--- Log opened Sat Sep 03 00:00:10 2016
20160903 00:05:56-!- fabi [~fabi@176.0.22.31] has joined #wesnoth-dev
20160903 00:12:01< fabi> hi
20160903 00:12:19< fabi> Aginor: ping
20160903 00:13:03-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 240 seconds]
20160903 00:24:33-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Remote host closed the connection]
20160903 00:34:09< vultraz> bah
20160903 00:34:20< vultraz> using ints has caused all these Wsigncompare warnings
20160903 00:34:24< vultraz> might just go back to size_t..
20160903 00:36:36< vultraz> ah, no, I got it
20160903 00:37:24-!- Bonobo [~Bonobo@2001:44b8:254:3200:4485:c528:47c6:870a] has joined #wesnoth-dev
20160903 00:38:50< vultraz> blah
20160903 00:38:55< vultraz> yeah i broke mapgen :|
20160903 00:39:09< vultraz> trying to convert the generator job to use the struct
20160903 00:42:00 * vultraz curses soundly
20160903 00:45:12< vultraz> perhaps I need a copy constructor for the struct
20160903 00:45:19< vultraz> celmin: what is the syntax for a copy constructor?
20160903 00:45:35< celmin> For declaring it?
20160903 00:45:39< vultraz> yes
20160903 00:45:51< celmin> (Your generator_data doesn't need one, though, unless you added something since I last saw it.)
20160903 00:45:59< celmin> class_name(const class_name&);
20160903 00:46:08< celmin> You can put a parameter name after the &, of course.
20160903 00:46:11< vultraz> i haven't added anything
20160903 00:46:29< vultraz> but for some reason mapgen is totally broken now :/
20160903 00:46:49< vultraz> was working before i made the job use the struct
20160903 00:47:15< celmin> Are you passing by reference or by value?
20160903 00:47:48< vultraz> value
20160903 00:47:50< vultraz> hmm..
20160903 00:47:54< vultraz> actually i think i see the problem
20160903 00:48:35< vultraz> mass find-replace error
20160903 00:49:37< vultraz> ahh seems to be fixed
20160903 00:50:50< vultraz> well, partly
20160903 00:54:36< vultraz> hmmm
20160903 00:54:49< vultraz> mp create seems to not remember the generator settings...
20160903 00:55:11< celmin> Generator?
20160903 00:55:15< celmin> Oh, map generator.
20160903 00:55:25< celmin> Would probably be a lot of work to change that though.
20160903 00:55:41< celmin> It doesn't remember in GUI1 either, right?
20160903 00:55:43< vultraz> do you have a second to test something
20160903 00:55:56< celmin> I'd say that even if you want to do it you should probably make it a separate commit.
20160903 00:56:05< vultraz> right
20160903 00:56:07< celmin> What do you want me to test? My build is currently broken though.
20160903 00:56:41< vultraz> but can you go to mp create, select random map, change an obvious setting that youcan see in the minimap (like setting the landform to Island), and click Regenerate
20160903 00:56:44< vultraz> a few times
20160903 00:56:49< vultraz> see if it's consistently an island
20160903 00:56:57< celmin> Maybe?
20160903 00:57:00-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160903 01:00:19< vultraz> hmm.. some values were passed to the generator job that weren't the data values..
20160903 01:01:12-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 250 seconds]
20160903 01:02:21< vultraz> anyway, if you could test that that'd be really useful
20160903 01:02:36< celmin> Can it be on 1.12?
20160903 01:02:51< celmin> Or does it have to be 1.13?
20160903 01:03:20< vultraz> 1.13
20160903 01:03:30< vultraz> i need to know if my changes have changed the behavior of something
20160903 01:04:00< celmin> Okay, um...
20160903 01:04:27< celmin> Like I said, my build is currently broken, so I'm not sure if I can do it right now.
20160903 01:05:20< vultraz> ok
20160903 01:06:01< vultraz> ill deal with it
20160903 01:06:49< vultraz> hmm
20160903 01:07:09< vultraz> looks like that's the behavior..
20160903 01:08:17< vultraz> in 1.12 the settings are remembered...
20160903 01:08:50< vultraz> in gui1 they are to
20160903 01:08:52< vultraz> o
20160903 01:08:55< vultraz> ok so this is an issue in Create
20160903 01:18:53-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160903 01:20:28< tad_> OK. So now I'm really confused. Perhaps I don't understand but it looks to me like the game you're playing when you start a new scenario by pressing End Scenario is quite a bit different than the one you're playing if you Load the scenario start. [ai] change, different factions. different translations .. and that's just what I see at a glance
20160903 01:21:26< celmin> Huh? Different translations?
20160903 01:22:02< tad_> Yep. And big changes in the [ai] .. in this case a lot of [goal] for Ethiliel gets deleted when you Load.
20160903 01:23:03< tad_> I'm trying to track down where "Undead" comes from instead of "Bandits" for user-team-name and decided maybe it's saved wrong so I diff'd the -Auto-Save1 from End Scenario against the same from after Load and LOTS DIFFERENT!!!!
20160903 01:24:06< tad_> I have been whacking the C++ and the best I can do is set all user_team_name to "Tad" but cannot for the life of me see where .. only if you press Load .. does it change to "Bandits"
20160903 01:24:44< celmin> Goals being deleted when you load might mean that they failed to be saved...
20160903 01:24:55< celmin> Which could in theory indicate a bug from the AI refactor...
20160903 01:25:15< celmin> …though… if it's a scenario start save, that might not apply...
20160903 01:25:44< tad_> So the save-the-scenario-start logic is VERY buggy. Changes the translations. Loses AI. Changes factions. Changes the loging player (human) name.
20160903 01:26:56< tad_> What I mean is it saves a buggy file, does not use the saved buggy file and playes the next scenario ubut it's not in the saved-file so if you quit and come back it's missing a lot of stiff???
20160903 01:27:13-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev
20160903 01:27:13< tad_> Because the turn 1 auto-save is VERY different
20160903 01:30:39< tad_> OTOH perhaps the auto-save is wrong for the play-through and corrected when you Load. Because that's the auto-save which has "Bandits" in it. Maybe it's garbage which should not have been saved the first time and when I used Load it wasn't there so it wasn't saved?
20160903 01:34:10< vultraz> gahhhhhh
20160903 01:34:11< vultraz> this.
20160903 01:34:13< vultraz> CODE
20160903 01:35:01< tad_> Anyway, I can see where the user_side_name is constructed (and if I whack it there I see the effect) but I cannot for the life of me see where it's set when we play-through.
20160903 01:35:59< tad_> It's correct in the .cfg and nothing about it in the scenario-start so must come from the .cfg (can't see where but must) so I'm leaning to this being part of the memory leaks I think I'm seeing if I play for hours.
20160903 01:36:35< celmin> When you say it saves a buggy file… are you examining the saved game file in a text editor?
20160903 01:37:06< tad_> I gunzip before and after and diff
20160903 01:37:15< tad_> copies
20160903 01:39:38< tad_> (1) cleared .cache .config and .local/share/../1.13 (2) played through to objectives for S04 (3) copied auto-save for turn 1 of S04. (4) killed process (5) start game, Load S04 scen-start (6) copied auto-save for S04 turn 1. gunzip and compare
20160903 01:40:42< tad_> IMHO the auto-save should be identical or very nearly identical. [ai] going missing and other stuff seems too different
20160903 01:40:50< celmin> (If gunzip is a pain you can tell Wesnoth not to compress them in prefs. Just noting.)
20160903 01:41:14< celmin> I see.
20160903 01:41:30< celmin> Indeed, it should be identical if you load the scenario start save...
