--- Log opened Fri Aug 26 00:00:37 2016
20160826 00:01:52-!- Shiki [~Shiki@141.39.226.227] has quit [Quit: Verlassend]
20160826 00:04:04< celmin> Discovered an interesting bug on the forum: https://forums.wesnoth.org/viewtopic.php?f=4&t=44356&view=unread#unread
20160826 00:13:00< vultraz> I have isolated the crash
20160826 00:19:34< vultraz> if(game_info* game = lobby_info_.get_game_by_id(id)) {
20160826 00:19:38< vultraz> that causes the crash
20160826 00:21:02< vultraz> question is why...
20160826 00:21:15< celmin> Okay.
20160826 00:21:30< celmin> So it crashes specifically at that line, or somewhere in get_game_by_id?
20160826 00:21:43< celmin> If the former, it suggests that lobby_info_ is somehow invalid...
20160826 00:21:52< vultraz> could be in get_game_by_id..
20160826 00:21:56< vultraz> let me debug further
20160826 00:24:37-!- jswensen [~jswensen@s48075040.temp.wsu.edu] has quit [Quit: jswensen]
20160826 00:32:39< celmin> Do you have anything specific that I should look at next?
20160826 00:33:01< vultraz> the thing where the filter doesn't work correctly once you change types in create?
20160826 00:33:19< celmin> I haven't understood the problem.
20160826 00:33:33< vultraz> example: set players filter to 2, switch to random maps. the proper number of random maps will show, but if you move the slider back to Any, none will show.
20160826 00:33:34-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 260 seconds]
20160826 00:34:45-!- celmin [~celticmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The minstrel departs, to spread the music to the masses!]
20160826 00:35:59< vultraz> make sense
20160826 00:36:01< vultraz> ?
20160826 00:39:30< vultraz> it's something in get_game_by_id..
20160826 00:39:57< celticminstrel> Well, it gives me something to work on, at least.
20160826 00:40:44< vultraz> https://github.com/wesnoth/wesnoth/blob/master/src/gui/dialogs/lobby/info.cpp#L240
20160826 00:40:45< vultraz> hmmm
20160826 00:41:22< celticminstrel> I can't see any way for that to go wrong though...
20160826 00:41:29< vultraz> neither can I
20160826 00:41:50< celticminstrel> Unless the lobb_info_ itself was somehow invalidated.
20160826 00:42:20< celticminstrel> Does the MP lobby store it directly or only hold a reference?
20160826 00:42:39< vultraz> reference
20160826 00:43:05< celticminstrel> Where is it held, then?
20160826 00:43:14< celticminstrel> Where is the actual declaration of the lobby info?
20160826 00:43:23< celticminstrel> The non-reference declaration.
20160826 00:44:12< vultraz> enter_lobby_mode
20160826 00:44:26< vultraz> game_init/multiplayer.cpp:655
20160826 00:45:52< celticminstrel> Okay, so it's safe to assume that it's never invalidated while the lobby is still being displayed.
20160826 00:47:46< Aginor> unless something is copying the lobby thing and passing that instead
20160826 00:48:35< vultraz> the lambda is capture-by-reference...
20160826 00:50:57< celticminstrel> It's capturing everything by reference. :(
20160826 00:51:25< vultraz> hm?
20160826 00:51:34< vultraz> you found a problem?
20160826 00:51:40< celticminstrel> No, it's not a problem.
20160826 00:52:11< celticminstrel> I don't actually know if [&] really means capture everything. It might mean capture anythign that's actually used.
20160826 00:53:22< celticminstrel> Idea: For the join and observe callbacks, try [this, get_game_index_from_id] instead of [&, this].
20160826 00:53:40< celticminstrel> ie, copying the lambda instead of referencing it.
20160826 00:54:17< celticminstrel> For the lambda itself I think you only need [this], right?
20160826 00:54:32< celticminstrel> Rather than [&]
20160826 00:55:04< celticminstrel> Seems unlikely that that would be the problem, but the other capture could be.
20160826 00:55:26< celticminstrel> Maybe the lambda was destroyed by the time it was called, and what should have been a reference to "this" had been overwritten with other data.
20160826 01:02:22< vultraz> gaahhhhhhhhh 
20160826 01:02:25< vultraz> still crashes
20160826 01:06:44< vultraz> not using a lambda either doesn'thelp
20160826 01:09:26< celticminstrel> So changing the capture method didn't help...
20160826 01:09:45< vultraz> nein
20160826 01:12:22< vultraz> wait a minute..
20160826 01:12:55-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 01:13:23< vultraz> hm
20160826 01:14:05< vultraz> game_list returns lobby_info_.gamelist()... the id of the correct game is passed select_game where it's set directly to selected_game_id_
20160826 01:14:10< vultraz> hm..
20160826 01:14:37< vultraz> I wish I could print the value of 'selected_game_id_' somewhere
20160826 01:18:00-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 276 seconds]
20160826 01:30:11< vultraz> ah!
20160826 01:30:15< vultraz> celticminstrel: I've found it
20160826 01:30:32< celticminstrel> Yay! What was it?
20160826 01:30:52< vultraz> the plugin was calling context.join({}). Using context.join() does not crash
20160826 01:31:03< celticminstrel> Uhh. What?
20160826 01:31:49< vultraz> I don't know
20160826 01:32:04< celticminstrel> It should work with context,join({}) though.
20160826 01:32:06< vultraz> perhaps since the plugin didn't have a cfg argument?
20160826 01:32:12< celticminstrel> Huh?
20160826 01:32:36< vultraz> then again, that shouldn't be a problem..
20160826 01:32:50< vultraz> anyway, it doesn't crash
20160826 01:33:36< vultraz> unfortunately now it doesn't atually join the game
20160826 01:36:01< vultraz> celticminstrel: I suspect this is because mp connect uses the gui1 context
20160826 01:36:39< celticminstrel> I don't think that would matter.
20160826 01:36:53< vultraz> different objects?
20160826 01:36:55< celticminstrel> The context isn't linked to GUI1 or GUI2, only called by them.
20160826 01:37:19< celticminstrel> And they were already different objects in the GUI1 case.
20160826 01:37:30< celticminstrel> Remember that GUI1 calls .reset() each time.
20160826 01:37:37< vultraz> hm
20160826 01:39:00< vultraz> i wonder if this has to do with the registered users only setting..
20160826 01:39:07< vultraz> that seems to be enabled
20160826 01:39:30< celticminstrel> Edit the host plugin to call .update_settings{registered_users = false} or something?
20160826 01:39:39< celticminstrel> Whatever the key was.
20160826 01:40:01< celticminstrel> I wonder if there's a flag to ignore prefs.
20160826 01:40:49< celticminstrel> Doesn't seem to be. It'd help for this sort of thing, though on Travis it doesn't really matter.
20160826 01:41:02< celticminstrel> Since it's always run on a blank slate.
20160826 01:42:58< celticminstrel> Oh, you could just set --userdata-dir to point to an empty folder or something, I guess.
20160826 01:43:59< vultraz> hm
20160826 01:44:43< vultraz> blah
20160826 01:44:45< vultraz> I added  context.update_settings({registered_users = false})
20160826 01:44:55< vultraz> but that just stops you at the mp configure screen :|
20160826 01:45:02< vultraz> create*
20160826 01:45:08< vultraz> the host does not proceed
20160826 01:45:34< celticminstrel> Eh...?
20160826 01:46:06< vultraz> i dunno
20160826 01:46:39< vultraz> that is the correct format, is it not?
20160826 01:47:02< celticminstrel> Looks like it, yeah.
20160826 01:47:49< vultraz> calling context.launch afterwards doesn't eotk either
20160826 01:49:11< vultraz> ideas?
20160826 01:50:26< vultraz> oh wait
20160826 01:50:28< vultraz> wrong place
20160826 01:51:00< vultraz> ok, join still doesn't join
20160826 01:51:53< vultraz> last chat message is "found test game"
20160826 01:52:36< celticminstrel> So select_game doesn't crash anymore, but doesn't seem to work either?
20160826 01:52:57< celticminstrel> I wonder if it's really necessary to yield between select_game and join.
20160826 01:52:58< vultraz> no, it gets past select_game fine...
20160826 01:53:13< vultraz> I added another message right before join
20160826 01:53:19< vultraz> it prints
20160826 01:53:25< vultraz> so it's something in the call to join()
20160826 01:53:32< vultraz> (not join({}), since that crashes
20160826 01:53:33< vultraz> go figure
20160826 01:53:38< celticminstrel> And "did not find join..." does not print?
