests/test_map_location.cpp: Merge branch 'test_relative_dir' http://git.io/ILElxQ
20140518 12:19:01< irker252> wesnoth: Chris Beck wesnoth:master 14c8b4728753 / src/tests/test_map_location.cpp: disable the get_relative_dir visualizer http://git.io/qr1brQ
20140518 12:19:03< irker252> wesnoth: Chris Beck wesnoth:master 9f8dfc381965 / src/map_location.cpp: Merge branch 'master' of git://github.com/wesnoth/wesnoth http://git.io/3hlvCw
20140518 12:37:09-!- Samual [~dioteckte@xonotic/core-team/Samual] has quit [Ping timeout: 252 seconds]
20140518 12:54:12-!- stikonas__ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20140518 12:54:25-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has quit [Ping timeout: 258 seconds]
20140518 13:08:02< mattsc> iceiceice: hmm?
20140518 13:09:33< mattsc> mordante: I’m here now, but only for a little while.
20140518 13:13:49< mattsc> or Coffee_irc: ^ (two lines up)
20140518 13:23:46< mordante> mattsc, could you test whether this patch causes an assertion failure on the Mac? http://paste.debian.net/100410/
20140518 13:26:41< mattsc> mordante: sure (in about 20-30 min)
20140518 13:27:28< mordante> ok thanks
20140518 13:28:10< mattsc> mordante: 1.12 or master (or does it matter)?
20140518 13:31:16< mordante> I did it on master, but expect the same result on 1.12
20140518 13:37:24-!- horon [~horon@nttkyo210231.tkyo.nt.ngn2.ppp.infoweb.ne.jp] has joined #wesnoth-dev
20140518 13:47:31< mattsc> mordante: did you add any files recently (last 10 days or so)?  I get a linker error (before applying the patch) with a gui2 error message.
20140518 13:47:44< mattsc> iceiceice also renamed a file, but I caught that one.
20140518 13:48:12< mattsc> mordante: http://pastebin.com/zFNsXWE2
20140518 13:54:38< mordante> mattsc, I didn't add that file, I think fabi_ added it
20140518 13:55:34< mattsc> mordante: which one is “that file”?
20140518 13:55:36< mordante> mattsc,  it's in gui/dialogs/core_selection.cpp
20140518 13:55:57< mattsc> mordante: okay, thanks.  Let me see ...
20140518 13:56:11< mordante> you're welcome
20140518 13:56:40-!- kex [~kex@78.157.29.205] has joined #wesnoth-dev
20140518 13:59:23-!- kex [~kex@78.157.29.205] has quit [Remote host closed the connection]
20140518 13:59:53< mattsc> mordante: yes, thanks, that did it.  Sorry, hadn’t compiled in a while and apparently I am having some stupidity issues this morning.  :P
20140518 14:00:01-!- kex [~kex@78.157.29.205] has joined #wesnoth-dev
20140518 14:00:20< mattsc> In a moment I’ll now actually be able to test the patch.  I think…
20140518 14:01:11< irker252> wesnoth: mattsc wesnoth:master b9a86a656af6 / projectfiles/Xcode/Wesnoth.xcodeproj/project.pbxproj: Update Xcode project http://git.io/SVemrA
20140518 14:03:52< mattsc> mordante: indeed, going into the help menu triggers that assert,
20140518 14:04:12< mattsc> Well, not the main help menu, but once I am trying to look at a unit’s help.
20140518 14:04:14-!- kex [~kex@78.157.29.205] has quit [Ping timeout: 240 seconds]
20140518 14:09:01< mordante> mattsc, that confirms it indeed is an issues with a non-neutral surface
20140518 14:09:26< mordante> will look at a solution
20140518 14:09:36< mattsc> mordante: thanks!
20140518 14:09:50< mattsc> I wouldn’t even know where to start ...
20140518 14:10:56< mordante> no problem, I probably ask you test another patch
20140518 14:12:24< mattsc> Sure.  I’ll be afk more than here throughout the day, but will check for my nick when I get back.
20140518 14:16:14-!- stikonas__ is now known as stikonas
20140518 14:18:43< mordante> mattsc, which of the two asserts did trigger?
20140518 14:19:25< mattsc> mordante: oh, right, sorry.  “Assertion failed: (is_neutral(src)), function blit_surface, file /mats/misc/Wesnoth/wesnoth/src/sdl_utils.cpp, line 1919.”
20140518 14:19:45< mordante> ok the one I expected :-)
20140518 14:24:00< mordante> mattsc, a bit ugly, but I expect it does the trick http://paste.debian.net/100428/
20140518 14:29:26< mattsc> mordante: woohoo, I can see my orcs again!  :)
20140518 14:32:38< mordante> and they no longer have the blues?
20140518 14:32:53< mattsc> Nope, they look quite happy and menacing.
20140518 14:33:22< mattsc> So yes, that patch seems to be doing the trick.  Thank you.
20140518 14:33:35< mordante> you're welcome, thanks for testing
20140518 14:33:50< mordante> I'll polish it a bit and commit for 1.12 and master
20140518 14:34:59< fabi_> mattsc: yes, the file is mine, maybe I didn't change the copyright message because it isn't really very different from the file I copied, thus claiming the copyright seemed somehow wrong.
20140518 14:35:05< mattsc> Great!  Thanks.  Let me know if you want me to reply to the bug report and forum post after that, or if you’ll do it yourself.
20140518 14:35:49< mordante> mattsc, if you can tell in the forum it will be fixed in the next release then I'll handle the report
20140518 14:35:51< mattsc> fabi_: no worries.  I was only trying to find which file it was (while rushing around doing other things), so that I could get Xcode to build.
20140518 14:36:15< mattsc> mordante: will do.
20140518 14:36:25< fabi_> mattsc: Yeah, I can easily fix scons and cmake but other build systems are out of my scope.
20140518 14:36:52< mordante> thanks
20140518 14:37:38< fabi_> mattsc: Maybe some sort of mechanism could be introduced, some place where you can leave a note that build systems need some love.
20140518 14:37:39< mattsc> fabi_: that’s fine, I have no problems doing it for Xcode.  I just hadn’t built in almost two weeks, so there were lots of changes and finding new files in Xcode is semi-painful.
20140518 14:38:04< mattsc> fabi_: sure, but it doesn’t happen all that often, so I wouldn’t worry about.
20140518 14:38:31< mattsc> Actually, I could probably just look at the history for the scons and cmake build files, couldn’t I?
20140518 14:40:15-!- horon [~horon@nttkyo210231.tkyo.nt.ngn2.ppp.infoweb.ne.jp] has quit [Quit: Leaving...]
20140518 14:44:02< mattsc> fabi_: indeed: https://github.com/wesnoth/wesnoth/commits/master/src/CMakeLists.txt
20140518 14:44:40< mattsc> So there’s the mechanism already.  That made way too much sense for me to come up with that idea myself.  :P
20140518 14:45:04< fabi_> mattsc: Yes, sometimes my commit messages make a little sense, but only sometimes :-)
20140518 14:46:10< mattsc> Well, independent of the messages, I can just check what changed in the file.  :)
20140518 14:46:34< mattsc> since all I need to know if which file changed, and not why.