20160903 01:41:49< tad_> Note, it's not ALL the [ai] gone missing .. just the specifics targetting Ehiliel [goal] and such
20160903 01:42:08-!- irker925 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160903 01:42:08< irker925> wesnoth: Charles Dang wesnoth:master f66884cf018a / src/gui/dialogs/editor/ (edit_side.cpp edit_side.hpp): Edit Side: simplify group setup code https://github.com/wesnoth/wesnoth/commit/f66884cf018a798a1ae69c359ca99a12a4133e3e
20160903 01:42:09< irker925> wesnoth: Charles Dang wesnoth:master 0d07c1a8151b / src/generators/ (4 files): Default Map Generator: refactored generator setting parameters into a struct https://github.com/wesnoth/wesnoth/commit/0d07c1a8151be4ad12c0812c5e96e64c8a08a6f9
20160903 01:42:10< irker925> wesnoth: Charles Dang wesnoth:master 42423a4a57c4 / src/gui/widgets/slider.cpp: Slider: removed spurious calls to set_item_position when setting min/max value https://github.com/wesnoth/wesnoth/commit/42423a4a57c40742e87933b3b9dfd5302e811a7f
20160903 01:42:12< irker925> wesnoth: Charles Dang wesnoth:master 56e8222bb096 / / (8 files in 6 dirs): Converted Generator Settings dialog to GUI2 https://github.com/wesnoth/wesnoth/commit/56e8222bb096e2dc95d142011ded4021ba5c2f9e
20160903 01:42:14< vultraz> DONE
20160903 01:42:21< vultraz> JESUS
20160903 01:42:22< vultraz> 8 hours
20160903 01:42:34< celmin> tad_: Perhaps, all [goal] and nothing else?
20160903 01:42:48< celmin> Unless these [goal] are actually removed in prestart or something.
20160903 01:42:51< vultraz> a thousand curses upon whoever did the mapgen code
20160903 01:43:00< celmin> No, I guess that's not relevant.
20160903 01:43:09< vultraz> a thousand curses, I say!
20160903 01:43:14< celmin> Because it's the difference between the auto-save and the auto-save.
20160903 01:43:27< tad_> A lot of other changes.  Scope name for translations. Player login name. user_team_name .. I'd have to do a list but it's large
20160903 01:43:52< celmin> Scope name? Meaning textdomain, or something else?
20160903 01:44:06-!- Sirp [~Sirp@u17402953.onlinehome-server.com] has quit [Ping timeout: 250 seconds]
20160903 01:44:07< tad_> right .. textdomain .. forgot the name
20160903 01:44:54< tad_> I have a wife wanting a midnight wally-world run and am getting it from several directions and want to gripe about this before I forget more :)
20160903 01:45:38< mattsc> Ooo, you live in a place where wally-world is open at midnight!
20160903 01:46:12-!- Sirp [~Sirp@u17402953.onlinehome-server.com] has joined #wesnoth-dev
20160903 01:46:24 * mattsc mattsc just got the first new CA in … ages added to the default AI (locally)
20160903 01:46:35< tad_> We have our choice of 4 in town. Eat your heart out. Of course, 2 of the four are "Bring your Uzi after 10PM" 24hr operations
20160903 01:46:36< mattsc> I better figure out how to do PRs now.
20160903 01:47:10< mattsc> I see; I once was at a BK where the clerks where behind bullet-proof glass
20160903 01:47:17< tad_> mattsc: checkout -b or whatever to make a subject branch for just the changes. push them to your github fork.  press Compare and Create PR
20160903 01:47:35< vultraz> . . . 
20160903 01:47:40 * vultraz flips
20160903 01:47:55< vultraz> now mapgen doesn't work properly even though I TESTED IT
20160903 01:48:11< mattsc> tad_: thanks, yeah, I know it’s something like that; I just haven’t done it in so long I need to look up the exact steps
20160903 01:48:14< tad_> welcome to wesnoth development ...
20160903 01:48:46< vultraz> ;_;
20160903 01:48:53< tad_> ok. gtg. I'll get back to trying to figure out what's up with user_team_name once the wife is placated.
20160903 01:48:57-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160903 01:50:33< vultraz> celmin: don't forget to update xcode
20160903 01:55:03< celmin> Oh right...
20160903 01:55:39< celmin> mattsc: What do you need a PR for?
20160903 01:55:52 * celmin curiosity
20160903 01:56:00< mattsc> Add the hig-xp_attack CA to the default AI
20160903 01:56:03< mattsc> *Adding
20160903 01:56:22< mattsc> and *high   (sigh)
20160903 01:56:29< irker925> wesnoth: Charles Dang wesnoth:master 50e09bab3035 / src/generators/default_map_generator.cpp: Default Map Generator: fixup  0d07c1a8151b (island size not being considered) https://github.com/wesnoth/wesnoth/commit/50e09bab30358bc8811952d927c1e954be9b3705
20160903 01:56:35< vultraz> seems to fix
20160903 01:56:45< vultraz> doesn't fix the *@&*(# map sometimes not regenerating though
20160903 01:56:49< vultraz> bc not enough space
20160903 01:56:53< vultraz> or some crap 
20160903 01:57:02< vultraz> will need to look later
20160903 01:57:38< mattsc> celmin: I could just commit it, but since it’s a 400-line piece of Lua that’s supposed to go into the default AI, I figured I’d ask for people’s opinions first
20160903 01:57:56< celmin> Always a good idea.
20160903 01:59:28< mattsc> I just noticed that I still wanted to add a very minor bit to the rating function though
20160903 02:00:53-!- travis-ci [~travis-ci@ec2-54-159-8-120.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 02:00:54< travis-ci> wesnoth/wesnoth#10682 (master - 56e8222 : Charles Dang): The build failed.
20160903 02:00:54< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157248069
20160903 02:00:54-!- travis-ci [~travis-ci@ec2-54-159-8-120.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 02:02:30< celmin> Hmm, I wonder if the other map generators have a user_config function...
20160903 02:03:27< celmin> vultraz: It'd be great if you stopped reflexively adding new dialogs to the exclusion list.
20160903 02:03:38< celmin> The random map settings dialog is trivially testable.
20160903 02:04:11< vultraz> I rely on you to add the tests :)
20160903 02:05:25< celmin> :(
20160903 02:10:14-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has joined #wesnoth-dev
20160903 02:10:16< vultraz> have discovered why the settings aren't preserved through clicks of the regenerate button in Create
20160903 02:10:20< vultraz> now to figure out a good fix
20160903 02:16:42-!- travis-ci [~travis-ci@ec2-54-159-8-120.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 02:16:43< travis-ci> wesnoth/wesnoth#10683 (master - 50e09ba : Charles Dang): The build failed.
20160903 02:16:43< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157249349
20160903 02:16:43-!- travis-ci [~travis-ci@ec2-54-159-8-120.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 02:22:32< irker925> wesnoth: Charles Dang wesnoth:master 8c65e49f7a19 / src/gui/dialogs/editor/generator_settings.cpp: Generator Settings: specify class name in bind https://github.com/wesnoth/wesnoth/commit/8c65e49f7a197c26c9aea869aae02f376eb8959e
20160903 02:24:34< celmin> Pretty sure it wouldn't compile without that.
20160903 02:25:00< vultraz> does for me
20160903 02:25:11< celmin> What's your language standard setting?
20160903 02:25:21< celmin> Ah, C++11...
20160903 02:25:22< vultraz> c++11
20160903 02:25:24< celmin> So not gnu++11...
20160903 02:25:52< vultraz> -std=c++11
20160903 02:25:56< celmin> Yeah.
20160903 02:26:03< celmin> I looked at the CB project and saw that.