20160826 01:54:18< vultraz> no
20160826 01:54:22< vultraz> it found the test game
20160826 01:54:57< celticminstrel> [Aug 25@9:32:06pm] vultraz: perhaps since the plugin didn't have a cfg argument?
20160826 01:54:58< celticminstrel> What did you mean by that?
20160826 01:56:07< vultraz> the function argument list was just 'const config&'
20160826 01:56:17< vultraz> then i realized that shouldn't matter
20160826 01:56:47< celticminstrel> That means it has an argument but doesn't use it... though, I wonder if it should be...
20160826 01:57:22< celticminstrel> Maybe those configs should be taken by value instead of reference.
20160826 01:57:34< celticminstrel> Not sure though...
20160826 01:58:28< celticminstrel> The typedef indicates taking them by value though.
20160826 01:58:49-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has joined #wesnoth-dev
20160826 01:59:03< vultraz> still crashes either way with join({})
20160826 01:59:21< celticminstrel> Weird.
20160826 01:59:42< vultraz> i can't figure this out :|
20160826 02:01:15< vultraz> why should it crash with an empty table!
20160826 02:01:17< vultraz> why!
20160826 02:23:46-!- irker894 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160826 03:12:25-!- JyrkiVesterinen [~JyrkiVest@87-100-242-136.bb.dnainternet.fi] has joined #wesnoth-dev
20160826 03:22:49< celticminstrel> I'm not sure whether there's something wrong with the theme items Lua API, or if the test scenario just uses it wrong...
20160826 03:23:02< celticminstrel> Shouldn't unit_status take the unit as a parameter?
20160826 03:25:24-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has quit [Ping timeout: 276 seconds]
20160826 03:25:26-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160826 03:35:36-!- irker725 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160826 03:35:36< irker725> wesnoth: Jyrki Vesterinen wesnoth:master c438a2a21ae0 / src/units/unit.cpp: Fix: it wasn't possible to remove invulnerability with the shorthand https://github.com/wesnoth/wesnoth/commit/c438a2a21ae064b80b24e5e300c7389b962af039
20160826 03:36:26< JyrkiVesterinen> I had realized subconsciously that something was wrong with the code even as I was writing it.
20160826 03:36:40< JyrkiVesterinen> I understood what it was when I went to bed.
20160826 03:37:39< JyrkiVesterinen> I had thought that the invulnerability status would reset because the ;unit command copies the unit, and the top-level "invulnerable" key won't be true in the config used for copying.
20160826 03:38:10< JyrkiVesterinen> I forgot that the actual status is still carried in the "status" child table.
20160826 03:38:13< JyrkiVesterinen> Fixed now.
20160826 03:54:39-!- Kwandulin [~Miranda@p200300760F6F2FD334CC181A855407A2.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160826 03:55:01-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has joined #wesnoth-dev
20160826 03:59:50-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 04:01:54-!- JyrkiVesterinen [~JyrkiVest@87-100-242-136.bb.dnainternet.fi] has quit [Quit: .]
20160826 04:04:33-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 265 seconds]
20160826 04:33:40< celticminstrel> I don't understand what scoped_dialog is doing at all...
20160826 04:47:54-!- jswensen [~jswensen@cpe-98-145-147-33.natnow.res.rr.com] has joined #wesnoth-dev
20160826 04:50:55< jswensen> I’m assuming that if I can get the SDL_GPU working on ios (which it has support for) that the performance will be vastly improved?
20160826 04:52:29< celticminstrel> No idea. The SDL_GPU code hasn't been maintained though.
20160826 04:52:48< celticminstrel> You're welcome to try.
20160826 04:57:57< jswensen> It looks like it has a crazy big effect.  See http://nivrigdev.tumblr.com/post/137158222896/game-engine-benchmarking, especially on later Apple devices that have gles3 support (iphone 5s and newer). It is still double the performance even on their older devices. 
20160826 04:58:30< jswensen> When you say it hasn’t been maintained, does that mean that no one has tested whether it even still builds if SDL_GPU is enabled?
20160826 04:59:15< celticminstrel> As far as I know, no-one has tested that. Even if someone did verify that it compiles, I don't know if they went any further than that.
20160826 04:59:26< celticminstrel> The person most likely to know is Aginor, I think.
20160826 05:03:05< jswensen> OK…that can be a later task. I think I have all the dependencies built now (in Apple’s “fat” format so I can test on simulator and real device). Now the arduous task of sorting out which source files actually need to go into the XCode project. The only reason I got headed in the SDL_GPU direction was because it was throwing a thousand error and I realized that it was a completely separate library that shouldn’t be incl
20160826 05:03:06< jswensen> in the base project. I think I will need to make a careful traversal of all the CMakeLists.txt files to get an accurate listing.
20160826 05:03:28< celticminstrel> You could check the existing XCode project file.
20160826 05:03:58< jswensen> I think the code bases of 1.10 and 1.13+ look different enough that I might just rather look at the current CMAkeLists
20160826 05:04:16< jswensen> Oh, you mean the OSX Xcode project.
20160826 05:04:21< jswensen> That would work.
20160826 05:04:26< celticminstrel> Yes, under projectfiles/
20160826 05:11:52-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has joined #wesnoth-dev
20160826 05:15:35< jswensen> ugh…I need to figure out how to make XCode do search paths differently. There is a settings.h at the toplevel and a settings.h in the gui/widgets directory and it is getting confused. I think I have it set to only look in the top level and not recursively, but for some reason anything that includes setting.h is finding the one in gui/widgets. Same for image.h
20160826 05:17:12< celticminstrel> XCode's habit of flattening the filesystem hierarchy has annoyed me in the past and sometimes results in broken code being commited in Wesnoth. :/
20160826 05:17:44< JyrkiVesterinen> jswensen: IIRC, SDL_gpu was abandoned because it didn't improve performance.
20160826 05:18:44< JyrkiVesterinen> Wesnoth's rendering pipeline relies heavily on "unrendering", i.e. reading graphics from the framebuffer, storing it in a surface, and blitting the surface again when whatever was originally renderer needs to disappear.
20160826 05:19:11< jswensen> I may end up just changing the name of the ones in gui/widget that have the same name. I did that with image.h and image.cpp to timage.h and timage.cpp and it worked great.
20160826 05:19:25< JyrkiVesterinen> Unrendering speeds up software rendering, but slows down hardware rendering. Reading the framebuffer is a very slow operation with hardware rendering.
20160826 05:19:54-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 05:20:07-!- fabi [~fabi@176.4.64.232] has joined #wesnoth-dev
20160826 05:21:10< jswensen> JyrkiVesterinen: I suspect that just going from SDL1.3 to SDL2 should have improvements. I’m not sure what SDL_GPU does above and beyond SDL2, but supposedly SDL2 has some sort of OpenGL harware acceleration for OpenGL ES2
20160826 05:21:10-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Read error: Connection reset by peer]
20160826 05:21:37-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 05:22:27< celticminstrel> jswensen: We aren't currently taking advantage of SDL2 hardware acceleration.
20160826 05:23:04-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 05:23:22-!- fabi_ [~fabi@176.0.125.135] has quit [Ping timeout: 244 seconds]
20160826 05:23:43< ancestral> jswensen: SDL2 did have an imporvement from what I can tell
20160826 05:23:52< ancestral> in the realm of graphics performance
20160826 05:23:56-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has quit [Read error: Connection reset by peer]
20160826 05:24:14< ancestral> But that’s simply a subjective observation, I have no hard facts
20160826 05:24:20-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 05:25:40-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has quit [Read error: Connection reset by peer]
20160826 05:25:45< jswensen> Regardless, SDL2 has better iOS support (e.g. it is part of the maintained package, rather than a third-party add-on)
20160826 05:25:58-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 05:26:04-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 250 seconds]
20160826 05:26:31< ancestral> SDL2 has better a lot suport
20160826 05:26:35< ancestral> *support
20160826 05:26:52< ancestral> Interesting though
20160826 05:27:05-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has quit [Read error: Connection reset by peer]
20160826 05:27:17-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 05:32:07-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 265 seconds]
20160826 06:07:46< jswensen> Sweet! I have the iOS build down to a single error (though it is occuring about 50 times). This is still a far cry from getting the OpenGL context set up right, SDL drawing to it, and all the sizing adjusted for iPhone/iTouch and iPad devices, but at least the majority of the build woes are out of the way.
20160826 06:08:04 * celticminstrel curious what the error is.
20160826 06:09:08< celticminstrel> Regarding sizing of UI elements, I'm not sure what the size of the iOS screens are these days, but I've been trying to ensure the game remains playable at 800x600.
20160826 06:09:27< celticminstrel> So you might not have a lot of work there.