20140518 14:49:35-!- sachith500 [~kvirc@112.135.156.3] has joined #wesnoth-dev
20140518 15:16:18-!- DHost [~Pcy@vps.pyws.fr] has quit [Ping timeout: 240 seconds]
20140518 15:22:34< irker252> wesnoth: Chris Beck wesnoth:master f9acf4d95799 / src/ (map_location.cpp map_location.hpp terrain_builder.cpp): Move legacy map_location arithmetic to terrain_builder.cpp http://git.io/H3-M9Q
20140518 15:22:36< irker252> wesnoth: Chris Beck wesnoth:master bf1dbe2b13d4 / projectfiles/Xcode/Wesnoth.xcodeproj/project.pbxproj: Merge branch 'master' of git://github.com/wesnoth/wesnoth http://git.io/H3AYdg
20140518 15:23:07< iceiceice> Coffee_irc: ^^
20140518 15:25:09-!- sachith500 [~kvirc@112.135.156.3] has quit [Read error: Connection reset by peer]
20140518 15:26:29-!- sachith500 [~kvirc@112.135.156.3] has joined #wesnoth-dev
20140518 15:28:00-!- sachith500 [~kvirc@112.135.156.3] has quit [Read error: Connection reset by peer]
20140518 15:28:40-!- sachith500 [~kvirc@112.135.156.3] has joined #wesnoth-dev
20140518 15:32:56-!- happygrue [~happygrue@wesnoth/developer/wintermute] has joined #wesnoth-dev
20140518 15:34:49-!- prophile [~alynn@oftn/member/prophile] has joined #wesnoth-dev
20140518 15:52:43< irker252> wesnoth: Mark de Wever wesnoth:1.12 91da7fe94bd9 / changelog src/sdl_utils.cpp: Only blit neutral surfaces. http://git.io/07WURA
20140518 15:57:06-!- Duthlet [~Duthlet@wesnoth/mp-mod/Duthlet] has quit [Quit: leaving]
20140518 16:02:26< irker252> wesnoth: Mark de Wever wesnoth:master ede743d4aa5f / changelog src/sdl_utils.cpp src/sdl_utils.hpp: Only blit neutral surfaces. http://git.io/wlt4Xg
20140518 16:02:33< mordante> mattsc, ^
20140518 16:09:53< iceiceice> mattsc: yeah sorry the xcode file got messed up,
20140518 16:09:59< iceiceice> i really seriously did try to fix the project files
20140518 16:10:06< iceiceice> but that was the only one i couldnt figure out how to
20140518 16:11:19< Necrosporus> http://pastebin.com/cLjjAi7X what am I doing wrong? Nested moveto event doesn't fire even though I set fire_event=yes 
20140518 16:12:57< Necrosporus> also I want to remove overlay, not touching base terrain but it doesn't seem to work either
20140518 16:15:03-!- gfgtdf [~chatzilla@e177122105.adsl.alicedsl.de] has joined #wesnoth-dev
20140518 16:19:08< gfgtdf> iceiceice: was there a reason why fight_on_without_leader isnt modifiable via lua ?
20140518 16:19:22< iceiceice> i think just pure laziness
20140518 16:19:28< iceiceice> is it modifiable in wml?
20140518 16:19:56< iceiceice> i'm not sure if i even realized those lists are different
20140518 16:20:06< zookeeper> 1) the nested event will surely fire 2) $x2,$y2 have no meaning in a moveto event 3) you want delayed_variable_substitution=yes in the nested event but you can avoid that by simply having it not be a nested event
20140518 16:20:09< zookeeper> Necrosporus, ^
20140518 16:20:33< gfgtdf> iceiceice: i dotn knwo how to modify a side in wml at all.
20140518 16:20:41< iceiceice> i think theres a [modify_side] tag
20140518 16:21:41< gfgtdf> iceiceice: oh
20140518 16:21:51< gfgtdf> hm i think its not
20140518 16:22:07< gfgtdf> ok i'll add it then
20140518 16:22:23< Necrosporus> zookeeper, wiki says (EventWML afair) x2y2 is source of movement
20140518 16:22:44< gfgtdf> and also on 1.12 since i consider it a fixup of the fight_on_without_leader tag.
20140518 16:26:14< iceiceice> ok thx
20140518 16:26:25< iceiceice> y thats a good thing
20140518 16:28:15< zookeeper> Necrosporus, oh right, new addition
20140518 16:28:29< zookeeper> 3 applies regardless
20140518 16:28:48< Necrosporus> so it won't work in 1.10?
20140518 16:29:11< Necrosporus> I guess I have to move to 1.11 event though I do not want to
20140518 16:44:02< Necrosporus> zookeeper, but delayed_variable_substitution is said to be default to yes
20140518 16:52:11-!- vultraz_ [~chatzilla@124.109.10.167] has joined #wesnoth-dev
20140518 16:52:31-!- vultraz [~chatzilla@124.109.10.167] has quit [Read error: Connection reset by peer]
20140518 16:52:43-!- vultraz_ is now known as vultraz
20140518 16:53:33-!- vultraz [~chatzilla@124.109.10.167] has quit [Changing host]
20140518 16:53:33-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20140518 16:55:21-!- wesbot changed the topic of #wesnoth-dev to: 1.11.14 (1.12 beta 5) is too buggy to be announced | string+feature freeze active on 1.12 | 239 bugs, 344 feature requests, 28 patches | Logs: http://irclogs.wesnoth.org | Alternate logs: http://wesnoth.debian.net | Don't paste on IRC! Use a pastebin: http://pastebin.com | http://imagebin.org
20140518 16:56:17< zookeeper> Necrosporus, really? i guess i don't remember anything anymore then
20140518 16:58:37< Necrosporus> http://wiki.wesnoth.org/EventWML#Delayed_Variable_Substitution_2
20140518 17:01:14-!- ancestral [~ancestral@12.23.74.29] has joined #wesnoth-dev
20140518 17:02:43< Necrosporus> zookeeper, you might at least know how to remove overlay?
20140518 17:09:21-!- prophile [~alynn@oftn/member/prophile] has quit [Quit: The Game]
20140518 17:10:01-!- ancestral [~ancestral@12.23.74.29] has quit [Quit: i go nstuf kthxbai]
20140518 17:11:08< zookeeper> Necrosporus, well there's this in UtBS: http://pastebin.com/CWxC8Mna
20140518 17:12:32< zookeeper> obviously the [and] there for filtering reasons.
20140518 17:12:35< zookeeper> +is
20140518 17:20:01< gfgtdf> iceiceice: im thinking about also makign set_lost available to lua, "side.lost = false"
20140518 17:20:37< iceiceice> so if you do i think you should make the handler also preserve whatever invariant that thing ahs
20140518 17:20:40< iceiceice> like...
20140518 17:20:41< gfgtdf> iceiceice: but im nore sur whether "lost" is the best attributenaem since in only effects the carryover i am right
20140518 17:20:50< iceiceice> lost / defeated isk
20140518 17:20:51< iceiceice> *idk
20140518 17:21:02< iceiceice> i think it is called both ways in the C++ vs wml
20140518 17:21:23< iceiceice> will you make it so that setting it false removes the carryover / restores it?
20140518 17:22:15< gfgtdf> iceiceice: ye , at lest that was teh plan
20140518 17:23:31< gfgtdf> the 2 bugs that blocked 1.11.14 are they alredy fixed ?
20140518 17:24:11-!- sachith500 [~kvirc@112.135.156.3] has quit [Read error: Connection reset by peer]
20140518 17:28:08< iceiceice> no i have been neglecting that stuff
20140518 17:28:11< iceiceice> i should get back to it
20140518 17:28:39< iceiceice> i was very pleased with the last patch we did a few days ago
20140518 17:28:49< gfgtdf> iceiceice: whcih last patch ?
20140518 17:28:51< gfgtdf> which
20140518 17:28:52< iceiceice> i got to like scenario 9 of LoW without getting a blocker :)
20140518 17:28:54< iceiceice> the sync patch
20140518 17:29:12< gfgtdf> iceiceice: the traits of units in [side9 ?