20160903 02:26:17< celmin> I wonder if CB has improved drastically since I last tried it.
20160903 02:26:47< celmin> Though I guess that CB project is currently non-portable.
20160903 02:27:04< celmin> I wonder if it's possible for a CB project to be portable.
20160903 02:34:59< vultraz> hmm
20160903 02:35:18< vultraz> trying to figure out what i did that makes islands get progressively smaller..
20160903 02:35:29< vultraz> until there's not enough land..
20160903 02:37:07< vultraz> it's now the opposite problem of 50e09bab3035  :|
20160903 02:42:50< vultraz> data_.island_size seems to get huge...
20160903 02:42:54< vultraz> but that shouldn't happen :/
20160903 02:43:11< vultraz> i mean, now the job uses the data_ member instead of a custom paremeter..
20160903 02:43:28< vultraz> but I set the data member to what the parameter would have been...
20160903 02:46:27< vultraz> idgi :|
20160903 02:47:16< celmin> I assume data_.island_size is initialized?
20160903 02:47:31< vultraz> yes
20160903 02:48:20< vultraz> for some reason the value's always going 0, 12, 5, 20...
20160903 02:48:23< vultraz> blah ill figure it out later
20160903 03:16:31-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 03:16:32< travis-ci> wesnoth/wesnoth#10684 (master - 8c65e49 : Charles Dang): The build has errored.
20160903 03:16:32< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157251558
20160903 03:16:32-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 03:22:57< irker925> wesnoth: mattsc wesnoth:ai_high_xp_attack 6950ee1dd08a / / (5 files in 5 dirs): New high_xp_attack candidate action for default AI https://github.com/wesnoth/wesnoth/commit/6950ee1dd08a429394f7b0c38340f298c344d57f
20160903 03:24:31< mattsc> Hmm ...
20160903 03:24:49< mattsc> Didn’t know that irker reports on this even if it’s in a branch
20160903 03:25:00< celmin> Yup, it does.
20160903 03:26:16< celmin> A CA score can vary between evaluations on the same turn, right?
20160903 03:26:55< mattsc> absolutely
20160903 03:27:14< Aginor> fabi: pong
20160903 03:27:48< Aginor> fabi: I've been using it mainly in linux, but Ive tried it in windows 10 as well
20160903 03:28:04< celmin> I think get_units() is faster than a filter because the filter has a specialization for when it's empty...
20160903 03:28:49< celmin> Either way I'm pretty sure it's looping over all units though.
20160903 03:29:27< celmin> I guess get_child(…, "attack") is meant to check if it has attacks?
20160903 03:29:37< celmin> Any reason you can't do #u.attacks  > 0?
20160903 03:29:42< Aginor> hey, do we allow force-push on master?
20160903 03:29:44< celmin> u -> unit
20160903 03:29:47< celmin> Aginor: No, why?
20160903 03:29:56< Aginor> just checking
20160903 03:30:15< Aginor> so if I were to do push --force, github would reject it?
20160903 03:30:43< mattsc> celmin: yes, that’s it’s purpose
20160903 03:30:43-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160903 03:30:52< mattsc> and #u.attacks works?  I didn’t know that.
20160903 03:30:55< celmin> Aginor: Yes.
20160903 03:30:59< celmin> mattsc: Should work, yeah.
20160903 03:31:05< mattsc> okay, cool
20160903 03:31:16< Aginor> celmin: ok, grat
20160903 03:31:20< Aginor> great even
20160903 03:31:45< mattsc> celmin: you’re commenting already before I’ve made it a PR.  :)
20160903 03:31:53< celmin> Is that okay?
20160903 03:31:59< mattsc> Of course
20160903 03:32:11< mattsc> Just that I am not ready to reply right now; I have to be afk for a little.
20160903 03:32:44< mattsc> (I didn’t know/remember that irker reports on branches, and was going to do the PR a little later)
20160903 03:33:06< mattsc> In any case, I am going to be back later (and I read the logs anyway, in case I am not tonight any more)
20160903 03:40:14< tad_> Huh? #define WRN_NG LOG_STREAM(warn, log_engine) \n #define ERR_NG LOG_STREAM(warn, log_engine)
20160903 03:40:21< celmin> ?
20160903 03:40:34< tad_> error logging is warning level?
20160903 03:40:43< celmin> …oh wow.
20160903 03:41:24< tad_> Well, warning level works for me so it's jsut "Huh?"
20160903 03:41:36< celmin> Yeah, that should probably be changed.
20160903 03:44:31< tad_> I'll do a PR on it if I change any actual changes.  team.cpp   Back to trying to grok why it's not loading user_team_name when coming from End Scenario.
20160903 03:46:49-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160903 03:52:12< mattsc> celmin, zookeeper, all: now the new high_xp_attack AI candidate action is in a pull request: https://github.com/wesnoth/wesnoth/pull/764
20160903 03:52:36< mattsc> Note specifically that there is a test scenario for checking out what this does.
20160903 03:56:43-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160903 04:00:28< celmin> hp_chance[i] is the chance of hp reaching i?
20160903 04:01:24< celmin> I wonder whether that's an actual table, or a userdata/table with __index...
20160903 04:01:56< celmin> Latter seems preferable, but from what I've heard around the Monte Carlo thing, I wouldn't be surprised to hear it's a plain array.
20160903 04:02:04< mattsc> It’s the chance (percentage) that the unit ends up with i HP after the combat; I think that’s what you’re saying, yes
20160903 04:02:15< celmin> Yeah.
20160903 04:02:41< celmin> Percentage is key - it'd be perfectly normal for the chance to be in the range [0.0,1.0].
20160903 04:03:08-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160903 04:03:16< celmin> …wait, it's compared to aggression, which generally is in that range, isn't it?
20160903 04:03:23< tad_> celmin: PR 765 for that silly #define 
20160903 04:04:09< celmin> I was thinking I'd search LOG_STREAM to see if there were any other similar.
20160903 04:04:48< tad_> Did that. Only place for err_ng
20160903 04:04:51 * celmin 's comments on the AI PR are actually more "Hmm, I didn't know about this" things.
20160903 04:04:56< mattsc> celmin: dbms say this about hp_chance: lua_var is of type table, value table: 0x7f9fc3514ad0, length 38, but no WML table
20160903 04:05:04-!- travis-ci [~travis-ci@ec2-54-159-8-120.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 04:05:05< travis-ci> wesnoth/wesnoth#10685 (ai_high_xp_attack - f7d06f6 : Lari Nieminen): The build has errored.
20160903 04:05:05< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157255886
20160903 04:05:05-!- travis-ci [~travis-ci@ec2-54-159-8-120.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 04:05:10< celmin> Lenght 38, so that means it really is a plain array.
20160903 04:05:14< celmin> ^Length
20160903 04:05:19 * mattsc appreciates celmin’s comments
20160903 04:05:32< celmin> If it were sparse, it'd probably show length 0.
20160903 04:05:41< celmin> Or say it's a userdata rather than a table.
20160903 04:05:54< celmin> So is it in the range [0.0,1.0] or in the range [0,100]?
20160903 04:06:12< mattsc> celmin: [0.0, 1.0], adding up to unity
20160903 04:06:15< celmin> Okay.
20160903 04:06:25< celmin> When you said "percentage" I assumed the latter.
20160903 04:06:40< mattsc> Oh, yes, sorry.  I should have said fraction ...
20160903 04:07:27< celmin> 0.001 penalty for occupied attack hex, huh...
20160903 04:07:40< mattsc> There’s a reason for that ...
20160903 04:07:44< celmin> Minor indeed.