20160826 06:13:27< jswensen> Everything since iPhone4 has been 640x960 or larger. Since about 93% of iOS users are on >iOS8 and >iphone4/itouch4thgen, I think we could just dispense with supporting older devices than that.
20160826 06:14:43< JyrkiVesterinen> Well, the Wesnoth UI has been designed for mouse and keyboard.
20160826 06:14:55< celticminstrel> Ah, that's true.
20160826 06:15:04< JyrkiVesterinen> Even though it fits to the iPhone screen, it's likely hard to use with a touchscreen.
20160826 06:15:06< jswensen> the error was something to do with BOOST_STATIC_ASSERT in arg.h. I saw a post on the forums about making a change in arg.h. It looks like that solved it.
20160826 06:15:35< celticminstrel> Oh, that error, okay.
20160826 06:16:01< jswensen> JyrkiVesterinen: The nice thing is that SDL2 already has standard touch to mouse mappings. As far as I can tell, the old iOS version used the SDL1.3 touch to mouse mappings as-is. You can adjust them, but might not need to.
20160826 06:16:35< JyrkiVesterinen> I meant that the hitboxes are likely too small for touch input.
20160826 06:16:51< JyrkiVesterinen> Usable, but not a very good experience, I suspect.
20160826 06:17:40< jswensen> I have played the older iOS version and it feels fine to me. My only complaint about the touch interface on the 1.10 iWesnoth is that there isn’t a good way to cancel a move after I have selected a unit. I tend to have to make a move and then undo.
20160826 06:21:59< vultraz> Are you the person I was speaking to about the iOS port on the forums, or was that someone different?
20160826 06:22:43-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 06:22:58< jswensen> Compile finished, but a bunch of linker errors. Looks like I will need readline as a lua dependency. Most of the others have to do with boost::iostream::zlib, boost::iostream::gzip, boost::iostream::bz2. Also a few other missing Boost symbols. Those ones seem strange to me because I built the whole Boost tamale.
20160826 06:23:01< jswensen> vultraz: Yep
20160826 06:23:30< celticminstrel> You can disable readline.
20160826 06:23:32< vultraz> You are?
20160826 06:23:53< celticminstrel> Lua doesn't depend on it, and the Lua console support in Wesnoth is also optional.
20160826 06:23:59< jswensen> I had just wanted to push forward and get all the dependencies built. I thought that was going to be a big hurdle, but the GStreamer cerbero dependency build tool for GStreamer did almost all the work.
20160826 06:25:16< vultraz> celticminstrel: so i tried passing a non-entry config to join(), and it still crashes
20160826 06:25:28< vultraz> obviously removing {} is not the solution
20160826 06:27:10-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 250 seconds]
20160826 06:31:01< vultraz> hm
20160826 06:31:03< vultraz> interesting
20160826 06:31:24-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has joined #wesnoth-dev
20160826 06:31:24< vultraz> removed the get_game_index_from_id, still fails
20160826 06:33:14-!- fabi [~fabi@176.4.64.232] has quit [Quit: Konversation terminated!]
20160826 06:33:17< vultraz> ok come on
20160826 06:35:43-!- irker725 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160826 06:39:39< vultraz> I really cannot figure this out for the life of me
20160826 06:53:53-!- jswensen [~jswensen@cpe-98-145-147-33.natnow.res.rr.com] has quit [Quit: jswensen]
20160826 06:58:35-!- irker175 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160826 06:58:35< irker175> wesnoth: Charles Dang wesnoth:master 9870a94b7591 / src/gui/dialogs/lobby/lobby.cpp: MP Lobby: some fixes to plugin context https://github.com/wesnoth/wesnoth/commit/9870a94b7591572d0ba558bc09a3d66695e8220f
20160826 06:58:36< irker175> wesnoth: Charles Dang wesnoth:master 8d0a75a54a8c / host.lua: Ensure registered_users_only is disabled in mp tests https://github.com/wesnoth/wesnoth/commit/8d0a75a54a8cb23946399e76a672be8517d01181
20160826 07:01:51-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has quit [Ping timeout: 276 seconds]
20160826 07:02:39< celticminstrel> That's going to break Travis.
20160826 07:02:48< vultraz> it is?
20160826 07:03:03< celticminstrel> Yes, because the GUI1 version doesn't add the update_settings function to the context.
20160826 07:03:07< vultraz> oh
20160826 07:03:09< vultraz> blah
20160826 07:03:11< vultraz> :|
20160826 07:03:21< celticminstrel> At least, I'm assuming a Lua error in the plugin would break Travis. I could be wrong.
20160826 07:03:52< celticminstrel> You can either test it first (if context.update_settings then ...) or add it in the GUI1 version.
20160826 07:05:31< celticminstrel> In select_game you might want to check that the index is actually valid.
20160826 07:05:44 * celticminstrel sleep now because 3am.
20160826 07:05:54-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!]
20160826 07:06:00< vultraz> none of that explains the crash in join()
20160826 07:28:30-!- boucman_work [~boucman@229.29.205.77.rev.sfr.net] has joined #wesnoth-dev
20160826 07:30:15-!- Kwandulin [~Miranda@p200300760F6F2FD334CC181A855407A2.dip0.t-ipconnect.de] has quit [Ping timeout: 258 seconds]
20160826 07:34:37-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20160826 07:38:03-!- Kwandulin [~Miranda@p200300760F6F2FEC219AD9F53B09C16D.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160826 07:53:12-!- travis-ci [~travis-ci@ec2-54-81-200-97.compute-1.amazonaws.com] has joined #wesnoth-dev
20160826 07:53:13< travis-ci> wesnoth/wesnoth#10565 (master - 8d0a75a : Charles Dang): The build has errored.
20160826 07:53:13< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/155259105
20160826 07:53:13-!- travis-ci [~travis-ci@ec2-54-81-200-97.compute-1.amazonaws.com] has left #wesnoth-dev []
20160826 08:11:00-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 08:15:16-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 244 seconds]
20160826 08:20:53-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Quit: wedge009]
20160826 08:21:10-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160826 08:40:57-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has joined #wesnoth-dev
20160826 08:41:20-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20160826 09:00:34< zookeeper> btw, if we want some kind of titlescreen logo animation back at some point, something like this would be easy to do: https://dl.dropboxusercontent.com/u/63964618/wesnoth/logo_pixel_fade.gif
20160826 09:01:06< zookeeper> dunno if we do, though, considering that you already see the logo in the loading screen, so it'd be weird to have a fancy animation for it afterwards.
20160826 09:01:34< Aginor> zookeeper: we should make a widget that has animation stubs and can be subclassed
20160826 09:03:11< zookeeper> well i have no opinion on that :p
20160826 09:03:56 * Aginor lends zookeeper his
20160826 09:06:54< zookeeper> hmm...
20160826 09:07:01< zookeeper> i think we should make a widget that has animation stubs and can be subclassed
20160826 09:07:12< Aginor> hehe
20160826 09:07:23< Aginor> well played
20160826 09:13:52-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has quit [Quit: .]
20160826 09:24:58-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160826 09:27:58-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Read error: Connection reset by peer]
20160826 09:28:28-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Client Quit]
20160826 09:32:25-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has joined #wesnoth-dev
20160826 09:38:42-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160826 09:40:52< tad_> Anyone know what this means? "ALSA lib pcm.c:8251:(snd_pcm_recover) underrun occurred" I'm seeing several dozen copies of that line before the game dies with a segmentation fault. I was working fine and cleared my .cache and .local/share and .config for wesnoth 1.13 and now it won't start.
20160826 09:42:05< JyrkiVesterinen> ALSA = Advanced Linux Sound Architecture, a sound system.
20160826 09:42:21< JyrkiVesterinen> I'm guessing that the underrun means an audio buffer underrun.
20160826 09:43:00< JyrkiVesterinen> That is, that ALSA drains an audio buffer and complains that it hasn't gotten new audio samples to play.
20160826 09:43:32< JyrkiVesterinen> I think it's not the cause of a crash, just a symptom.
20160826 09:43:57< tad_> Hmm. Well I can disable sound and see if that makes that message go away. The segfault is something to do with lua integers. I'm re-building from the current master before digging into it.
20160826 09:46:33< tad_> Disabling sound made the underrun messages go away. Now to find that segfault.
20160826 09:48:57< tad_> The segfault is in a C++ lua helped when it's trying to load the starting locations from the map. ISTM celmin was working on that in the past day .. hmm ..