20140518 17:29:16< iceiceice> yeah
20140518 17:29:56< iceiceice> i think its like maybe one or two more patches and probably it works completely
20140518 17:30:30-!- thunderstruck [~zaibotren@cpc13-sgyl31-2-0-cust696.18-2.cable.virginm.net] has quit [Quit: leaving]
20140518 17:31:44< irker252> wesnoth: gfgtdf wesnoth:master 7390a991d78a / src/team.hpp: add setter for team.fight_on_without_leader http://git.io/aHDJfg
20140518 17:34:38< irker252> wesnoth: gfgtdf wesnoth:master ff3b68636e24 / src/scripting/lua.cpp: expose side.lost and fight_on_without_leader to lua http://git.io/WXnNqQ
20140518 17:40:46< irker252> wesnoth: Chris Beck wesnoth:master 3501d438b034 / src/tests/test_map_location.cpp: add 3-way tests for map_location, like associativity. refactor. http://git.io/uI3ypw
20140518 17:40:48< irker252> wesnoth: Chris Beck wesnoth:master ff004167203a / src/ (scripting/lua.cpp team.hpp): Merge branch 'master' of git://github.com/wesnoth/wesnoth http://git.io/NK8DIA
20140518 17:41:41< iceiceice> Coffee_irc: i decided also to add some 3 way checks to the unit test thing, so we know that vector addition is associative like that comment claims (it was only a few lines)
20140518 17:41:50< iceiceice> and i also added a check that get_relative_dir is translation invariant
20140518 17:41:57< iceiceice> so dir(a,b) = dir(a+c, b+c)
20140518 17:42:03< iceiceice> but that one doesnt seem to hold atm :/
20140518 17:42:35< iceiceice> i think it means that if we drew the picture centered at places other than 7,8, (8,8), it could look different...
20140518 17:43:07< iceiceice> i dont think i want to think about it for a while though
20140518 17:43:34< iceiceice> my feeling is that it should be possible to translate to a good spot and then run your code if we want
20140518 17:47:44-!- kex [~kex@89.205.75.19] has joined #wesnoth-dev
20140518 18:02:34-!- Sulfur [~Miranda@p5B327CD8.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20140518 18:14:48< irker252> wesnoth: gfgtdf wesnoth:1.12 bbb2746528d3 / src/team.hpp: add setter for team.fight_on_without_leader http://git.io/VOyyZg
20140518 18:16:43< irker252> wesnoth: gfgtdf wesnoth:1.12 228b2afa880a / src/scripting/lua.cpp: expose side.lost and fight_on_without_leader to lua http://git.io/WK1HTg
20140518 18:20:35-!- ancestral [~ancestral@12.23.74.29] has joined #wesnoth-dev
20140518 18:24:38< iceiceice> oh btw, gfgtdf did you see slowthinker's last comment here?
20140518 18:24:38< iceiceice> https://gna.org/bugs/index.php?21800
20140518 18:25:01< iceiceice> he thinks that something is wrong in the integrity of debug mode in mp i guess, like it changes the teams of the sides or something
20140518 18:28:03-!- ancestral [~ancestral@12.23.74.29] has quit [Quit: Smell ya later!]
20140518 18:28:04< gfgtdf> iceiceice: to me it looks lie all this takes place befroe the game is crates maing in the mo connect & co files.
20140518 18:28:21< iceiceice> y...
20140518 18:28:30< iceiceice> but if its true it means i should have been testing with debug mode i guess
20140518 18:28:58< gfgtdf> iceiceice: you want ot say you tested an you got different results ?
20140518 18:29:11< iceiceice> idk i havent confirmed the bug
20140518 18:29:41< iceiceice> it is pretty minor though
20140518 18:30:07< iceiceice> i guess i imagine there could be a side like that in a campaign, if the side is populated during an event at some later turn or smth?
20140518 18:31:39< gfgtdf> iceiceice: i think it is possibe to activate a null controlled side per wml
20140518 18:32:36< gfgtdf> makign controller=ai with modify_side
20140518 18:41:28< iceiceice> yi havent tried to reproduce it, its on my todo list
20140518 18:41:34< iceiceice> s/yi/y i
20140518 18:41:35< mattsc> iceiceice: no worries, you can’t do anything about the Xcode project file without acutally using Xcode (we’ve tried) and it’s easy to do
20140518 18:42:32< gfgtdf> iceiceice: i just testes playing a scenario with empty controller in debug mode and it wors liek expected
20140518 18:42:36< gfgtdf> tested
20140518 18:42:40< gfgtdf> 1.13-dev
20140518 18:42:53< iceiceice> even with the teams thing like he said?
20140518 18:43:02< iceiceice> w/e i guess its a nonissue then
20140518 18:43:46< gfgtdf> iceiceice: which teams thing ?
20140518 18:44:19-!- ancestral [~ancestral@12.23.74.29] has joined #wesnoth-dev
20140518 18:44:35< mattsc> mordante: thanks!  I confirmed that it works in both 1.12 and master.  (Not that I doubted it, but better safe than sorry.)
20140518 18:44:40< iceiceice> "But I still see a little bug there: The team of the side with
20140518 18:44:40< iceiceice> 'allow_player=no'/'controller=null' is automatically changed to the side1's
20140518 18:44:40< iceiceice> team, and maybe there are more changes in the side's properties."
20140518 18:46:30< gfgtdf> iceiceice: i testes with caves of basiliks whcih have a side with 'allow_player=no' and 'controller=null'. The side showed up normal in teh mp connect dialog as "empty"
20140518 18:48:52< gfgtdf> iceiceice: but i dnt realy knwo what he means with 'team' and 'side'
20140518 18:48:54< gfgtdf> don't
20140518 18:48:57< iceiceice> so i see the bug
20140518 18:48:59< iceiceice> in mp connect
20140518 18:49:04< iceiceice> under Team/Color
20140518 18:49:13< iceiceice> see 1 is North
20140518 18:49:15< iceiceice> 2 is South
20140518 18:49:17< iceiceice> 3 is North again
20140518 18:49:25< iceiceice> idk if that means 1 will have vision of the statues?
20140518 18:51:16< iceiceice> hmmm ok
20140518 18:51:19< iceiceice> i made two saves,
20140518 18:51:21< iceiceice> one in debug mode and one not
20140518 18:51:27< iceiceice> start of scenario, same scenario you talked about
20140518 18:51:37< iceiceice> even though mp_connect says side 3 will be team North,
20140518 18:51:39< iceiceice> its not true,
20140518 18:51:43< iceiceice> it ends up as team=statues
20140518 18:52:09< iceiceice> so i guess its a minor glitch or smth
20140518 18:52:36< iceiceice> maybe the mp connect doesnt know to look at all the team strings when populating its list when debug mode is on?
20140518 18:53:01< iceiceice> but i'm not worried about it messing up the testing
20140518 18:53:21< iceiceice> at least i havent seen anything to suggest that
20140518 18:54:37< gfgtdf> iceiceice: ye it uses player_teams https://github.com/wesnoth/wesnoth/blob/46850f271a3a6c4fb2025aa50c70b5c59ecbce35/src/multiplayer_connect.cpp#L76 whcih is all temas whose sides don't have allow_player=no
20140518 18:56:57< gfgtdf> iceiceice: but teh result miht be that as soon as you change it in debug mode you cannot change it back
20140518 18:57:00< gfgtdf> the
20140518 18:57:59< iceiceice> maybe the engine_.player_teams() should take into account if its debug mode
20140518 18:59:45< iceiceice> maybe add a check here: https://github.com/wesnoth/wesnoth/blob/46850f271a3a6c4fb2025aa50c70b5c59ecbce35/src/multiplayer_connect_engine.cpp#L154
20140518 18:59:51< iceiceice> "|| game_config::debug" 
20140518 19:00:04< iceiceice> or whatever the appropriate thing is
20140518 19:02:45< gfgtdf> iceiceice: uhm i dont realy know much about the mp connect code sry.
20140518 19:03:38< iceiceice> i might commit something like that i think in a bit
20140518 19:03:47< iceiceice> will test though
20140518 19:07:29< gfgtdf> iceiceice: did i recently saw a commit you yours makign bad_lexical_cast  devide from std::exception or am a confused ?
20140518 19:08:37< iceiceice> i did that
20140518 19:09:10< iceiceice> https://github.com/wesnoth/wesnoth/commit/49dc82b5d8f2e3f07258247d700458baa7c73902
20140518 19:10:35< gfgtdf> iceiceice: seems we have 2 times taht calss https://github.com/wesnoth/wesnoth/blob/1.12/src/util.hpp#L76,  https://github.com/wesnoth/wesnoth/blob/49dc82b5d8f2e3f07258247d700458baa7c73902/src/lexical_cast.hpp#L98
20140518 19:11:18< gfgtdf> iceiceice: that is very confusing and shoudl be fixed i think
20140518 19:11:23< iceiceice> :O
20140518 19:12:27< iceiceice> y i think util.cpp should include lexical_cast.hpp
20140518 19:12:34< iceiceice> and maybe util.hpp shoudl just have a stub?