20160903 04:08:21< mattsc> Don’t want to do another attack if this one is better; however, if another one is equal, do that one instead, so that you don’t use up MP for moving a unit out of the way.
20160903 04:08:42< celmin> Whoa, what's this ai_helper function do...
20160903 04:08:59< celmin> movefull_outofway_stopunit
20160903 04:09:17< mattsc> lot’s of things, but exactly what it says
20160903 04:09:47< mattsc> do a full move; but move unit-in-way out of the way first if needed; if no move possible, do an ai.stop_unit instead
20160903 04:10:09< celmin> stop_unit on the main unit, not the unit-in-way?
20160903 04:10:14< mattsc> yes
20160903 04:10:23< mattsc> the move out of the way is definitely a partial move
20160903 04:10:35< celmin> I'd expect that, yeah.
20160903 04:10:45-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160903 04:11:03< mattsc> It has to be checked that the move out of the way is possible for that, otherwise you get at the very least a warning, more likely an error
20160903 04:11:32< celmin> Your max_score seems to be wrong in the config.
20160903 04:11:39< celmin> Missing braces around the macro invocation.
20160903 04:12:02< celmin> I wonder what the game does when it finds a non-numeric max_score.
20160903 04:12:32< mattsc> Oops, yes indeed
20160903 04:12:34< celmin> I think the name is unused for Lua CAs, though maybe it doesn't hurt to have it anyway.
20160903 04:13:12-!- JyrkiVesterinen [~JyrkiVest@87-100-157-208.bb.dnainternet.fi] has joined #wesnoth-dev
20160903 04:13:18< mattsc> Well, the AI works, so it must be treating it as non-existent (rather than 0)
20160903 04:14:09< celmin> Yeah, probably assigns a max score of MAX_INT or similar.
20160903 04:14:22< celmin> Something very large, in other words.
20160903 04:15:29< mattsc> right
20160903 04:16:07< mattsc> celmin: I have to be off do some household chores now.  I’ll likely just fall into bed after that.  I’ll try to catch up tomorrow.
20160903 04:16:18< celmin> I actually have nothing more to say.
20160903 04:16:31< mattsc> Cool; thanks for doing that so quickly!
20160903 04:16:42< irker925> wesnoth: Gregory A Lundberg wesnoth:master 0696da44c97a / src/team.cpp: Fix bug: Error typically means error https://github.com/wesnoth/wesnoth/commit/0696da44c97a8bcae9a037d697f0cba279cb91df
20160903 04:16:44< irker925> wesnoth: Celtic Minstrel wesnoth:master e1e2d43b9ebc / src/team.cpp: Merge pull request #765 from GregoryLundberg/GL_error_means_error https://github.com/wesnoth/wesnoth/commit/e1e2d43b9ebccd1a1df0b9d15843600501806cb5
20160903 04:17:21< mattsc> celmin: So IIRC, the missing braces and simplifying checking whether the unit has attacks is all I need to do so far?
20160903 04:17:33< celmin> Yeah.
20160903 04:17:50< celmin> Using #unit.attacks should be faster since it doesn't need to convert the unit to a config, I guess.
20160903 04:18:20< mattsc> Oh, yeah, if that works (not that I doubt it), it should definitely be faster.
20160903 04:18:28< celmin> Speaking of converting the unit to a config, I think it does this anew every time you access .__cfg
20160903 04:18:44< celmin> So if you needed to do it a lot I guess it'd be faster to assign .__cfg to a variable first.
20160903 04:18:49< mattsc> Yes, I think so.  And it’s horribly slow.
20160903 04:18:56< celmin> (Just a random observation, has nothing to do with your new AI.)
20160903 04:19:06< mattsc> That (what you suggest there) is what I do in my Fred AI.
20160903 04:19:16< celmin> I see.
20160903 04:35:52< vultraz> I think I understand the generator issue I introduced
20160903 04:36:37< vultraz> the data is reused every time the map is generated
20160903 04:36:42< vultraz> and modified each time
20160903 04:38:45< celmin> Understanding is a good start!
20160903 04:38:52< celmin> Provided it's correct understandint, I guess.
20160903 04:39:40< vultraz> so what was happening before was temp variables were passed to the generator job 
20160903 04:42:05< vultraz> then the generator job takes those temp variables + a copy of the data...
20160903 04:42:46< vultraz> I think I could fix this by making and modifying a copy of the class data
20160903 04:43:09< vultraz> and passing that by const reference to the generator job
20160903 04:43:40< celmin> I don't really get it, but if you think it'll work, then try it!
20160903 04:46:05< vultraz> it's also safer, since then the internal settings are never modified except by the settings dialog
20160903 04:46:34< JyrkiVesterinen> Hmm, in my "debug MP test failures" branch the test isn't getting even as far as before.
20160903 04:46:47< JyrkiVesterinen> Host logs on and enters lobby mode.
20160903 04:46:56< JyrkiVesterinen> Let's see if it gets any further than that...
20160903 04:49:43< JyrkiVesterinen> "running lobby loop" is the final output from the host.
20160903 04:49:55< JyrkiVesterinen> The host is likely crashing because of the logging I added.
20160903 04:50:03< JyrkiVesterinen> DBG_MP << "current plugin context: " << ui.get_plugins_context()->get_name() << "\n";
20160903 04:50:22< celmin> So you think the context might be null
20160903 04:50:23< JyrkiVesterinen> I should have checked the plugins_context for null pointers...
20160903 04:50:47< JyrkiVesterinen> The context is never null in the guest, but I didn't check the host before pushing.
20160903 04:52:40-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 04:52:41< travis-ci> wesnoth/wesnoth#10686 (ai_high_xp_attack - 6950ee1 : mattsc): The build has errored.
20160903 04:52:41< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157256617
20160903 04:52:41-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 04:58:01< vultraz> yyyesss this works :D
20160903 04:58:28< vultraz> amazing what taking a step back can do
20160903 05:00:37< JyrkiVesterinen> Huh? In my local test the plugins_context is not null.
20160903 05:00:50< JyrkiVesterinen> Yet another difference between Travis and my computer.
20160903 05:00:57< JyrkiVesterinen> And it's the exact same code.
20160903 05:02:05< JyrkiVesterinen> It doesn't really matter in this case because the guest is what causes tests to fail.
20160903 05:02:19< JyrkiVesterinen> But it's a potentially useful observation anyway.
20160903 05:02:41-!- Kwandulin [~Miranda@p200300760F4241433074E3EB233DE29F.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160903 05:07:43< irker925> wesnoth: Charles Dang wesnoth:master 501c50fcea8b / src/generators/ (4 files): Default Map Generator: some refactoring to fix issues with map generation introd https://github.com/wesnoth/wesnoth/commit/501c50fcea8b644a7aeb292458e7b2fb86a994a4
20160903 05:11:00< irker925> wesnoth: Jyrki Vesterinen wesnoth:debug-multiplayer-test-failures 1d973b702026 / src/game_initialization/multiplayer.cpp: Attempted fix for the host crashing (?) https://github.com/wesnoth/wesnoth/commit/1d973b702026d2add78ec78d3585abf8501172fe
20160903 05:12:44< JyrkiVesterinen> Thinking about it... could it be that the plugins_context is also null in the guest between scenarios?
20160903 05:13:18< vultraz> there are two plugin contexts, aren't there
20160903 05:13:20-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160903 05:13:39< JyrkiVesterinen> That would explain why the script apparently doesn't exit the loop in lines 128-131.
20160903 05:13:44< vultraz> the gui2 dialog one, and the gui1 one mp::ui
20160903 05:13:52< JyrkiVesterinen> https://github.com/wesnoth/wesnoth/blob/1d973b702026d2add78ec78d3585abf8501172fe/join.lua#L128-L131
20160903 05:13:58< celmin> vultraz: Currently he's only dealing with the GUI1 one, I think.