20160826 09:56:13-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160826 09:59:16-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 10:03:33-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 240 seconds]
20160826 10:16:16-!- irker175 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160826 10:52:11-!- Kwandulin [~Miranda@p200300760F6F2FEC219AD9F53B09C16D.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160826 10:57:46-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160826 11:07:33-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has quit [Ping timeout: 240 seconds]
20160826 11:08:44-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has joined #wesnoth-dev
20160826 11:13:41-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has quit [Ping timeout: 244 seconds]
20160826 11:14:39-!- TheJJ [~rofl@ipbcc36896.dynamic.kabel-deutschland.de] has joined #wesnoth-dev
20160826 11:32:37-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160826 11:46:56-!- Kwandulin [~Miranda@p200300760F6F2FEC5469FAB7AF64FDB7.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160826 11:47:35-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 11:52:25-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 265 seconds]
20160826 12:17:26-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160826 12:23:17-!- Shiki [~Shiki@141.57.58.245] has joined #wesnoth-dev
20160826 12:25:10-!- VultCave [~chatzilla@124.109.10.167] has joined #wesnoth-dev
20160826 12:25:59-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Read error: Connection reset by peer]
20160826 12:26:10-!- VultCave is now known as vultraz
20160826 12:51:11-!- Shiki [~Shiki@141.57.58.245] has quit [Quit: Verlassend]
20160826 12:57:29-!- JyrkiVesterinen [~JyrkiVest@nblzone-242-23.nblnetworks.fi] has quit [Quit: .]
20160826 13:01:36-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has joined #wesnoth-dev
20160826 13:03:12-!- boucman_work [~boucman@229.29.205.77.rev.sfr.net] has quit [Ping timeout: 244 seconds]
20160826 13:10:04-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160826 13:18:16-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.1 Aria]
20160826 13:23:16-!- Shiki [~Shiki@141.57.58.245] has joined #wesnoth-dev
20160826 13:28:37-!- Kwandulin [~Miranda@p200300760F6F2FEC5469FAB7AF64FDB7.dip0.t-ipconnect.de] has quit [Quit: Kwandulin]
20160826 13:56:21-!- boucman_work [~boucman@229.29.205.77.rev.sfr.net] has joined #wesnoth-dev
20160826 14:15:34-!- irker651 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160826 14:15:34< irker651> wesnoth: Charles Dang wesnoth:master a04390223653 / src/gui/dialogs/multiplayer/mp_create_game.cpp: MP Create: merge Create and Configure plugin contexts https://github.com/wesnoth/wesnoth/commit/a04390223653b2dd283d0c782e107398a86cdeb3
20160826 14:15:35< irker651> wesnoth: Charles Dang wesnoth:master 793b8491ade1 / host-gui2.lua host.lua join-gui2.lua utils/travis/mp_test_executor-gui2.sh: Added GUI2-compatible versions of mp tests and plugin scripts https://github.com/wesnoth/wesnoth/commit/793b8491ade1ebe1d1b2b1f3bf0b595c4bda465c
20160826 14:16:35-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160826 14:30:07< irker651> wesnoth: Charles Dang wesnoth:master 6332bd2d8d92 / src/gui/dialogs/multiplayer/mp_create_game.cpp: MP Create: properly handle select_game/select_mod plugins callbacks https://github.com/wesnoth/wesnoth/commit/6332bd2d8d92cfd60d83cfd56d614e0edad3ed90
20160826 14:30:10< irker651> wesnoth: Charles Dang wesnoth:master faab5c814e7d / src/gui/dialogs/multiplayer/mp_create_game.cpp: MP Create: minor cleanup to custom options setup https://github.com/wesnoth/wesnoth/commit/faab5c814e7d37504723cbd3d292bba6844a4037
20160826 14:47:50< mattsc> zookeeper: I am mostly done optimizing (more or less) the high_xp_attack AI for speed.  I am planning to turn that into a PR after that, so that people can comment on the code more easily.  Do you want to check out the behavior before or after that?
20160826 14:48:06< mattsc> The behavior is not going to change at all from the version you can test right now.
20160826 14:48:58< mattsc> And the speed differences will really only be noticable when there are hundreds of units on the map, I think.
20160826 14:49:45-!- Kwandulin [~Miranda@p200300760F6F2FEC6899A4A98FD44423.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160826 14:50:02-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160826 14:53:14-!- travis-ci [~travis-ci@ec2-54-144-252-156.compute-1.amazonaws.com] has joined #wesnoth-dev
20160826 14:53:15< travis-ci> wesnoth/wesnoth#10566 (master - 793b849 : Charles Dang): The build has errored.
20160826 14:53:15< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/155353484
20160826 14:53:15-!- travis-ci [~travis-ci@ec2-54-144-252-156.compute-1.amazonaws.com] has left #wesnoth-dev []
20160826 14:55:44-!- higgins [~higgins@68.ip-149-56-14.net] has quit [Quit: Leaving]
20160826 15:01:31-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc]
20160826 15:12:44-!- Shiki [~Shiki@141.57.58.245] has quit [Ping timeout: 244 seconds]
20160826 15:16:20< vultraz> whyy is clang failing
20160826 15:17:36-!- boucman_work [~boucman@229.29.205.77.rev.sfr.net] has quit [Remote host closed the connection]
20160826 15:18:48-!- higgins [~higgins@68.ip-149-56-14.net] has joined #wesnoth-dev
20160826 15:24:01-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160826 15:26:05-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160826 15:27:21< tad_> Thread 1 "wesnoth" received signal SIGSEGV, Segmentation fault. 0x00000000010a7b2e in luaH_setint(lua_State*, Table*, int, lua_TValue*) ()
20160826 15:27:33< zookeeper> mattsc, ah, probably easier after
20160826 15:27:35< vultraz> oh dear
20160826 15:28:28< mattsc> zookeeper: sounds good
20160826 15:29:56< tad_> Segfault occurs launching any campaign after story screens, before map first displays. Reading the backtrace it looks to be trying to find the starting positions on the map.
20160826 15:31:09< tad_> I can see a couple possible commits as culprit. But it takes a long time per test, so I'm hoping someone will say "wait, try this ..."
20160826 15:32:58< vultraz> wait until celticminstrel returns
20160826 15:33:22< vultraz> tad_: btw, what is your thought on the new leader display in Load Game?
20160826 15:35:22< tad_> vultraz: I liked it, in general, but wanted to clear out that file you mentioned in case it was the cause of the error I'm seeing before telling you. So, since I can't test if it was a data corruption .. I will tell you I'm seeing the leader duplicate over and over (some with inage, some with "?" image) as I reload the scenario while testing.
20160826 15:36:07< vultraz> the ? image is intentional if the image can't be found
20160826 15:36:56< tad_> Which begs the question of why sometimes, yes, sometimes, no, when it's adding yet another copy of the leader to the listbox because I loaded again.
20160826 15:37:19< vultraz> another copy of the leader, you say?
20160826 15:37:21< vultraz> this is odd..
20160826 15:38:44< tad_> I'm beating up on TSG S05 and all myriad paths to victory and defeat. It fog, no shroud, at EASY and every cycle it's adding a new copy of Deoran to the leader list on Load panel.
20160826 15:39:17< vultraz> this is odd..
20160826 15:40:16< tad_> Plus the scenario-start (which always has none listed) shows the same scroll-bar as the Turn 1 save I arrowed past (it's always turn 1 because I am debugging and never end turn while testing another path through the code)
20160826 15:40:22< tad_> oh ..
20160826 15:41:07< tad_> zookeeper: If you want to see the mess I made of TSG S05 it should be up and working in my TSG-elven fork. 
20160826 15:41:19< irker651> wesnoth: Charles Dang wesnoth:master d27ff5768260 / data/core/ (35 files in 2 dirs): New baseframes and animations for the Heavy Infantryman line https://github.com/wesnoth/wesnoth/commit/d27ff576826093112d8897050d3057ca83a6840a
20160826 15:41:24 * zookeeper will check a bit later
20160826 15:41:51< zookeeper> vultraz, btw, did you run those through an optimizer?
20160826 15:41:56< vultraz> No
20160826 15:42:26< zookeeper> okay
20160826 15:42:27< tad_> zookeeper: thanks. actually, the commit is pretty clean especially if you ignore whitespace changes, it's just way huger than I aim for,
20160826 15:43:17< vultraz> tad_: can you send me a savegame?
20160826 15:43:38< tad_> Nope. I cleared out .cache, .config and .local/share
20160826 15:43:51< vultraz> what you describe should not be possible unless there are actually multiple copies of the leader
20160826 15:43:54< tad_> It was working fine until I did that.
20160826 15:45:30< tad_> I will, if the problem recurs when running from clean starting-point.
20160826 15:45:50< vultraz> ok
20160826 15:47:44-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 15:49:25< vultraz> have confirmed your crash
20160826 15:50:59< tad_> Well, that's good. I'm back to touch-one-file-makes-half-the-engine so it's forever testing.