20140518 19:12:36< iceiceice> does that make sense?
20140518 19:13:30< gfgtdf> iceiceice: https://github.com/wesnoth/wesnoth/commit/a0a25fb48f99a204d0eb1717e38b1db2f25af39c
20140518 19:13:52< gfgtdf> iceiceice: seems like mordante wanted to replace teh lexical cast
20140518 19:14:28< gfgtdf> maybe mordante knows why he didnt completed this.
20140518 19:15:36< shadowm> mordante: What about the neutral surface thing mattsc/Alarantalara mentioned?
20140518 19:15:38< iceiceice> y i had no idea lexical_cast.hpp wasnt actually used
20140518 19:15:52< iceiceice> is that message still accurate?
20140518 19:16:01< iceiceice> maybe it was completed... ?
20140518 19:16:28< mattsc> shadowm: https://github.com/wesnoth/wesnoth/commit/ede743d4aa5f8397e1bc24f41941f354ba78155f
20140518 19:17:32< gfgtdf> iceiceice: i found no includes of leical_cast.hpp
20140518 19:17:53< mattsc> I need to be off.  Bye for now.
20140518 19:18:27< shadowm> mattsc: The implementation and declaration of the function now have mismatched parameter names then. :p
20140518 19:18:54< shadowm> So did it work?
20140518 19:18:57-!- shadowm_desktop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20140518 19:19:04< mattsc> shadowm: yes, tested on both 1.12 and master.
20140518 19:19:22< mattsc> okay, really got to go not though.  BBL.
20140518 19:19:28-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: Ciao]
20140518 19:58:15< mordante> thanks mattsc
20140518 19:59:39< mordante> gfgtdf, guess I forgot about it
20140518 20:02:16< mordante> I'm off bye
20140518 20:04:54-!- mordante [~mordante@wesnoth/developer/mordante] has quit [Quit: Leaving]
20140518 20:11:40< gfgtdf> iceiceice: you said "Killing all the units is a fairly common thing to do in WML it seems" where did you know this ?
20140518 20:12:03-!- aquileia [2edf524c@gateway/web/freenode/ip.46.223.82.76] has joined #wesnoth-dev
20140518 20:14:19< aquileia> Anyone with a sound knowledge of khalifate timeline here? http://wiki.wesnoth.org/Timeline_of_Wesnoth has no info at all, and http://r.wesnoth.org/p503997 isn't too precise
20140518 20:14:32< aquileia> wesbot: seen noy
20140518 20:14:32< wesbot> aquileia: The person with the nick noy last spoke 43d 3h ago.  2d 18h ago was here and on the channels #wesnoth and #wesnoth-umc-dev with the message: Quit: noy
20140518 20:14:54< aquileia> happygrue: ^
20140518 20:15:15< happygrue> aquileia: that is a question for noy
20140518 20:15:47< happygrue> I do the balancing but none of the art or fluff, and I have not seen the draft of their history he has
20140518 20:15:59< aquileia> I have to sync them with http://wiki.wesnoth.org/Garardine_Dynasty for our campaign
20140518 20:16:10< happygrue> though I hope it comes to the development branch soon :D
20140518 20:16:31< happygrue> I would email or send him a forum PM
20140518 20:16:34-!- iceiceice [~chris@cpe-66-108-20-80.nyc.res.rr.com] has quit [Ping timeout: 240 seconds]
20140518 20:18:37< aquileia> The issue is: We need a Wesnothian war of succession to place the campaign at, and (except for HttT) the time around 500 YW is far too stable
20140518 20:18:51-!- shadowm_desktop [ignacio@wesnoth/developer/shadowmaster] has quit [Ping timeout: 265 seconds]
20140518 20:19:19< aquileia> It's Asheviere, not Ashviere, right?
20140518 20:19:24< happygrue> honestly, I don't know anything about the actual timeline, I just click the units.  :D
20140518 20:20:05< happygrue> I am 98% focused on gameplay, with 2% leftover for fluff when it doesn't harm gameplay at all.  :D
20140518 20:20:18< aquileia> Hmm... she could have hired khalifate mercenaries... that might fit perfectly...
20140518 20:20:25< aquileia> happygrue: Thanks nontheless
20140518 20:20:27-!- ancestral [~ancestral@12.23.74.29] has quit [Quit: i go nstuf kthxbai]
20140518 20:20:27< happygrue> could be
20140518 20:20:45< happygrue> you can check the unit descriptions of the units at units.wesnoth.org
20140518 20:20:54< aquileia> No chance to get them into the HttT overhaul? =P
20140518 20:20:56< happygrue> the level 1s are pretty good, the rest will probably change
20140518 20:21:03< happygrue> maybe
20140518 20:21:20< happygrue> eventually they will get worked into varios things... but we need the history first!
20140518 20:21:33< happygrue> *various
20140518 20:21:41< aquileia> Because "The March of the ten thousand" would be the story of those khalifate mercs
20140518 20:22:26< happygrue> It could fit really well, depending on how exactly they got to Wesnoth - but IDK
20140518 20:23:00< aquileia> Though I guess I'll need a different dispute to match the first scenario
20140518 20:23:28-!- ancestral [~ancestral@12.23.74.29] has joined #wesnoth-dev
20140518 20:26:00< zookeeper> well, either you need the history first so you can base your campaign on it, or you can create the history with your campaign. i mean sure, if you can get more than that 3 years old draft from noy then you should base it on that, but if no one fills in the gaps for you then surely it's fine if you do so.
20140518 20:26:48< aquileia> zookeeper: Ok, nice
20140518 20:27:13< aquileia> Would tiny adjustments to the existing timeline be acceptable?
20140518 20:27:24< happygrue> I would agree, except that so far when someone has taken the time to write something about them it tends to get rewritten by noy, which has been frustrating for some.
20140518 20:27:35< happygrue> though also a good catylist for actually getting stuff written...
20140518 20:27:46< zookeeper> aquileia, well if they don't interfere with anything important then it's probably ok
20140518 20:28:03< zookeeper> can't really give any general guidelines for that
20140518 20:28:19< aquileia> zookeeper: HttT... well, not THAT important =P
20140518 20:28:39< aquileia> It would spice it up!
20140518 20:28:50< happygrue> I agree, surely some changes to the timeline will probably have to be made as they get worked in, but I think it would be best to have a talk with noy before trying that.  Perhaps get a list of all the changes and throw it out to the ML in one go?
20140518 20:28:57< happygrue> or maybe people don't care about it that much
20140518 20:29:04< happygrue> I sure don't, but some people probably do?
20140518 20:29:55< happygrue> or as zookeeper says, maybe if you just do it, then it will become the offical history ;)
20140518 20:30:04< aquileia> Then I'll try to get our plans in sync with the timeline
20140518 20:30:16< happygrue> seeing no movement on it for some time now I would feel fine if you did that.
20140518 20:31:59< aquileia> BTW, I noticed that http://wiki.wesnoth.org/Garardine_Dynasty spells Asheviere wrongly
20140518 20:32:37< zookeeper> it does, feel free to fix if you want :>
20140518 20:33:00< aquileia> well, it's in the image as well
20140518 20:33:20< aquileia> I'll modify the text
20140518 20:33:33< zookeeper> oh right, so it is
20140518 20:33:34< zookeeper> how silly
20140518 20:36:15-!- [Relic] [~relic@99-58-54-211.lightspeed.milwwi.sbcglobal.net] has joined #wesnoth-dev
20140518 20:39:35-!- ancestral [~ancestral@12.23.74.29] has quit [Quit: i go nstuf kthxbai]
20140518 20:42:03-!- cib_ [~cib@p5DC75E1A.dip0.t-ipconnect.de] has quit [Ping timeout: 265 seconds]
20140518 20:46:15< irker252> wesnoth: gfgtdf wesnoth:master 71cf3e96605d / src/util.hpp: devide bad_lexical_cast from std::exception http://git.io/aFeMRA
20140518 20:46:48-!- shadowm_desktop [ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20140518 20:47:14< shadowm>  Derive?