20160903 05:14:13< vultraz> but the mp faction select dialog is gui2
20160903 05:14:22< celmin> True.
20160903 05:14:25< vultraz> meaning the general plugins context he added gets activated
20160903 05:14:49< JyrkiVesterinen> If the plugin isn't getting control even once while the guest is showing the lobby between scenarios, then it's going to remain in that loop even for the second scenario.
20160903 05:15:03< JyrkiVesterinen> And then it would spin in the second loop when the game is over.
20160903 05:15:10< celmin> Makese sense.
20160903 05:15:14< celmin> ^makes
20160903 05:15:23< celmin> What's this lobby between scenarios?
20160903 05:15:33< celmin> I thought there was just a loading screen and that's it.
20160903 05:16:28< JyrkiVesterinen> The lobby is opened here: https://github.com/wesnoth/wesnoth/blob/c3220676d76349cfdb2a50600fa4513ec6408179/src/game_initialization/playcampaign.cpp#L361-L370
20160903 05:17:32-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 240 seconds]
20160903 05:17:51< celmin> And that happens between scenarios?
20160903 05:18:04< JyrkiVesterinen> At least in my local tests.
20160903 05:18:20< JyrkiVesterinen> I confirmed that both in the debugger, and the logs I captured show it too.
20160903 05:18:34< JyrkiVesterinen> https://gist.github.com/jyrkive/92c08d22178d12b6df2b3f7423f52440 (end of log)
20160903 05:26:06-!- celmin [~celticmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The minstrel departs, to spread the music to the masses!]
20160903 05:29:52< irker925> wesnoth: Charles Dang wesnoth:master 55cc6886878a / src/gui/dialogs/multiplayer/mp_create_game.cpp: Mp Create: call set_current_level on game select instead of on updating details https://github.com/wesnoth/wesnoth/commit/55cc6886878a635cb2aae57e4479ee3c0a2b5745
20160903 05:31:02< celticminstrel> Wow, timing.
20160903 05:31:15< celticminstrel> That was just in time for my pull to catch it.
20160903 05:31:39< celticminstrel> Generator settings are under editor/?
20160903 05:38:14< vultraz> yes
20160903 05:38:33< celticminstrel> Why?
20160903 05:39:30< vultraz> to keep it next to the generate_map dialog
20160903 05:44:28< irker925> wesnoth: Celtic Minstrel wesnoth:master 473076104c60 / projectfiles/Xcode/Wesnoth.xcodeproj/project.pbxproj: Update XCode project https://github.com/wesnoth/wesnoth/commit/473076104c609a509aa737011f839d1c433056b1
20160903 05:48:47< irker925> wesnoth: Charles Dang wesnoth:master a232b0c2cbe5 / src/ (3 files in 2 dirs): MP Create: disable Settings button if generator doesn't allow settings https://github.com/wesnoth/wesnoth/commit/a232b0c2cbe576441522cc7bd6ce7e33d0be734c
20160903 05:49:10< vultraz> (currently only the Cave Generator seems to end up in that category)
20160903 05:49:34< celticminstrel> Lua generators allow settings?
20160903 05:50:46< vultraz> code says no
20160903 05:51:28< vultraz>  // TODO: Add support for user configurability (via defining a gui2 dialog in lua) 
20160903 05:51:29< vultraz>  // What's missing is that you need access to the 'wesnoth' object to call show dialog
20160903 05:51:31< vultraz>  // at the moment.
20160903 05:51:51< celticminstrel> Pretty sure that was resolved whenever lua_kernel_base was introduced.
20160903 05:52:07< vultraz> there is still in lua_map_generator::user_config, though..
20160903 05:52:15< vultraz> not sure, since I don't have a lua generator to test with
20160903 05:52:33< vultraz> s/still/stuff
20160903 05:53:40< vultraz> URM uses wml
20160903 05:53:42< vultraz> not lua
20160903 05:54:11< vultraz> (ultimate random maps)
20160903 05:54:15< celticminstrel> Sigurd's?
20160903 05:54:56< vultraz> yes
20160903 05:57:12< vultraz> hmm
20160903 05:57:23< vultraz> 55cc6886878a might have caused a problem..
20160903 05:57:38 * vultraz debugs
20160903 05:59:09< vultraz> oh no, my last commit
20160903 06:01:46< vultraz> hmm
20160903 06:03:23< vultraz> (and that filter bug is still there, *sigh* )
20160903 06:04:19-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.1 Aria]
20160903 06:14:20-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev
20160903 06:14:20-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection]
20160903 06:21:54-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 06:21:55< travis-ci> wesnoth/wesnoth#10689 (master - e1e2d43 : Celtic Minstrel): The build has errored.
20160903 06:21:55< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157259926
20160903 06:21:55-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 06:25:15-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity]
20160903 06:44:25-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 06:44:26< travis-ci> wesnoth/wesnoth#10690 (master - 501c50f : Charles Dang): The build has errored.
20160903 06:44:26< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157263289
20160903 06:44:26-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 06:51:34-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!]
20160903 07:00:03-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has joined #wesnoth-dev
20160903 07:01:36-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160903 07:06:03-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 265 seconds]
20160903 07:12:20-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has quit [Quit: ChipmunkV]
20160903 07:20:08-!- JyrkiVesterinen [~JyrkiVest@87-100-157-208.bb.dnainternet.fi] has quit [Quit: Going somewhere. I should be back in an hour.]
20160903 07:24:12-!- mjs-de [~mjs-de@x4db5144b.dyn.telefonica.de] has joined #wesnoth-dev
20160903 07:33:03-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 07:33:05< travis-ci> wesnoth/wesnoth#10691 (debug-multiplayer-test-failures - 1d973b7 : Jyrki Vesterinen): The build has errored.
20160903 07:33:05< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157263627
20160903 07:33:05-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 07:40:01-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 07:44:42-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 276 seconds]
20160903 07:44:42-!- wedge010 is now known as wedge009
20160903 07:47:18-!- mjs-de [~mjs-de@x4db5144b.dyn.telefonica.de] has quit [Ping timeout: 276 seconds]
20160903 07:47:45-!- mjs-de [~mjs-de@x4db5144b.dyn.telefonica.de] has joined #wesnoth-dev
20160903 07:49:48-!- Duthlet [~Duthlet@dslb-188-104-247-024.188.104.pools.vodafone-ip.de] has joined #wesnoth-dev
20160903 07:51:33-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has joined #wesnoth-dev
20160903 08:08:11-!- JyrkiVesterinen [~JyrkiVest@87-100-157-208.bb.dnainternet.fi] has joined #wesnoth-dev
20160903 08:09:07-!- horrowind [~Icedove@2a02:810a:8380:10a8:21b:fcff:fee3:c3ff] has joined #wesnoth-dev
20160903 08:10:52-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 08:10:53< travis-ci> wesnoth/wesnoth#10692 (master - 55cc688 : Charles Dang): The build has errored.
20160903 08:10:53< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157264798
20160903 08:10:53-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 08:15:32< JyrkiVesterinen> The host gets much further now.
20160903 08:15:49< JyrkiVesterinen> Yet, the new "no plugin context" message doesn't appear anywhere in the log. Strange.
20160903 08:18:35< JyrkiVesterinen> The host gets to hosting the game, but not starting it.
20160903 08:24:33< JyrkiVesterinen> Okay, the problem is that the guest apparently doesn't get the leader select dialog.