20160826 15:51:10< tad_> My vote for the culprit is 9dbae3944839f2e55086f0e7c221ef7083c6be06
20160826 15:51:49< vultraz> possibly
20160826 15:52:36< tad_> I'm trying to revert it now. We'll see how long it takes to recompile and test on my lame-ass system which is recompiling files not touched.
20160826 15:52:52< vultraz> I doubt your system is worse than some of ours :P
20160826 15:53:08< vultraz> (though I'll admit I do have a 'medium' system)
20160826 15:54:38< tad_> I'm on an Arch in Virtualbox. And it's on my wifes old, slow HP laptop running Windows 10 .. both patched current. So it likes to take things ssllooww and every now and then Windows decides to update in the background.
20160826 15:54:54< tad_> But it's a VM so I rarely lose work like most Windwos 10 users.
20160826 15:55:24 * vultraz is on windows 10 and doesn't lose work
20160826 15:55:48-!- Shiki [~Shiki@141.57.58.245] has joined #wesnoth-dev
20160826 15:56:30< tad_> No? Wait until you've been editing a file for 10 hours and windows decides two days is long enough and reboots under you edit .. because you never left full-screen edit to see Windows had been asking for a while now.
20160826 16:03:25-!- DeFender1031 [~DeFender1@46-116-114-128.bb.netvision.net.il] has quit [Quit: I'm not back now.]
20160826 16:04:33< tad_> I love the new feature of windows 10 "quiet hours" where you tell it when you're on the computer. Except Microsoft has determined that I cannot possibly be using the computer for more than 12 hours at a stretch and, therefore, cannot prevent Windows from rebooting without permission while I'm working.
20160826 16:11:18-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160826 16:12:35< tad_> Speak 'o da devil.
20160826 16:13:02-!- Kwandulin [~Miranda@p200300760F6F2FEC6899A4A98FD44423.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160826 16:13:23< celticminstrel> ?
20160826 16:14:29< tad_> I hit a segfault, vultraz has confirmed he can reproduce it. I'm voting for one of your recent commits to be the culprit and started a revert of it compiling while waiting for your appearance.
20160826 16:14:41< celticminstrel> Which one?
20160826 16:14:52< tad_> [10:51] <tad_> My vote for the culprit is 9dbae3944839f2e55086f0e7c221ef7083c6be06
20160826 16:15:56< tad_> I cleared .cache, .config and .local/share of all wesnoth files and it segfaults on startup a few subroutines inside getting the starting map positions.
20160826 16:15:58< celticminstrel> vultraz: Didn't I say we shouldn't merge the create and configure contexts until the MP test was working? (The reasoning was just to get it working first without updating the plugin.) Or did you get the MP test working?
20160826 16:16:43< celticminstrel> Where's the crash?
20160826 16:16:51< celticminstrel> Oh.
20160826 16:17:01< tad_> [10:27] <tad_> Thread 1 "wesnoth" received signal SIGSEGV, Segmentation fault. 0x00000000010a7b2e in luaH_setint(lua_State*, Table*, int, lua_TValue*) ()
20160826 16:17:33< celticminstrel> What about the first Wesnoth code stack frame?
20160826 16:18:26< celticminstrel> vultraz: Why did you remove the dummy loop in the UPDATE_ATTRIBUTE macro? It was there for a good reason.
20160826 16:18:58-!- Shiki [~Shiki@141.57.58.245] has quit [Quit: Verlassend]
20160826 16:18:58-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has joined #wesnoth-dev
20160826 16:19:01< tad_> The backtrace is many lines.  All in the C++ .. don't know where you're asking about but about 5 or 6 frames up its in a call to get the starting positions. I can stop the make and get the frames, I guess.
20160826 16:19:47< celticminstrel> Well, a function beginning with lua?_ is not in the Wesnoth code, except for luaW_
20160826 16:20:46< tad_> #10 0x0000000000ebe1b8 in game_lua_kernel::run_wml_action(std::__cxx11::basic_string<char, std::char_traits<char>, std::allocator<char> > const&, vconfig const&, game_events::queued_event const&) ()
20160826 16:21:11< celticminstrel> Ugh, why do we need two separate plugins now... :|
20160826 16:21:19< tad_> #3  0x0000000000ec55eb in int dispatch<&game_lua_kernel::intf_get_starting_location>(lua_State*) ()
20160826 16:22:28< celticminstrel> So it's in intf_get_starting_location I guess...
20160826 16:23:05< tad_> And inside there it's doing a push which looks like one of those you touched in that commit .. hence my guess as to the culprit
20160826 16:23:37< celticminstrel> Right...
20160826 16:24:24< vultraz> celticminstrel: what was the good reason?
20160826 16:24:46< vultraz> celticminstrel: and I added a second version of the tests for gui2 
20160826 16:24:47< tad_> So I decided to revert that commit but the darned makefile decided I meant 'make all' and it's taking FOREVER!
20160826 16:25:10< celticminstrel> vultraz: It makes the macro be a single standalone statement, so that you can treat it like an ordinary function call without worrying about bizarre side-effects.
20160826 16:25:13< vultraz> celticminstrel: host had no problems going through Create and the game was hosted correctly
20160826 16:25:40< vultraz> celticminstrel: therefor I concluded I was safe to merge the contexts
20160826 16:25:49< vultraz> the problem with the tests is in join, and the lobby
20160826 16:26:40< celticminstrel> Oh, you're right, sorry.
20160826 16:27:09< celticminstrel> I guess you did confirm that it works with the existing plugin, even if something else doesn't.
20160826 16:27:34< vultraz> i have confirmed the new host-gui2 plugin works with the changes
20160826 16:27:56< celticminstrel> But the join plugin still crashes?
20160826 16:28:07< vultraz> yes
20160826 16:28:44< celticminstrel> Anyway, about the macro - suppose you wanted to do if(some_condition) UPDATE_ATTRIBUTE(whatever); else UPDATE_ATTRIBUTE(other_attribute);
20160826 16:29:08< celticminstrel> At least this example would be a compile error rather than silently unexpected behaviour, though.
20160826 16:30:23< celticminstrel> vultraz: on_mod_toggle is not the correct implementation of select_mod in my opinion.
20160826 16:30:37< celticminstrel> Actually, on_mod_toggle was a thorn in my side earlier, if I recall correctly.
20160826 16:30:39< vultraz> celticminstrel: neither was your implementation
20160826 16:30:51< vultraz> all you did was select the checkbox :|
20160826 16:31:13< celticminstrel> True.
20160826 16:31:17< vultraz> anyway, I still don't see your point about the 'do' block
20160826 16:31:38< vultraz> how is if() { if() {} } else { if() {} } invalid?
20160826 16:31:58< celticminstrel> Well, if you use the full braces, it would be fine, of course,
20160826 16:32:14< vultraz> I've been advised to always use full braces
20160826 16:33:34< celticminstrel> https://gcc.gnu.org/onlinedocs/cpp/Swallowing-the-Semicolon.html#Swallowing-the-Semicolon
20160826 16:33:42< vultraz> and plus, why would wrapping the macro's contents in scope braces instead of a do-loop be invalid?
20160826 16:34:08< celticminstrel> Because scope braces do not have a terminating semicolon.
20160826 16:34:58< vultraz> I see
20160826 16:35:13< vultraz> well, either way I foresee no need for conditionals here
20160826 16:35:17< celticminstrel> This can also be solved by adding a trailing else in the macro declaration, but my vague memory suggests that that still might have some pitfalls.
20160826 16:35:25< tad_> The idea of that suggestion for a macro is to make it into a 'statement' so the user-programmer does not get unexpected warnings/errors about semicolons 
20160826 16:35:47< celticminstrel> True, but I think it's good to follow good macro practice in case of future changes that you didn't foresee.
20160826 16:36:15< vultraz> c++ macros really feel clunky
20160826 16:36:35< celticminstrel> Anyway, about select_mod, I wanted it to set the mod's status, not toggle it.
20160826 16:36:38< tad_> c++ macro are c macros and, therefore, REALLY clunky
20160826 16:37:03< celticminstrel> I notice you didn't touch select_eras though... is that implementation correct?
20160826 16:37:17< vultraz> oh
20160826 16:37:22< vultraz> missed that one
20160826 16:38:13 * tad_ sighs. 
20160826 16:38:29< vultraz> hm?
20160826 16:38:55< tad_> make finally finished and reverting that commit did not solve the crash. So .. should I start a bisect (GAK)?