20140518 20:48:12< gfgtdf> ye i noticed too
20140518 20:48:18< gfgtdf> it shodul bee derive
20140518 20:48:21< gfgtdf> i think
20140518 20:51:10< irker252> wesnoth: gfgtdf wesnoth:1.12 2aebad5dac3a / src/util.hpp: derive bad_lexical_cast from std::exception http://git.io/1Tpfqw
20140518 20:52:42< gfgtdf> shadowm: i corrected it on the 1.12 branch :)
20140518 20:53:50< shadowm> Which reminds me I need to go over that addition to the Lua documentation and fix a few things.
20140518 20:56:55< gfgtdf> shadowm: my addition ?
20140518 20:57:15< shadowm> Yes.
20140518 20:59:12-!- ancestral [~ancestral@12.23.74.29] has joined #wesnoth-dev
20140518 20:59:12-!- iceiceice [~chris@cpe-66-108-20-80.nyc.res.rr.com] has joined #wesnoth-dev
20140518 21:00:28< gfgtdf> shadowm: hm i really tried to use less typos for that.
20140518 21:01:18< iceiceice> gfgtdf: idk, i guess it happens in conquest for example when you have to set up the scenario...
20140518 21:01:24< iceiceice> i think it happens frequently in cutscenes
20140518 21:01:51< iceiceice> i think dugi often makes all the units exit a scenario dramatically (disappearing)
20140518 21:01:53< gfgtdf> iceiceice: well al song as you place units in teh same evnt is shoudl be no problme
20140518 21:01:54< iceiceice> shadowm mentioned
20140518 21:02:01< gfgtdf> hm ok
20140518 21:02:21< iceiceice> i had a branch
20140518 21:02:28< iceiceice> to make a check_victory blocker for lua
20140518 21:02:36< iceiceice> but i didnt test it yet
20140518 21:02:36< shadowm> I mentioned what?
20140518 21:02:44< irker252> wesnoth: gfgtdf wesnoth:master aed6d9482015 / src/synced_context.cpp: use use_undo=false for [do_command] http://git.io/J-1Urw
20140518 21:03:00< iceiceice> shadowm: i remember the last time we talked about this you were like ... hmm do i kil all the units in AtS
20140518 21:03:07< iceiceice> and then you "unpaniced" or somethign
20140518 21:03:25< shadowm> Yeah, I kill all units repeatedly for story reasons.
20140518 21:04:04< gfgtdf> iceiceice: maybe we could replace the "fight_on_without_leader" with, defeat_condition=no_leader(default)/no_units/never
20140518 21:05:00< iceiceice> why do we want to change it? i think it serves a purpose as it is
20140518 21:05:15< iceiceice> oh i see
20140518 21:05:29< iceiceice> yeah actually i kind of like your idea
20140518 21:05:49< gfgtdf> :)
20140518 21:06:04< iceiceice> hmm so
20140518 21:06:17< iceiceice> if the side is not defeated
20140518 21:06:25< iceiceice> then i guess it wont be removed from carryover
20140518 21:06:45< iceiceice> i'm trying to decide if, assuming we did this, if there woudl still be a use case for skip_check_victory
20140518 21:06:47< iceiceice> i think probably not
20140518 21:07:28< iceiceice> y this is probably a very good idea
20140518 21:08:34< gfgtdf> iceiceice: we currenty distinguish between lost for carryover and defeated for losing a scenario, but i this is somehow confusin 
20140518 21:11:01< iceiceice> i agree
20140518 21:11:14< iceiceice> the lost for carryover thing is new in 1.11.7 by the way
20140518 21:11:19-!- prophile [~alynn@oftn/member/prophile] has joined #wesnoth-dev
20140518 21:11:41< iceiceice> i think the idea was that we wanted to make sure that if a side was defeated, adn then we saved and reloaded, that the carryover wouldnt get corrupted
20140518 21:11:50< iceiceice> but i can point you to a commit where that was introduced
20140518 21:12:19< gfgtdf> iceiceice: you mean reloadign durignlingremode ?
20140518 21:13:08< iceiceice> ok, i can usually figure out what the words mean but i really have no idea what that one is "durignlingremode"
20140518 21:13:47-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 252 seconds]
20140518 21:14:06< gfgtdf> during linger mode
20140518 21:15:01< gfgtdf> iceiceice: ^
20140518 21:15:12< iceiceice> ah
20140518 21:15:36< iceiceice> no i wasnt thinking of that actually
20140518 21:15:42< iceiceice> i guess that from my reading of the commit,
20140518 21:15:51< iceiceice> hmm...
20140518 21:15:58< iceiceice> i'm not sure if i remember anymore how it works tbh
20140518 21:17:43< iceiceice> https://github.com/wesnoth/wesnoth/commit/fe3224e9b428ab72a834fd22fc2443c159d11480
20140518 21:18:19< iceiceice> that's where "lost" came from
20140518 21:18:36< iceiceice> so i guess the idea is that actually,
20140518 21:18:47< iceiceice> we maintain a correct database of carryover_sides info ata ll time
20140518 21:18:56< iceiceice> so there is carryover_sides_start from the beginning of the level,
20140518 21:19:06< iceiceice> but we also have an intermediate thing maintained as we go
20140518 21:19:16< iceiceice> at least thats what i took away from this
20140518 21:19:25< iceiceice> presmuably the reason is in case of a middle of scenario save
20140518 21:19:54< iceiceice> i dont know all the details about this though
20140518 21:20:45< gfgtdf> iceiceice: to me it looks like the sided were never removed from the prsistent die list beofre this commit ?
20140518 21:22:09< gfgtdf> iceiceice: with the currentl implementation the wml dev is able to choose which sided shoudl persist dnamicly settign side.lost = false/lua from lua in teh victory event
20140518 21:22:29< gfgtdf> al least that was my plan when expsing exposing it to lua
20140518 21:22:42< iceiceice> yeah that sounds very good
20140518 21:23:11< iceiceice> but i dont know what assumptions are required to make it safe to save at various points
20140518 21:23:33< iceiceice> so i guess... also maybe risky for umc to mess around with it?
20140518 21:23:58< iceiceice> it seems like if the reason we are doing this is to make sure we keep saves correct,
20140518 21:24:08< iceiceice> then maybe we shouldnt remove anything from carryover until the end
20140518 21:24:17< iceiceice> and just whenever we save, check the side.lost flags
20140518 21:24:38< gfgtdf> iceiceice: hm i thought we onyl rmeove things from carryover at the end ?
20140518 21:24:54< iceiceice> so since that commit,
20140518 21:25:03< iceiceice> we remove a side from carryover as soon as it is defeated
20140518 21:25:15< gfgtdf> iceiceice: no we set lost at soon as its defeated
20140518 21:25:27< gfgtdf> then at teh end of teh scenario we check teh lost variable
20140518 21:25:29< gfgtdf> the
20140518 21:25:39< gfgtdf> at least thats what i thought
20140518 21:26:04< iceiceice> y you are right...
20140518 21:26:20< iceiceice> let me check the current state of the code, idk why im confused
20140518 21:31:54< gfgtdf> shadowm: did you already changed somethign in the wiki ?
20140518 21:32:22< shadowm> I'm preoccupied with other matters at the moment, so I have not.
20140518 21:33:20-!- trewe [~trewe@2001:8a0:d102:fb01:626c:66ff:fe92:9b7c] has joined #wesnoth-dev
20140518 21:33:23< gfgtdf> shadowm: hm ok im just curious
20140518 21:34:24< iceiceice> hmm ok first of all i can no longer quite follow the logic in that function... not a good sign :(
20140518 21:34:33< iceiceice> it looks like in my commit
20140518 21:34:39< iceiceice> https://github.com/wesnoth/wesnoth/blob/master/src/play_controller.cpp#L1426
20140518 21:34:58< iceiceice> i changed the test "if (!tm->no_leader() && ...)"