20160903 08:27:02-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev
20160903 08:35:36< zookeeper> mattsc, ohh, wonderful. so the high xp attacking works by default everywhere now?
20160903 08:35:58< zookeeper> (well, after that PR i mean)
20160903 08:39:16< JyrkiVesterinen> (continuing my ramblings abobe) From what I can tell, the guest gets stuck attempting to join the test game.
20160903 08:39:33< JyrkiVesterinen> I suspect that it can't find the game, or it can't join the game for whatever reason.
20160903 08:43:14< JyrkiVesterinen> Ahhh. The log shown in the website was just truncated. The raw log shows that the test got much further than I thought. :)
20160903 08:44:09< JyrkiVesterinen> (The log is monstrous, hundreds of thousands of lines. I need that level of accuracy for the brief moment between the scenarios.)
20160903 08:46:45 * JyrkiVesterinen notices that both the host and the guest win the scenarios in the test - not quite a normal MP game result 
20160903 08:48:46< JyrkiVesterinen> ...and here we are. These two lines.
20160903 08:48:47< JyrkiVesterinen> 20160903 06:52:21 debug mp/main: running lobby loop
20160903 08:48:55< JyrkiVesterinen> 20160903 06:52:21 debug mp/main: leaving lobby loop, result 6
20160903 08:49:23< JyrkiVesterinen> The loop doesn't run even a single iteration. It doesn't play a plugin slice.
20160903 08:49:45-!- irker925 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160903 08:49:59-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160903 08:50:13< JyrkiVesterinen> This is why the join script stays in the upper loop in Travis.
20160903 08:51:18< JyrkiVesterinen> There are hundreds of thousands of "lobby loop: playing plugins slice" lines in the log, but none between the two lines above.
20160903 08:53:00-!- Guest94942 is now known as matthiaskrgr
20160903 08:53:06-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 08:53:07< travis-ci> wesnoth/wesnoth#10693 (master - 4730761 : Celtic Minstrel): The build has errored.
20160903 08:53:08< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157265586
20160903 08:53:08-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 08:53:12< JyrkiVesterinen> That means, in turn, that the wait dialog returns a result other than mp::ui::CONTINUE in the first check.
20160903 08:53:59< JyrkiVesterinen> Result 6 is mp::ui::PLAY.
20160903 08:54:08-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 250 seconds]
20160903 08:58:22< JyrkiVesterinen> One possible place where the wait dialog result may be set before the plugin gets a chance to run: https://github.com/wesnoth/wesnoth/blob/1d973b702026d2add78ec78d3585abf8501172fe/src/game_initialization/multiplayer_wait.cpp#L108-L111
20160903 08:59:54-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160903 09:01:02-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Client Quit]
20160903 09:11:33< JyrkiVesterinen> Looking at the log, I see that the server prints the second "starting game" message before the guest starts downloading level data.
20160903 09:12:07< JyrkiVesterinen> This suggests that the second scenario has indeed already started when the guest joins it.
20160903 09:12:29< JyrkiVesterinen> Why doesn't that happen on my computer? Could it be a timing related issue?
20160903 09:16:50-!- Kwandulin [~Miranda@p200300760F4241433074E3EB233DE29F.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160903 09:20:26< JyrkiVesterinen> My current mood: https://www.youtube.com/watch?v=jPJE4l6j-hg :D
20160903 09:26:07-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160903 09:30:22-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 252 seconds]
20160903 09:30:39< JyrkiVesterinen> I was right. I reproduced the issue locally by adding a delay to the source code. AWWW YEAH. :D
20160903 09:31:14< JyrkiVesterinen> Now to think how to fix it. Our MP tests shouldn't depend on timing.
20160903 09:36:50-!- exciton [chuck-the-@89.208.170.132] has joined #wesnoth-dev
20160903 09:43:34-!- exciton [chuck-the-@89.208.170.132] has quit [Ping timeout: 252 seconds]
20160903 09:46:55-!- exciton [~exciton@89.208.170.132] has joined #wesnoth-dev
20160903 10:10:18-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160903 10:10:24-!- Duthlet [~Duthlet@dslb-188-104-247-024.188.104.pools.vodafone-ip.de] has quit [Ping timeout: 250 seconds]
20160903 10:21:23-!- Duthlet [~Duthlet@dslb-188-104-247-024.188.104.pools.vodafone-ip.de] has joined #wesnoth-dev
20160903 10:25:52-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160903 10:27:20-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Client Quit]
20160903 10:33:28-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has quit [Ping timeout: 255 seconds]
20160903 10:37:59-!- vultraz [~chatzilla@wesnoth/developer/vultraz] has joined #wesnoth-dev
20160903 10:41:47< JyrkiVesterinen> I'm working on a fix, namely explicitly waiting for the last scenario in the join script instead of assuming that the wait dialog either will or won't be shown between the scenarios.
20160903 10:42:07< JyrkiVesterinen> I haven't yet gotten it working without crashing.
20160903 10:42:21< JyrkiVesterinen> I need to go somewhere else for an hour or two, I'll resume after that.
20160903 10:42:23-!- JyrkiVesterinen [~JyrkiVest@87-100-157-208.bb.dnainternet.fi] has quit [Quit: .]
20160903 10:47:38-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160903 10:51:50-!- Kwandulin [~Miranda@p200300760F424143B173B851507635F3.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160903 11:08:29-!- mjs-de [~mjs-de@x4db5144b.dyn.telefonica.de] has quit [Remote host closed the connection]
20160903 11:39:00-!- JyrkiVesterinen [~JyrkiVest@87-100-170-47.bb.dnainternet.fi] has joined #wesnoth-dev
20160903 11:40:03-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.2 Aria]
20160903 11:51:36-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160903 12:08:47-!- hk238 [~kvirc@t224.ip7.netikka.fi] has joined #wesnoth-dev
20160903 12:08:55-!- hk238 [~kvirc@t224.ip7.netikka.fi] has quit [Changing host]
20160903 12:08:55-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160903 12:09:36-!- Greg-Boggs [~greg_bogg@173.240.241.83] has joined #wesnoth-dev
20160903 12:13:49-!- Greg-Boggs [~greg_bogg@173.240.241.83] has quit [Ping timeout: 255 seconds]
20160903 12:21:09-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.1 Aria]
20160903 12:31:40-!- horrowind [~Icedove@2a02:810a:8380:10a8:21b:fcff:fee3:c3ff] has quit [Quit: horrowind]
20160903 12:36:55-!- irker298 [~irker@uruz.ai0867.net] has joined #wesnoth-dev
20160903 12:36:55< irker298> wesnoth: Charles Dang wesnoth:master 74f1fe1e15fb / src/gui/dialogs/multiplayer/mp_create_game.cpp: MP Create: fixed games list not function correctly if switching type with filter https://github.com/wesnoth/wesnoth/commit/74f1fe1e15fbac56dcef1363d735284c927d7fe5
20160903 12:36:58< vultraz> finally
20160903 12:37:07< vultraz> need to ask celticminstrel something, though
20160903 12:40:18-!- gfgtdf [~chatzilla@x4e368e49.dyn.telefonica.de] has joined #wesnoth-dev
20160903 12:44:58< vultraz> gfgtdf: I think Create is fully functional now, except for campaign entry points.
20160903 12:45:08< gfgtdf> :D
20160903 12:45:46< vultraz> I also made a small mods-only configure screen for SP, dunno if you saw.
20160903 12:46:10< gfgtdf> vultraz: i saw the commit on github but didnt test it yet.
20160903 12:46:57< gfgtdf> vultraz: you intend to make a gui2 mp connect/wait next ?