20160826 16:39:00< vultraz> (i suppose I should also fix select_type to forgo the UI interaction)
20160826 16:39:21< vultraz> tad_: no need, we'll find the cause
20160826 16:39:30< celticminstrel> Can you also change select_mod so that it sets the mod status to true instead of toggling it?
20160826 16:39:41-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 16:39:46< vultraz> not without editing the create engine
20160826 16:40:18< celticminstrel> Eh... I'll look into it then, I guess...
20160826 16:40:23< tad_> Oh .. hard to tell .. you're thinking it's related to the MP stuff you've been working on, then?
20160826 16:40:43< celticminstrel> I think it's unlikely that a crash in Lua would be related to the MP stuff.
20160826 16:42:55-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 250 seconds]
20160826 16:45:22< irker651> wesnoth: Charles Dang wesnoth:master a015ca97fe61 / src/gui/dialogs/multiplayer/mp_create_game.cpp: MP Create: removed a few more cases of UI interaction in the plugins context https://github.com/wesnoth/wesnoth/commit/a015ca97fe616e5179652db5d4ad40abb696e8ad
20160826 16:45:26-!- Greg-Bog_ [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160826 16:46:00-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 16:47:53-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160826 16:48:19-!- Appleman1234 [~Appleman1@KD036012030198.au-net.ne.jp] has quit [Ping timeout: 244 seconds]
20160826 16:49:56< zookeeper> tad_, ok, so, i'm not opposed to restructuring the ending like that
20160826 16:52:29< zookeeper> i'm not sure if that much of the epilogue text should be transferred to ethiliel in-scenario, but i guess it can work. i'm not sure what the rationale is though.
20160826 16:55:40-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 16:57:42< tad_> you're referring to the elven end-game .. the comments in-code were that the original designer didn't like the abrupt ending. That's no problem to me either way; just a first-take on how to handle it and easy to remove.
20160826 16:58:16< tad_> zookeeper: my main concern today is all the changes in S05 to clean up the objectives and conversations for all the ways it can play out.
20160826 16:59:42< celticminstrel> I think I figured out why my custom placement new didn't work in the Lua plugins awhile ago - its definition wasn't included.
20160826 17:00:03< vultraz> what?
20160826 17:00:07< tad_> zookeeper: istm that's a two-stage change first was that Urza Afalas could be dead, then I noticed there were a LOT more ways it could be played ...
20160826 17:03:29< zookeeper> tad_, oh one little detail: we (try to) use ’ instead of ' as apostrophes in text (so "Defeat Mal M'Brin" should be "Defeat Mal M’Brin", etc). although i'm myself among those people who always tend to forget that.
20160826 17:03:56< vultraz> hm, the thing where it tries to download addon dependencies is in the lobby
20160826 17:04:31< vultraz> will have to add that
20160826 17:04:33< tad_> That one (Mal M'brin) is beating me up. I can't type it and so copy-and-paste it. Every time I just type it it says I didn't do it the same as everrywhere else.
20160826 17:05:49< tad_> zookeeper: I agree it should be the same. And I'm seriously thinking of scrogging the id="Mal M'Brin" and making it id=somethingMoreLikeAVariableName
20160826 17:06:26< zookeeper> oh, hmh, it's using ' in the original text it seems. maybe it's a different kind of apostrophe on purpose, i dunno.
20160826 17:07:39< tad_> zookeeper: global search and replace is no problem. My main beef is the id= is a string and has some form of punctuation in it.
20160826 17:08:22< zookeeper> well everything's a string, one can enclose them in quotes if it makes things easier
20160826 17:10:33-!- Kwandulin [~Miranda@p200300760F6F2FECB14B9DF3D1BF7A9C.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160826 17:10:42< tad_> zookeeper: I know. But some things really should never have been strings, like id= but done's done. As to Mal M'Brin and which apostrophe it uses I just copy and paste because it seems to change so I use what works .. but in [message] I probably did it by hand and should proof-read that I do consitent with elsewhere
20160826 17:11:33< zookeeper> well yes, contrary to what i just said please do use whichever form is already used in names
20160826 17:11:50< zookeeper> those can be mass-replaced separately at some point, if necessary
20160826 17:12:48< celticminstrel> Did anyone have opinions on the two-portrait message?
20160826 17:12:54< tad_> I'll check both for consistency within S05 and for consistency across all scenario. If it needs changing, I'll get with you. Most likely, though, any changes will be my mistakes.
20160826 17:13:00< vultraz> it's fine
20160826 17:13:07< zookeeper> but everything really is a string, there's no avoiding that. side=1 is the same as side="1".
20160826 17:13:11< tad_> celticminstrel: I didn't see the need, but was fine with the idea.
20160826 17:13:34< zookeeper> celticminstrel, neverheard
20160826 17:13:45< tad_> zookeeper: Until equals="yes" fails because you needed boolean_equals="yes" ... :P
20160826 17:14:25< celticminstrel> http://celmin.pwcsite.com/wesnoth/double_message.png
20160826 17:15:01< zookeeper> tad_, well still they're all strings. what the engine does with the string or how it interprets it is another matter.
20160826 17:15:24< vultraz> this is odd..
20160826 17:15:25< zookeeper> celticminstrel, well, doesn't sound like a bad idea to allow that
20160826 17:15:31< vultraz> if i pass -1 to do_game_join manually...
20160826 17:15:39< vultraz> it doesn't crash and passed the first join call
20160826 17:15:50< vultraz> that implies the index provided is invalid..
20160826 17:16:40< vultraz> but why in hell would itb e
20160826 17:17:12< celticminstrel> Huh? Are you questioning the invalidity of -1 as an index?
20160826 17:17:34< vultraz> no, I'm questioning why it does NOT crash if passed -1 :|
20160826 17:17:40< celticminstrel> Oh.
20160826 17:17:40< vultraz> and it DOES if passed 0
20160826 17:18:14< zookeeper> tad_, what's the need for this check? https://git.io/v6xqS is it just to avoid saying "we have merely defeated a specter" if technically it wasn't you but a bandit?
20160826 17:19:21< tad_> Yep. And that was the last change I made .. 4AM last night .. when I was looking at any place where it had inconsistencies.
20160826 17:19:24< vultraz> celticminstrel: thoughts?
20160826 17:20:19< celticminstrel> Where does it crash when passed an invalid index? do_game_join() shouldn't.
20160826 17:20:21< tad_> zookeeper: Of course, I have no problems with backing off any change. ISTM it's better to look for 'em all and remove some than miss some and regret it later.
20160826 17:20:27< celticminstrel> Since it checks the index, right?
20160826 17:21:16-!- JyrkiVesterinen [~JyrkiVest@78-27-114-202.bb.dnainternet.fi] has joined #wesnoth-dev
20160826 17:21:53< vultraz> dunno anymore
20160826 17:22:15< tad_> zookeeper: My feeling is this is the pivotal scenario for the campaign and needs the most attention as a result. Everything before was preparing for this and everything after is followup consequences to the way this scenario played out.
20160826 17:22:24< vultraz> last i checked it had someting to do wit h if(game_info* game = lobby_info_.get_game_by_id(game_id))
20160826 17:22:37< vultraz> but..
20160826 17:22:54< vultraz> if if(do_game_join(get_game_index_from_id(selected_game_id_), false)) is passed with 0 instead of get_game_index_from_id it still crashes
20160826 17:23:09< vultraz> 0 is valid in do_game_join
20160826 17:23:14< vultraz> which means the problem might be in const game_info& game = *lobby_info_.games()[idx];
20160826 17:23:20< celticminstrel> It's only valid if there are games in the list.
20160826 17:23:30< vultraz> implying games() maybe isn't updated yet..
20160826 17:23:31< zookeeper> tad_, yeah all that S05 stuff seems okay
20160826 17:23:57< celticminstrel> If games() isn't updated yet, wouldn't the index check in do_game_join fail?
20160826 17:24:54< zookeeper> tad_, so what's the status of TSG Bug fixes #703?
20160826 17:25:59< tad_> zookeeper: Wow. I'm relieved. It beat me up and I ended up doing a huge logic map because the first few passes I missed some cases and caused bugs when I play-tested a different path through the scene
20160826 17:26:38< zookeeper> tad_, well i didn't examine the logic to verify it or anything, but i'll take your word on it :P
20160826 17:26:56< tad_> zookeeper: As to #703 .. I want to merge me [filter-recall] work into TSG-elven and expect within the week to be pushing a mass-replace for #703
20160826 17:28:23< tad_> zookeeper: Or, if you prefer, I can push TSG-elven up to #703 now and do the store/unstore/debug-modes/filter-recall work as a follow-up
20160826 17:30:15-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Remote host closed the connection]
20160826 17:30:40< zookeeper> tad_, well yes, the smaller your PR's the better
20160826 17:31:09-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160826 17:31:24< tad_> zookeeper: oh, come up .. you really want 60 PRs .. I know .. I'm going overboard with all these niggling little fixes...