20140518 21:35:09< iceiceice> to "if(!tm->fight_on... && ...)"
20140518 21:35:28< iceiceice> but that doesnt actually make sense, fight_on was already incorporated earlier
20140518 21:36:25< iceiceice> tbh i dont really understand now why the no_leader check was there in the first place... we already know there was no leader or it would have been seen in the "seen_leaders" / "not_defeated" loop
20140518 21:36:58< iceiceice> in current implementation it looks like a side with "fight_on ... = true" can never have tm.set_lost() occur
20140518 21:37:02< iceiceice> and that was not the intention
20140518 21:37:39-!- mjs-de [~mjs-de@f048215213.adsl.alicedsl.de] has quit [Remote host closed the connection]
20140518 21:39:20< gfgtdf> iceiceice: y  it looks liek it should just be removed. the fight_on_without_leader in that line
20140518 21:39:51< iceiceice> y
20140518 21:39:54< gfgtdf> iceiceice: but tehen teh vairable remove_from_carryover_on_leaders_loss is named wrngly
20140518 21:39:57< gfgtdf> then the
20140518 21:40:13< gfgtdf> ot shold rather be remove_from_carryover_on_defeat
20140518 21:40:16< gfgtdf> should*
20140518 21:40:46< iceiceice> y i mean that thing was named before fight_on existed
20140518 21:47:58-!- Gambit [~derek@wesnoth/developer/grickit] has quit [Ping timeout: 240 seconds]
20140518 21:50:09< irker252> wesnoth: gfgtdf wesnoth:master 1fc27bb9e7ce / src/scripting/lua.cpp: replace wesnoth.is_synced() with wesnoth.current.synced_state http://git.io/rWCmdQ
20140518 21:52:35-!- Gambit [~derek@wesnoth/developer/grickit] has joined #wesnoth-dev
20140518 21:52:46-!- gfgtdf_ [~chatzilla@f054150090.adsl.alicedsl.de] has joined #wesnoth-dev
20140518 21:54:43-!- gfgtdf [~chatzilla@e177122105.adsl.alicedsl.de] has quit [Ping timeout: 240 seconds]
20140518 21:54:57-!- gfgtdf_ is now known as gfgtdf
20140518 21:56:01< gfgtdf> iceiceice: should i try to implement defeat_condition ?
20140518 21:56:40< iceiceice> y sure 
20140518 21:57:17< gfgtdf> iceiceice: do you knwo why we have most value in a team_info class and not inside the team class ?
20140518 21:57:18< iceiceice> we should maybe rename remove_from_carryover_on_leaders_loss also i guess as you mentioned
20140518 21:57:26< iceiceice> gfgtdf: no
20140518 21:59:29< Coffee_irc> iceiceice: the translation invariance can't be true
20140518 21:59:56< Coffee_irc> iceiceice: for the picture we generated and get_relative_dir
20140518 22:00:43< Coffee_irc> because hexes that are on odd columns have north east at a different relative location (on square coordinates) to even column hexes
20140518 22:00:52< iceiceice> y
20140518 22:01:22< Coffee_irc> iceiceice: the reason why it doesn't hold is the conversion from square coordinates to hex coordinates
20140518 22:01:46< iceiceice> hmmm
20140518 22:02:11< iceiceice> no but that's no what i mean by translation invariance
20140518 22:02:12-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20140518 22:02:19< iceiceice> if i have a.get_dir(b)
20140518 22:02:26< Coffee_irc> mmm
20140518 22:02:29< iceiceice> then i do, a+(northeast x 1)
20140518 22:02:32< iceiceice> and b+(northeast x 1)
20140518 22:02:36< iceiceice> the dir should be the same
20140518 22:02:39< iceiceice> i just shifted the hex grid
20140518 22:02:49< Coffee_irc> no
20140518 22:03:02< Coffee_irc> if you did by 2 or an even number it would work
20140518 22:03:25< iceiceice> ok i think we must be disagreeing what + means
20140518 22:04:44< iceiceice> ok so for the sake of discussion lets say the even number hexes are displayed one half step lower than the odd number
20140518 22:04:53< iceiceice> so its like working in C++ coordinates not WML coordinates
20140518 22:04:56< Coffee_irc> iceiceice: I'm happy to look at creating some more pictures after work today if needed :)
20140518 22:05:06< iceiceice> hehe ok
20140518 22:05:18< iceiceice> but seriously, if we dont have translation invariance, theres not much point in making the pciture
20140518 22:05:25< iceiceice> because it doesnt tell us what is happening at the other points then
20140518 22:05:37< Coffee_irc> iceiceice: ok
20140518 22:05:46< iceiceice> all i mean is 
20140518 22:05:51< iceiceice> if i hard coded 5 5 instea dof 8 8
20140518 22:05:53< iceiceice> i should get the same picture
20140518 22:05:57< Coffee_irc> I suppose I only looked at the 2 pictures centred on (7,8) and (8,8)
20140518 22:06:07< iceiceice> y i guess i assumed that we would end up with tranlsation invariance
20140518 22:06:23< iceiceice> and most likely i think we could like, auto enforce translation invariance by shifting to (8,8) or smth :)
20140518 22:06:31< iceiceice> and then refactor to make it not insane later
20140518 22:06:36< Coffee_irc> well..
20140518 22:06:46< Coffee_irc> it should work for all centres
20140518 22:06:50< iceiceice> yeah
20140518 22:06:57< Coffee_irc> but some hexes could be either one direction or another
20140518 22:06:58< iceiceice> so the thing i'm worired now though is
20140518 22:07:05< iceiceice> i think that the vector addition things are off
20140518 22:07:08< iceiceice> or may be off
20140518 22:07:10< iceiceice> like
20140518 22:07:13< Coffee_irc> some are inbetween and we need to discount them from the analysis
20140518 22:07:18< iceiceice> i think they may have more or less baked into the code that
20140518 22:07:22< iceiceice> odd is higher than even
20140518 22:07:29< iceiceice> but that somewhere when we got rid of border = 1,
20140518 22:07:33< iceiceice> it becmae the other way
20140518 22:07:34< iceiceice> or smth
20140518 22:07:43< iceiceice> idk im not sure but im suspecting something like that
20140518 22:07:49< iceiceice> didnt finish investigating though
20140518 22:08:19< iceiceice> i think that if we had a problem liek that it wouldnt cause the other tests to fail,
20140518 22:08:25< iceiceice> it should be commutative assocaitive etc. either way
20140518 22:08:54< Coffee_irc> iceiceice: and it isn't for hexes that are clearly only one possible direction?
20140518 22:09:14< iceiceice> so i guess i check right now that
20140518 22:09:19< iceiceice> hm what do i check
20140518 22:09:23< iceiceice> i think i have a check right now that
20140518 22:09:31< Coffee_irc> I think it should be good at the moment
20140518 22:09:53< iceiceice> L.get_direction(north) == L.vector_sum( map_location(0,0).get_direction(north))
20140518 22:10:07< iceiceice> and similar...