20160903 12:50:13< vultraz> yeah that needs to be done
20160903 12:50:19< vultraz> it's gonna be a ton of work though
20160903 12:50:28< vultraz> would appreciate help
20160903 12:51:58< vultraz> for one
20160903 12:52:06< vultraz> the chat area from the lobby needs to be widget-ized
20160903 12:52:52< vultraz> we're also missing the drag-and-drop behavior gui1 had
20160903 12:52:58< vultraz> dunno if we need that, though
20160903 12:53:57< gfgtdf> vultraz: hmm i persoalyl never knew it had the drag and drop beaviour until i read the code.
20160903 12:59:51< vultraz> yeah
20160903 12:59:57< vultraz> there's really no indication of it
20160903 13:00:04< vultraz> plus one can reassign players from the dropdown
20160903 13:00:47< vultraz> so that's not really a necessary feature
20160903 13:00:50< vultraz> the chatbox is though
20160903 13:02:04< vultraz> and I need to decide on a layout
20160903 13:02:29< gfgtdf> vultraz: what woudl you change ?
20160903 13:02:47< vultraz> well we want to merge connect and wait right?
20160903 13:04:43< vultraz> im thinking something similar to  the lobby
20160903 13:05:26< vultraz> layout-wis
20160903 13:05:27< vultraz> e
20160903 13:05:59< vultraz> only the host would be able to change player order
20160903 13:06:14< vultraz> but i think anyone should be able to select their faction/color/leader
20160903 13:07:21< vultraz> perhaps a button to bring up the flg dialog
20160903 13:17:39< mattsc> zookeeper: hi — yes, it’s part of the default AI, so it should be working everywhere
20160903 13:27:43< vultraz> gfgtdf: btw, there are only 2 "small" gui1 dialogs left (I think): Attack Prediction and Statistics. Any ideas how we could convert the former? (maybe adding a version of timage that takes a surface and constructing the surface in preshow? that still leaves the labels in the damage table, though..)
20160903 13:28:39< vultraz> (Statistics should be relatively simple to convert, just rather tedious)
20160903 13:30:37-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 13:30:38< travis-ci> wesnoth/wesnoth#10695 (master - 74f1fe1 : Charles Dang): The build has errored.
20160903 13:30:38< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157299812
20160903 13:30:38-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 13:33:23< gfgtdf> vultraz: well you coiudl alwayys make teh labels part of that 'surface' but i dont rally liek that solution
20160903 13:33:52< vultraz> yeah me neither
20160903 13:34:08< vultraz> it's possible but not optimal 
20160903 13:34:12< gfgtdf> vultraz: i think we shodul try to make some bar chart or generall chart widget
20160903 13:34:26< gfgtdf> vultraz: wasnt there soemone on he forum wo is working ona charrt widget?
20160903 13:34:35< vultraz> there was?
20160903 13:34:42< vultraz> i don't recall such a thing
20160903 13:35:24< vultraz> oh
20160903 13:35:26< vultraz> hmm
20160903 13:35:27< vultraz> https://forums.wesnoth.org/viewtopic.php?f=10&t=44404&p=600231&hilit=widget+chart#p600231
20160903 13:35:33< gfgtdf> vultraz: i menar this https://forums.wesnoth.org/viewtopic.php?f=10&p=600231 () but appeartnyl is adifferent type of chart?
20160903 13:35:48< vultraz> yeah that's what I found
20160903 13:36:01< vultraz> that's a line graph.. we need a bar graph.
20160903 13:37:55< vultraz> it doesn't look half bad, actually
20160903 13:38:26< gfgtdf> when i log in in the forum i auotmatically get redirected to the strting page, (instad of to the the tread im current viewwing). This alsone woudlnt be that anyoing but somehow i cannot even go back to my last visited thread by using the browsers 'back'. Do you alos have this problem?
20160903 13:39:01< vultraz> i dunno... i never log out of the forums
20160903 13:39:12< vultraz> or more accurately, i log in automatically
20160903 13:39:49< zookeeper> mattsc, super. good job :J
20160903 13:41:09< mattsc> zookeeper: thanks :)
20160903 13:41:50< vultraz> gfgtdf: come to think of it, this line graph might actually work for Attack Prediction... x could be percentage change and y could be hp
20160903 13:41:56< vultraz> chance
20160903 13:41:56< mattsc> zookeeper: did you try out the test scenario?  Should I wait for other comments?
20160903 13:42:12< zookeeper> i haven't tried it myself, no
20160903 13:42:15< vultraz> it just needs a few tweaks
20160903 13:42:18< vultraz> to the layout
20160903 13:42:42< vultraz> (y labels should be on the other side of the line, for one)
20160903 13:43:11< vultraz> and the lines should optimally be a little thicker and anti-aliased 
20160903 13:43:54< vultraz> gfgtdf: thoughts?
20160903 13:44:10< mattsc> zookeeper: I’d be interested to hear what you have to say about how it chooses to attack (or not); all you have to do is start Wesnoth with the -thigh_xp_attack CL option and there’s almost a dozen test cases set up right there
20160903 13:44:22-!- horrowind [~Icedove@2a02:810a:8380:10a8:21b:fcff:fee3:c3ff] has joined #wesnoth-dev
20160903 13:45:56< gfgtdf> vultraz: hmm that woudl be posible, but i think i like that current bar chrts more, but that coudl also just be becasue im just used to them.
20160903 13:47:14< vultraz> a line graph might actually be a better representation of the data
20160903 13:47:18< vultraz> im still pondering
20160903 13:47:51< irker298> wesnoth: Jyrki Vesterinen wesnoth:debug-multiplayer-test-failures 0d4bd206d96d / utils/travis/mp_test_executor.sh: Turn off excessive logging of MP tests https://github.com/wesnoth/wesnoth/commit/0d4bd206d96df1150fbcbe82315fc43f4a068055
20160903 13:47:53< irker298> wesnoth: Jyrki Vesterinen wesnoth:debug-multiplayer-test-failures 24df62476087 / join.lua src/play_controller.cpp: Fix multiplayer tests failing in Travis CI https://github.com/wesnoth/wesnoth/commit/24df62476087e30a85ab163f79711f26fe9d86f8
20160903 13:48:09< vultraz> let's see what celmin thinks
20160903 13:48:50< vultraz> JyrkiVesterinen: ah, you fixed it? :D
20160903 13:49:00< JyrkiVesterinen> Hopefully yes.
20160903 13:49:17< JyrkiVesterinen> I pushed the fix only to the debug-multiplayer-test-failures branch for now.
20160903 13:49:20< zookeeper> mattsc, well it's a bit awkward for me to try to test PR's locally...
20160903 13:49:32< JyrkiVesterinen> If the test passes there, I'll cherry-pick the commit to master.
20160903 13:50:23< vultraz> sweet :D
20160903 13:50:41< mattsc> zookeeper: okay, no worries; I’ll wait a little longer then to see if anybody else comments, and you can test it after it’s been merged.  We can always mogify it afterwards.
20160903 13:51:18< zookeeper> mattsc, yeah and i doubt there's any massive holes in it anyway
20160903 13:52:11< mattsc> I hope not; but I’ve several times encountered cases with other AIs where some situation happened that I just didn’t think of.
20160903 13:52:36< mattsc> We’ll mostly only find those once lots of people use it with lots of campaigns though.