20160826 17:32:10< tad_> zookeeper: OK. Once I get a runnable engine again I'll do a full run-through play-test and push what I have to #703.
20160826 17:33:49< tad_> zookeeper: Them I will shelve the cleanup for all the store/unstore mess and move on to getting the DM PR cleared. I can always do a follow-on PR for the cleanups and I suppose the more true bug-fixes the better.
20160826 17:38:45< zookeeper> tad_, regarding the "?" name change... the point of that was originally to make it more clear that it's not really him and that there's more to it, even before you manage to kill it. no one directly references/addresses that unit specifically at any point either, and it doesn't speak.
20160826 17:39:43< zookeeper> it's basically just a mindless specter so should it really have m'brin's name at all?
20160826 17:40:46< tad_> zookeeper: Yes, but it seemed clunky to me being "?" and, in the aftermath conversations we're finally told that **looked** like Mal M'Brin but was a stand-in. What I'd really prefer is some sort of exposition "Taht looks like Mal M'Brin!!!' when we first meet him, I suppose.
20160826 17:42:24< tad_> zookeeper: So, it's not the name= which was really my problem, but the "where did that come from?" I hit when we're told it was not him and we need to proceed underground ("And where did 'underground' come from?")
20160826 17:42:32-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has quit [Quit: End Transmission.]
20160826 17:44:19< zookeeper> ah, actually originally, if you sided with bandits, there was no indication that urza afalas knew that it _wasn't_ m'brin, so the following scenarios were kind of nonsensical because... you had reason to think that you had actually killed m'brin.
20160826 17:45:04< zookeeper> so changing the name and the dialogue when you kill it was to facilitate urza afalas actually knowing that it wasn't the real m'brin
20160826 17:45:58< zookeeper> i'm guessing i intended the "?" name to make someone like him be able to tell the difference more believable, or something
20160826 17:47:14< tad_> zookeeper: Perhaps what's needed is a bit more dialog, earlier on, to make it flow better for both paths.  Adding the name, or making it "?", neither really feel like they address the heart of the issue.
20160826 17:48:17< zookeeper> yeah, why not
20160826 17:48:29< zookeeper> i'm not sure what exactly you mean by "where did that come from?" though
20160826 17:48:32< tad_> zookeeper: I do like the idea that Urza is convinced that IS the malbrin .. but, then, why does he comment about the underground caverns, too?
20160826 17:48:53-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.1 Aria]
20160826 17:49:47< tad_> Ignoring the bandit-path .. Ethiliel tells it it wasn't really Mal M'Brin when we don't have any reason to think it was. Hence I added the name. But your comments make me think the name wasn't the issue .. the real problem is there is no introduction to the idea, at all.
20160826 17:50:47< zookeeper> yeah, the name was supposed to be a subtle introduction to the idea, but apparently it's not working that well? :p
20160826 17:50:55< zookeeper> or maybe it's too subtle
20160826 17:51:38< tad_> zookeeper: the only hint of the confusion is when we kill Gruth (the walking-dead leader) and that seemed OK but still a little "Why would we think that's 'the undead leader' we care about rather than 'some random undead leader'?"
20160826 17:52:09-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has joined #wesnoth-dev
20160826 17:53:08< tad_> How about we (bandits, elves, and Deoran) all think the lich **IS** Mal M'Brin until he dies and add a death-bed exposition he's a dopleganger?
20160826 17:53:21-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160826 17:53:41< zookeeper> hmh, yeah i guess that could work too
20160826 17:53:55-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 17:53:56-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160826 17:54:03< celticminstrel> Bah, XCode just crashed.
20160826 17:54:23-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 17:54:36< zookeeper> funny detail about gruth's death: if the killer is ethiliel, hylas or deoran, then they won't get to say the silly "The master of the undead is defeated!" line, but apparently gerrick is deemed stupid enough because he gets the deliver it in that case.
20160826 17:55:01< tad_> Yep.
20160826 17:55:28< zookeeper> so of all our heroes, only gerrick is as dumb as a common soldier
20160826 17:56:21< zookeeper> maybe just change that event so the dialogue only gets delivered if there's a generic low-level dummy unit available to say that? :P
20160826 17:56:24< tad_> When I'm playing normally, I usually run across the bandits, first, but, before I turn to go for Urza the undead appear. It becomes a toss-up which I kill first, Urza or Gruth.
20160826 17:57:04< tad_> zookeeper: Makes sense. I already changed it so we don't get the message at all if Urza's bandits kill him (as side 2).
20160826 17:59:18-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 258 seconds]
20160826 17:59:35< tad_> zookeeper: So, for me, Urza and Gruth both present a chance to expose the lich as "Mal M'Brin" and, when we finally do meet and/or kill the lich, Ethiliel can recognize he's a doppleganger and point to the caverns and their wards.
20160826 18:00:30< zookeeper> sure
20160826 18:01:22< tad_> zookeeper: But if we sided with the bandits, we don't find out he was an imposter until after we're out of the long march and sir gerrick is left to hold him off while deoran heads to weldyn. That means re-working a bit of Urza's comments in S05 to void mentioniung the caverns we don't know of and cannot enter without Ethiliel's magic
20160826 18:01:46< zookeeper> eh hmm
20160826 18:01:50< celticminstrel> How should side tables compare equal in Lua?
20160826 18:02:10< celticminstrel> Should I compare by pointer, or by side number?
20160826 18:03:04< tad_> celticminstrel: overload operator == and != ???
20160826 18:03:20< tad_> Oh, Lua
20160826 18:03:23< celticminstrel> == and ~- you mean.
20160826 18:03:25< celticminstrel> ~=
20160826 18:04:22< tad_> zookeeper: I'll give it some thought and see if I can come up with some dialog changes which can go either way until we've made a choice, then flow naturally from what we've learned and the choice we've made.
20160826 18:04:50< zookeeper> tad_, that bit seems rather tacky to me. for one, you have to have some kind of explanation for the undead horde of the long march.
20160826 18:05:29< celticminstrel> I guess there'll never be two sides with the same number, so I'll go with that...
20160826 18:05:57< tad_> zookeeper: actually, I agree about that on the bandit-path.
20160826 18:05:58< celticminstrel> Now I need to remember what the metafunction is actually called...
20160826 18:06:11< celticminstrel> Ah, __eq
20160826 18:06:31< tad_> celticminstrel: Actually the engine seems to get VERY upset if your [side] tags are not sequentially numbered from 1..n
20160826 18:06:58< celticminstrel> Even if it were changed to allow sparse numbering, I think we wouldn't allow duplicates.
20160826 18:09:15< zookeeper> tad_, TSG is a somewhat campaign historically. it really wasn't that well-designed originally, there was all sorts of weird crap in it.
20160826 18:09:22< zookeeper> somewhat interesting, i meant
20160826 18:09:37< zookeeper> i think the whole bandit branch was only added afterwards...
20160826 18:09:58< tad_> zookeeper: I found evidence of that in the commit logs.
20160826 18:10:51< tad_> zookeeper: My issues are that it does not flow well, conversationally, through S05 and the bandit-path seems like, as a noob player, "Wow, I should reload and choose the elven path."
20160826 18:11:10< celticminstrel> Why does it seem like that?
20160826 18:11:54< tad_> When I play the elven path, on EASY, I don't bother recruiting or recalling once I hit S05. The four heros are enough by themselves to finish the campaign.
20160826 18:12:51< tad_> But when I take the bandit side, I get my ass kicked and often lose and have to try again.
20160826 18:13:27< zookeeper> yeah it's probably considerably harder than the elvish path even on easy
20160826 18:14:00< celticminstrel> I wish I could get rid of all these linker warnings.
20160826 18:14:07< tad_> zookeeper: When I first played through as a noob I didn't even NOTICE the bandit side. It seemed obvious to go with the elves.
20160826 18:15:40< zookeeper> tad_, in principle i'm perfectly fine with the bandit choice being less obvious, actually harder and more of a "you screwed up" branch, but of course it shouldn't be _that_ much harder.
20160826 18:20:15< tad_> zookeeper: maybe it's that I don't play it enough. to me, it's scary as hell once I'm out of the long march and I regret going with the bandits when I have to try to get sir gerrick to live AT ALL .. and 40 turns for max benefit? You have GOT to be kidding!