20140518 22:10:33< iceiceice> idk i think the hex grid makes sense either way
20140518 22:10:43< iceiceice> no matter what columns you say are even or odd theres always two possibilities
20140518 22:10:49< iceiceice> its only when you display it that you get a possible issue
20140518 22:10:53< Coffee_irc> iceiceice: should always look like http://imgur.com/PMIKaLq
20140518 22:11:10< Coffee_irc> the gaps could go to either direction they border
20140518 22:11:15< iceiceice> y
20140518 22:11:23< Coffee_irc> as long as this works for all centres it is good
20140518 22:11:24< iceiceice> that should be the goal
20140518 22:11:37< Coffee_irc> translation invariance will not be kept
20140518 22:11:47< Coffee_irc> because of the ones that fall inbetween
20140518 22:12:21< iceiceice> hmm
20140518 22:12:46< iceiceice> ok
20140518 22:12:58< iceiceice> but still like,
20140518 22:13:08< iceiceice> basically every procedure works by getting the "diff"
20140518 22:13:21< iceiceice> movign the source vector to (1,1)
20140518 22:13:44-!- Netsplit *.net <-> *.split quits: gfgtdf, higgins, ancestral, crimson_penguin, Upth, happygrue, irker252, cjhopman, DDR, shadowm_desktop,  (+6 more, use /NETSPLIT to show all of them)
20140518 22:13:56< iceiceice> if that translation process is correct then it shoudlnt matter
20140518 22:13:56< iceiceice> and we should get translation invariance it hink
20140518 22:14:07< Coffee_irc> iceiceice: I don't think it is necessary
20140518 22:14:12< iceiceice> y ok
20140518 22:14:24< iceiceice> i guess could try to exclude those points somehow from the test
20140518 22:14:41< Coffee_irc> at any point the relative irection will be okay as it is, but warrants further investigation
20140518 22:14:45< Coffee_irc> *direction
20140518 22:16:03-!- trewe [~trewe@2001:8a0:d102:fb01:626c:66ff:fe92:9b7c] has quit [Ping timeout: 245 seconds]
20140518 22:19:36-!- Netsplit over, joins: shadowm_desktop, vultraz, happygrue, c74d, gfgtdf, prophile, ancestral, [Relic], irker252, timotei (+6 more)
20140518 22:22:35-!- _8680_ [~8680@2002:4404:712c:0:f56f:3bff:ed67:3f9d] has quit [Ping timeout: 265 seconds]
20140518 22:23:03-!- _8680_ [~8680@2002:4404:712c:0:25f2:c523:415c:67cc] has joined #wesnoth-dev
20140518 22:23:04< iceiceice> y idk where else the vector ops are used...
20140518 22:23:15< iceiceice> i grepped for it but i think its mostly for the editor afaict
20140518 22:23:25< mattsc> Did somebody change recently how images are cleared/updated in Lua when wesnoth.put_unit() is used?
20140518 22:23:55-!- ancestral [~ancestral@12.23.74.29] has quit [Quit: End Transmission.]
20140518 22:24:01< mattsc> I am suddenly getting all kinds of “ghost images” with the AIs that do things like that.  (master only, not 1.12)
20140518 22:25:24< mattsc> Oh, does this have to do with the MOVE_UNIT bug fix, maybe?
20140518 22:25:40< mattsc> In any case, it is extremely disturbing to watch.
20140518 22:26:40< AI0867> mattsc: er, I thought that was something I fixed a few days ago
20140518 22:26:49< AI0867> it's why 1.11.14 wasn't released
20140518 22:27:04< mattsc> AI0867: yeah, actually, I cannot be that because it only happens in master for me.
20140518 22:28:16< mattsc> What’s going on is that the AIs need to place units on the map to evaluate things like: “where could the enemies get if I put my units here”  (there are many other reasons too, but that’s one of the most obvious one)
20140518 22:28:35< mattsc> Previously, those units never showed up.  Now they do, apparently, at least some of them.
20140518 22:28:40< AI0867> ugh
20140518 22:28:48< mattsc> agreed
20140518 22:29:12< AI0867> did this show up in the last week? My bugfixes were 6 days ago
20140518 22:29:16< mattsc> Alarantalara’s recruiting algorithm does the same thing.
20140518 22:29:37< mattsc> Well, I saw it for the first time today, but I had not compiled master in 10 days.
20140518 22:30:04< mattsc> But again, it’s master only, not 1.12.  So it can’t be that (at least not by itself)
20140518 22:30:22-!- trewe [~trewe@2001:8a0:d132:ed01:626c:66ff:fe92:9b7c] has joined #wesnoth-dev
20140518 22:30:23< mattsc> I guess I need to do another bisecting …  :P
20140518 22:31:59< mattsc> Uh, actually, it’s much worth than that.
20140518 22:32:12< mattsc> Units aren’t moving in a straight line at all, but jump all over the place suddenly.
20140518 22:34:11< iceiceice> whoa holy crap
20140518 22:34:19< iceiceice> something happened to my repo that does not look good 
20140518 22:34:47< mattsc> Whoa, this is BAD !
20140518 22:35:20< iceiceice> the list of files that has been deleted is so long i can no longer see what my git status is :(
20140518 22:37:05< mattsc> Could somebody with a recently compiled master branch start HttT, and just end the turn on the first two moves and watch the AI units move, please?
20140518 22:37:20< mattsc> On normal speed.  And tell me if that all looks normal?
20140518 22:37:38< mattsc> iceiceice: :(
20140518 22:37:44< iceiceice> i think i figured it out...
20140518 22:37:53< gfgtdf> :)
20140518 22:37:54< iceiceice> i tried to git checkout 1.12 but i typed git checkout 1.1
20140518 22:37:58< iceiceice> and i hit ctrl-c
20140518 22:38:00< iceiceice> and it seeme dto be ok
20140518 22:38:05< iceiceice> but actually i think it got into a wierd spot
20140518 22:38:22< iceiceice> :( i didnt realize that kind of thing could be unsafe 
20140518 22:38:26-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20140518 22:38:29< iceiceice> i wonder if my repo is corrupted now
20140518 22:38:46< gfgtdf> iceiceice: i thougth you figured out masscs bug
20140518 22:38:53< iceiceice> oh no not that
20140518 22:38:54< mattsc> gfgtdf: you have a recently compiled master, I presume?
20140518 22:39:00< iceiceice> i'm still just triyng to get my repo
20140518 22:39:13< gfgtdf> mattsc: my master is curently compiling coudl take a whiel :)
20140518 22:39:40< iceiceice> wow this is super annoying i have like 1000 untracked files in my root now
20140518 22:39:46< mattsc> Ah, ok.  I’d like to know whether this is a common problem, so if somebody could test it, that would be nice.
20140518 22:39:49< iceiceice> MANUAL.catalan...
20140518 22:40:22< mattsc> It is both very amusing and disturbing.  :P
20140518 22:40:44< AI0867> iceiceice: you can get the working copy (and even the index) into weird states, but the actual repository isn't touched by these operations
20140518 22:41:19< iceiceice> its really annoying though...
20140518 22:41:29< iceiceice> i guess i have to just remote the whole working direcotyr and reset hard?
20140518 22:41:44< iceiceice>  i have like 1000 old image and data files
20140518 22:41:46< AI0867> iceiceice: git checkout master && git reset --hard && (git status --porcelain|awk '/^??/{ print $2 }'|xargs rm
20140518 22:41:48< iceiceice> that are untracked
20140518 22:41:51< AI0867> )
20140518 22:42:15< AI0867> though files with spaces could screw that up
20140518 22:42:21< iceiceice> and directories apparently
20140518 22:42:28< iceiceice> i know that if i do"rm -r ." i will lose allt he .git 's
20140518 22:44:28< gfgtdf> iceiceice: https://github.com/gfgtdf/wesnoth-old/commit/26d60c0a8fb3a2d32e667f15f69915375b48f77b
20140518 22:44:50< gfgtdf> iceiceice: didnt test yet have to wait for compilation
20140518 22:46:41< iceiceice> damn this really sucks
20140518 22:46:45< iceiceice> maybe i should just clone again
20140518 22:47:30< iceiceice> hmm wait
20140518 22:47:35< iceiceice> are there .git files inside the subdirectories?
20140518 22:47:37< iceiceice> or only at root
20140518 22:48:34-!- apoi [~andi@85-126-180-242.volume.xdsl-line.inode.at] has quit [Ping timeout: 240 seconds]
20140518 22:49:11< iceiceice> huzzah i guess i'm saved, unless i'm actually corrupted...
20140518 22:51:11< iceiceice> gfgtdf your commit looks good
20140518 22:51:13< iceiceice> i will test though
20140518 22:51:34< gfgtdf> iceiceice: can you test/push it on 1.12 ?
20140518 22:51:53< iceiceice> yeah ok
20140518 22:53:39< iceiceice> hmm i think its making me do a full rebuild though because i deleted most of the repo :/
20140518 22:53:58< iceiceice> idk did you change stuff that would cause a full rebuild?
20140518 22:54:01< iceiceice> i guess you changed team hmm...
20140518 22:54:47< iceiceice> oh 
20140518 22:54:49< iceiceice> one thing i thought to do
20140518 22:54:58< iceiceice> i had wrote a lua check_victory blocker a while ago
20140518 22:55:07< iceiceice> but i got stuck trying to write a unit test for the thing iw as supposed to be fixing :/
20140518 22:55:18< iceiceice> i dont actually know how to reproduce the thing that happened to slowthinkers scenario
20140518 22:55:38< iceiceice> how do i induce a synced context?
20140518 22:55:47< iceiceice> is anyting in [lua] automatically in its own synced context?
20140518 22:56:40< gfgtdf> iceiceice: no lua is a morlan wml action
20140518 22:57:05< gfgtdf> a synced context is vasicly a player action liek move, attack, wml_menu_item lcik
20140518 22:57:08< gfgtdf> basicly
20140518 22:57:14< iceiceice> ok
20140518 22:57:32< iceiceice> so i can definitely reproduce like,
20140518 22:57:33< gfgtdf> iceiceice: but you cannot invoke it with [do_command] if you are alrady in a synced context
20140518 22:57:39< iceiceice> go to 1.11 wesnoth,
20140518 22:57:42< iceiceice> download conquest- add-on
20140518 22:57:48< iceiceice> try to start, it just is defeated
20140518 22:58:05< gfgtdf> iceiceice: immideately after start =
20140518 22:58:08< gfgtdf> ?
20140518 22:58:11< iceiceice> its not immediate
20140518 22:58:18< iceiceice> he has this sequence,
20140518 22:58:38< iceiceice> you can set up your game, and then the random seed is chosen
20140518 22:58:42< iceiceice> and the host has an option to save
20140518 22:58:46< iceiceice> basically its like your deterministic mode
20140518 22:59:03< iceiceice> during that time all the units are removed, the host controls a spearman or something in a tiny map
20140518 22:59:09< iceiceice> and he jumps to a pool of water to start
20140518 22:59:20< iceiceice> but in the 1.12 it like ends in defeat immeaditely
20140518 22:59:25< iceiceice> i think maybe as soon as you click on him or smothing
20140518 22:59:33< iceiceice> dont remember xactly
20140518 23:00:57< gfgtdf> iceiceice: hm i currenty cannot test becasue teh cmpile might blame "cannot write to file"
20140518 23:01:03< gfgtdf> iceiceice: i think i'll add a defeat_condition "always" too 
20140518 23:01:24< iceiceice> ok makes sense
20140518 23:01:55< iceiceice> so i mean you can see if travis built it and passed the unit tests
20140518 23:01:57< iceiceice> most likely he did
20140518 23:02:22< gfgtdf> iceiceice: travis does unti test always ?
20140518 23:02:23< Necrosporus> http://pastebin.com/hbw3jnvx This is supposed to kill units on volcano and chasm, plus non-aquatic ones in deep water but it doesn't
20140518 23:02:29< Necrosporus> it kills everything
20140518 23:02:45< iceiceice> y
20140518 23:03:16< iceiceice> gfgtdf: https://github.com/wesnoth/wesnoth/blob/master/.travis.yml#L11
20140518 23:03:35< gfgtdf> iceiceice: hmm  does tavis take more time mow ?
20140518 23:04:30< iceiceice> https://travis-ci.org/wesnoth/wesnoth
20140518 23:06:59< iceiceice> wow all the unit tests failed :O
20140518 23:08:24< iceiceice> the only one that passed was the one that breaks the replay
20140518 23:09:35< iceiceice> hmm well for some reason they always pass on my machine and fail on travis :hmm:
20140518 23:10:14< iceiceice> at least for 3 days...
20140518 23:10:20< iceiceice> i should have been checking these logs i guess :/
20140518 23:11:12< gfgtdf> iceiceice: they worked before on travis ?
20140518 23:11:36< iceiceice> hmm 9 days ago they were fine
20140518 23:11:40< iceiceice> but 3 days ago they were failing
20140518 23:15:13< iceiceice> i think theres something wrong with the new script, i'm not sure
20140518 23:15:25-!- shadowm_laptop [~ignacio@186.11.45.72] has joined #wesnoth-dev
20140518 23:15:27< iceiceice> maybe travis doesnt like the unix timeout or something?
20140518 23:22:29-!- travis-ci [~travis-ci@ec2-54-87-145-36.compute-1.amazonaws.com] has joined #wesnoth-dev
20140518 23:22:29< travis-ci> [travis-ci] gfgtdf/wesnoth-old#98 (defeat_condition - 26d60c0 : gfgtdf): The build passed.
20140518 23:22:29< travis-ci> [travis-ci] Build details : http://travis-ci.org/gfgtdf/wesnoth-old/builds/25476920
20140518 23:22:29-!- travis-ci [~travis-ci@ec2-54-87-145-36.compute-1.amazonaws.com] has left #wesnoth-dev []
20140518 23:24:50-!- shadowm_laptop [~ignacio@186.11.45.72] has quit []
20140518 23:25:29-!- shadowm_laptop [~ignacio@wesnoth/developer/shadowmaster] has joined #wesnoth-dev
20140518 23:27:51< mattsc> In case somebody wants to see the funny way in which units move now:  https://www.dropbox.com/s/hppj11nuzs6whpg/zigzag2.mov
20140518 23:31:00< iceiceice> hmm...
20140518 23:31:26< iceiceice> mattsc try building this commit maybe
20140518 23:31:27< iceiceice> 741051b0b34b4de7cf768569c6fd8afe8f8138a7
20140518 23:33:03< mattsc> iceiceice: okay - has to wait though, I’m on my way out the door again.  
20140518 23:34:43< iceiceice> ok
20140518 23:35:42< shadowm> How does that even happen?
20140518 23:36:07< shadowm> Did the pathfinder break or...?
20140518 23:38:38< gfgtdf> mattsc: i just tests and i dont have this problem on my master.
20140518 23:42:03< iceiceice> it doesnt happen at this commit 6475cd1ad2d72c1201944dcaf3e011ebbdc2d604
20140518 23:42:12< iceiceice> but it did happen on my master
20140518 23:43:40< iceiceice> shadowm: we observed some weirdness in the get_relative_dir function last night
20140518 23:43:49< iceiceice> and also it was relying on legacy map_location arithmetic
20140518 23:43:57< iceiceice> so we stayed up and made a replacement
20140518 23:44:03< iceiceice> but i think there might be an off by one error somewhere
20140518 23:44:19< iceiceice> idk it might be something else but my best guess is that was it,
20140518 23:44:36< iceiceice> and i think its entirely cosmetic if that makes you feel any better althoguh i guess formula ai may be using that function
20140518 23:46:02< iceiceice> ok it doesnt affect this commit either: 741051b0b34b4de7cf768569c6fd8afe8f8138a7
20140518 23:46:12 * shadowm slowly backs away.
20140518 23:48:39< iceiceice> ok it was introduced in our commit 829b74c2beaa18eda42710c364b12c987f9caed5
20140518 23:49:25< iceiceice> we were going to look at this again anyways...
20140518 23:50:20< gfgtdf> iceiceice: hm i still see nothing.
20140518 23:51:56< iceiceice> gfgtdf:
20140518 23:52:02< iceiceice> you could build the test executable?
20140518 23:52:18< gfgtdf> iceiceice: whcih test executable ?
20140518 23:52:18< iceiceice> and so validate the map_location output?
20140518 23:52:27< iceiceice> idk if you build target test
20140518 23:53:16< gfgtdf> i think the default msvc projectfile doent containg that.
20140518 23:53:21< gfgtdf> contain
20140518 23:56:42< iceiceice> hmm can you show me the impl of get_relative_dir
20140518 23:56:49< iceiceice> in src/map_location.cpp?
20140518 23:58:07< iceiceice> heres the one thats on master now, that coffee wrote last night
20140518 23:58:08< iceiceice> http://pastebin.com/DuS5XdJA
--- Log closed Mon May 19 00:00:27 2014