20160826 18:20:26< zookeeper> tad_, so you're free to suggest alternatives WRT how to handle m'brin and his doppleganger, but i'd think the clearly easiest path would be to let urza remain aware of it and reveal it only after it has been killed
20160826 18:20:41< zookeeper> is it 40 turns? uh, yeah... that's too long
20160826 18:21:17< vultraz> i really, really can;t figure out this crash
20160826 18:21:26< zookeeper> although it should be practically impossible, and you don't want the maximum to be quite feasible to reach on most playthroughs
20160826 18:21:44< zookeeper> because there should always be a reason to try to last just one last turn longer
20160826 18:22:04< zookeeper> so having the max be unreachable should be okay
20160826 18:25:02-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Ping timeout: 264 seconds]
20160826 18:27:26< zookeeper> (besides, what 40 turns? it looks like beyond turn 22 you get an extra 20 starting gold per turn, without limitation)
20160826 18:28:10-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 18:32:39< vultraz> celticminstrel: I give up
20160826 18:32:41< vultraz> this makes no sense
20160826 18:33:01< vultraz> again, all the evidence seems to indicate it's somewhere in if(game_info* game = lobby_info_.get_game_by_id(game_id)) 
20160826 18:33:12< vultraz> but the the hell that should be caused if an empty config if passed
20160826 18:33:16< vultraz> to join
20160826 18:33:24< vultraz> i have NFI
20160826 18:33:36< celticminstrel> Maybe I'll look into it later.
20160826 18:33:42< celticminstrel> Right now I'm in the middle of some Lua work.
20160826 18:33:53< celticminstrel> Apparently I broke the unit_type metatable earlier.
20160826 18:34:22< celticminstrel> Because I assumed unit types were userdata when they were just a table with a custom metatable.
20160826 18:34:36< celticminstrel> So I'm making them userdata and expanding the metatable.
20160826 18:34:47< celticminstrel> Attacks seem not to work either.
20160826 18:44:16-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160826 19:23:58-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 19:29:25-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 255 seconds]
20160826 19:41:02< mattsc> celticminstrel: uh, in 1.12, do aspects not work in MP in [side][ai] ?
20160826 19:41:17< mattsc> standard aspects
20160826 19:41:49< mattsc> I think I vaguely remember that being the case at some point during some version, but I didn’t know that that is a general bug in the stable series.
20160826 19:42:56< celticminstrel> What?
20160826 19:43:17< celticminstrel> key=value and full composite aspects should both work in 1.12.
20160826 19:43:20< mattsc> Yes, that was my feeling.
20160826 19:43:51< celticminstrel> In MP?
20160826 19:44:30< mattsc> Yes.
20160826 19:44:35< mattsc> I am checking out this: https://forums.wesnoth.org/viewtopic.php?f=21&t=44534
20160826 19:44:55< mattsc> And it looks like the OP did everything correctly, but the aggression value does not take effect.
20160826 19:45:17< mattsc> I just added a [modify_side][ai] to it and then it works.
20160826 19:45:26< mattsc> That’s pretty bad if that is a general bug!
20160826 19:45:41-!- irker651 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160826 19:45:48< celticminstrel> Gimme a minute to launch Firefox and check the thread.
20160826 19:48:32< mattsc> celticminstrel: hmm, I just checked with Dark Forecast and it seems to work there
20160826 19:49:45< celticminstrel> It's not clear to me whether he knows that he needs to restart Wesnoth (or press F5 at the title screen) to make changes to the WML take effect.
20160826 19:49:55< celticminstrel> That aside though, I can't see anything wrong with what he's doing.
20160826 19:50:32< mattsc> Well, I do know that though and I started his scenario and checked aggression for the AI sides.  It’s at 0.4.
20160826 19:50:43< celticminstrel> I see.
20160826 19:51:01-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has quit [Ping timeout: 252 seconds]
20160826 19:52:52< mattsc> I cannot see anything wrong with the code.
20160826 19:52:58< celticminstrel> Nor I.
20160826 19:53:12< celticminstrel> No console warnings or anything either?
20160826 19:55:09-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity]
20160826 19:56:55< mattsc> something complaining about unknown attribute “unrenameable” (sic) and “no_recruit”, but not for those sides
20160826 19:59:27-!- atarocch [~atarocch@93.56.160.28] has joined #wesnoth-dev
20160826 20:00:32< mattsc> I don’t understand …  Esp. since it works in Dark forecast.
20160826 20:02:58< mattsc> Oh, wow, once I copy the Dark Forecast side definition into this guy’s scenario, it works there too ...
20160826 20:04:47< celticminstrel> So, what's different between the two?
20160826 20:07:32< mattsc> I’m currently trying to figure that out.
20160826 20:07:43< mattsc> I mean, which of the difference triggers this.
20160826 20:12:00-!- Shiki [~Shiki@141.39.226.227] has joined #wesnoth-dev
20160826 20:13:32-!- JyrkiVesterinen [~JyrkiVest@78-27-114-202.bb.dnainternet.fi] has quit [Quit: .]
20160826 20:16:54< mattsc> Wow …
20160826 20:17:26< mattsc> So, Dark Forecast has a line ‘allow_player=no’ in the side definition, Zombie Survival 2 does not.
20160826 20:17:42< mattsc> And that’s the difference that make the aspect be set correctly in the former, and not in the latter.
20160826 20:17:59< celticminstrel> That's kinda odd...
20160826 20:18:01< mattsc> Must be something in the way the sides get set up one way or the other.
20160826 20:18:04-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 20:18:06< mattsc> Indeed.
20160826 20:18:19< mattsc> At least I feel a little better about us not knowing about this.
20160826 20:18:38< mattsc> And there’s an easy workaround.
20160826 20:22:39-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 258 seconds]
20160826 20:33:28-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has joined #wesnoth-dev
20160826 20:35:27< celticminstrel> The debugger keeps crashing XCode... :(
20160826 20:42:17-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 20:55:59-!- celmin [~celticmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160826 20:57:47< celmin> Question!
20160826 20:57:56< celmin> Someone interested in Lua API stuff.
20160826 20:58:37< celmin> In the Lua view of a unit type (obtained from wesnoth.unit_types), I am adding a field "variations" which contains the unit's variations as further unit type views.
20160826 20:59:53< celmin> I want sorcerer.variations.male to refer to the male variations and sorcerer.variations.female to refer to the female variation, if present.
20160826 21:00:45< celmin> The question is, should pairs(sorcerer.variations) loop over those two, or do nothing? It's easiest if I make it only cover non-gender variations, like with the walking corpse.
20160826 21:01:11< celmin> (I also have sorcerer.variations.base referring back to sorcerer, though that's more for internal use.)
20160826 21:01:55< celmin> I also made #sorcerer.variations return 2 (and #zombie.variations return 11), but just now I realized that that's actually kind of weird since # is normally for array-like tables.
20160826 21:02:08< celmin> So I might revert that.
20160826 21:06:25-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160826 21:11:39-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160826 21:28:13-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160826 21:31:33-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 276 seconds]
20160826 21:31:33-!- wedge010 is now known as wedge009
20160826 21:35:04-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has quit [Ping timeout: 244 seconds]
20160826 21:37:26-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20160826 21:38:55-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.]
20160826 21:39:53-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has joined #wesnoth-dev
20160826 21:57:36 * celmin tests for aeth.
20160826 21:59:15-!- Kwandulin [~Miranda@p200300760F6F2FECB14B9DF3D1BF7A9C.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160826 22:41:11-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160826 22:49:25< celmin> I wonder if there's other info that could be added to the gamestate inspector.
20160826 22:49:58-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20160826 22:49:58-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Read error: Connection reset by peer]
20160826 23:00:50-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has joined #wesnoth-dev
20160826 23:07:16-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection]
20160826 23:16:05-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160826 23:17:34-!- RatArmy [~RatArmy@om126229092047.12.openmobile.ne.jp] has quit [Ping timeout: 255 seconds]
20160826 23:24:05-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160826 23:27:22-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160826 23:27:43-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160826 23:32:17-!- stikonas_ is now known as stikonas
20160826 23:32:44-!- Shiki [~Shiki@141.39.226.227] has quit [Quit: Verlassend]
20160826 23:38:02-!- Lohengramm [sid1929@gateway/web/irccloud.com/x-gktygytvpdchahen] has quit [Ping timeout: 244 seconds]
20160826 23:40:03-!- ancestral [~ancestral@67-4-224-82.mpls.qwest.net] has quit [Quit: i go nstuf kthxbai]
20160826 23:40:29-!- Lohengramm [sid1929@gateway/web/irccloud.com/x-fndpcoyxtcbslavt] has joined #wesnoth-dev
20160826 23:43:41-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 250 seconds]
20160826 23:45:44-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
--- Log closed Sat Aug 27 00:00:49 2016