20160903 13:59:28-!- edgrey [~edgrey@178.204.108.159] has joined #wesnoth-dev
20160903 14:05:31-!- gfgtdf [~chatzilla@x4e368e49.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.92 [Firefox 48.0.2/20160823121617]]
20160903 14:06:36-!- Rhonda [~rhonda@anguilla.debian.or.at] has quit [Remote host closed the connection]
20160903 14:06:52< irker298> wesnoth: mattsc wesnoth:ai_high_xp_attack 390d1fdc10cd / data/ai/lua/ca_high_xp_attack.lua: High XP CA: simplify attack conditions with 1.13 syntax https://github.com/wesnoth/wesnoth/commit/390d1fdc10cd9fb878186fb1c123cd1e343bdb6d
20160903 14:06:54< irker298> wesnoth: mattsc wesnoth:ai_high_xp_attack af0f49ae9a8c / data/core/macros/ai_candidate_actions.cfg: High XP CA: add missing braces to CA definition https://github.com/wesnoth/wesnoth/commit/af0f49ae9a8c907ff420c9e851fd44188c1ebb5d
20160903 14:10:18-!- Rhonda [~rhonda@anguilla.debian.or.at] has joined #wesnoth-dev
20160903 14:25:16-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160903 14:25:54-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 14:25:55< travis-ci> wesnoth/wesnoth#10696 (debug-multiplayer-test-failures - 24df624 : Jyrki Vesterinen): The build passed.
20160903 14:25:55< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157307601
20160903 14:25:55-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 14:26:19< JyrkiVesterinen> THE BUILD PASSED OMG MY FIX WORKS
20160903 14:26:22< JyrkiVesterinen> :D
20160903 14:28:36< mattsc> yay
20160903 14:31:33< irker298> wesnoth: Jyrki Vesterinen wesnoth:master 2c4d72f544b4 / join.lua src/play_controller.cpp: Fix multiplayer tests failing in Travis CI https://github.com/wesnoth/wesnoth/commit/2c4d72f544b447d5a225f349b2078923e2f767cf
20160903 14:31:55< celticminstrel> \o/
20160903 14:46:49-!- Bonobo [~Bonobo@2001:44b8:254:3200:4485:c528:47c6:870a] has quit [Ping timeout: 255 seconds]
20160903 14:47:08-!- Bonobo [~Bonobo@2001:44b8:254:3200:5508:246e:1386:1752] has joined #wesnoth-dev
20160903 14:54:16< mattsc> celticminstrel: FYI, max_score is set to 10^7 when not given
20160903 14:55:12< celticminstrel> max_score only affects order of evaluation, right?
20160903 14:55:20< mattsc> yes
20160903 14:59:00-!- exciton [~exciton@89.208.170.132] has quit [Read error: Connection reset by peer]
20160903 14:59:30-!- Bonobo [~Bonobo@2001:44b8:254:3200:5508:246e:1386:1752] has quit [Quit: Leaving]
20160903 15:04:16-!- exciton [chuck-the-@89.208.170.132] has joined #wesnoth-dev
20160903 15:08:17< irker298> wesnoth: Jyrki Vesterinen wesnoth:master 357cbfe35aea / projectfiles/VC12/ (wesnoth.vcxproj wesnoth.vcxproj.filters): Update Visual Studio project https://github.com/wesnoth/wesnoth/commit/357cbfe35aea49b9010e6d5d0a848f48cf7a6880
20160903 15:14:33-!- Kwandulin [~Miranda@p200300760F424143B173B851507635F3.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160903 15:15:56-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 15:15:57< travis-ci> wesnoth/wesnoth#10697 (ai_high_xp_attack - af0f49a : mattsc): The build has errored.
20160903 15:15:57< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157310596
20160903 15:15:57-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 15:17:20< celticminstrel> So I decided against converting addon progress to loading screen, because the loading screen currently can't be cancelled, and addon {down,up}load should be cancellable.
20160903 15:17:49< celticminstrel> shadowm, loonycyborg: Are there any reasons not to merge the asio campaignd?
20160903 15:30:38-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160903 15:30:58< tad_> vultraz: you up?
20160903 15:31:33< celticminstrel> Hi
20160903 15:31:42< tad_> Hey.
20160903 15:32:04< tad_> OK. I am closing in on the culprit for the corrupted games
20160903 15:32:18< tad_> I am expanding the description, you note.
20160903 15:33:16< tad_> The question is how to fix it. The feature is badly done and screams to be rewritten from the ground up. But that's not gonna fly.
20160903 15:33:54< celticminstrel> Which feature? Saved games?
20160903 15:34:09< tad_> Something called "Multiplayer"
20160903 15:34:15< celticminstrel> Ah.
20160903 15:34:15 * tad_ grinZ
20160903 15:35:04< tad_> Actually, it looks like the culprit is the side-shuffle which appears to be a MP feature and is being done for SP, too.
20160903 15:36:48< tad_> To do it's job it is loading a second copy of the entire campaign, doing the side-shuffle for the scenario, then casting away const-ness and overwriting the perfectly good, but not shuffled, scenario with it's buggy, badly done with errors all over the place, scenario.
20160903 15:37:03< celticminstrel> Hmm, I wonder if the ellipsis should be included in the translatable strings rather than being appended...
20160903 15:37:19< tad_> I'm thinking this is going to also turn out to be the memory leak I'm seeing.  This sort of crap is very memory leak-ish.
20160903 15:37:35< celticminstrel> Ooh, const_cast. 
20160903 15:37:49< celticminstrel> I removed a lot of uses, but I guess that was one of the ones I couldn't decipher or something.
20160903 15:38:07< tad_> Not sure how it's doing its job but it's a const object being whacked.
20160903 15:38:54< celticminstrel> Rewriting MP may not be an option, but rewriting side shuffle might me... >_>
20160903 15:39:02< tad_> ellipsis is a filename, right? Do we really need localized filenames?
20160903 15:39:21< celticminstrel> Huh? No, this is in the loading screen.
20160903 15:39:50< celticminstrel> (We do have some localized images, though I'm not sure if those are passed through gettext. Probably not.)
20160903 15:39:51-!- horrowind [~Icedove@2a02:810a:8380:10a8:21b:fcff:fee3:c3ff] has quit [Remote host closed the connection]
20160903 15:40:10< tad_> I have a red/blue error logger on either side of where the error **appears** and it's the copy function for config-to-state
20160903 15:40:17< celticminstrel> (Speaking of localized images, do we have all the logo text images yet?)
20160903 15:40:22< celticminstrel> vultraz: ^
20160903 15:40:33< celticminstrel> Red/blue?
20160903 15:40:41< tad_> Pass/fail
20160903 15:41:17< tad_> 'tis good, now 'tis bad .. I added a divide-by-zero to the bad so I'd not miss it and could check the backtrace.
20160903 15:41:28-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has joined #wesnoth-dev
20160903 15:42:16-!- horrowind [~Icedove@2a02:810a:8380:10a8:21b:fcff:fee3:c3ff] has joined #wesnoth-dev
20160903 15:43:13< tad_> So .. I've found where the bad data is copied live. This morning I've going to try to find where the good data becomes bad. It _may_ be fixable there. But the code is a mess and modifying a const object is never a good sign.
20160903 15:43:40< celticminstrel> Yeah.
20160903 15:44:26-!- Kwandulin [~Miranda@p200300760F424143C8F64111927C52C4.dip0.t-ipconnect.de] has joined #wesnoth-dev
20160903 15:44:37< celticminstrel> I hope that const object is at least a copy of the original scenario.
20160903 15:45:02< tad_> This is where the user_team_name goes bad. It's also where the [ai] gets lost, the translation scope changes .. probably more I've not noticed
20160903 15:46:13< tad_> Well, it looks like it's a full reload-from-disk but I may be wrong. If it's not, it's somehow recovering the translation scope .. #wesnoth is changing to #wesnoth-tsg at this point.
20160903 15:47:07< celticminstrel> The textdomain is supposed to be wesnoth-tsg, I assume?
20160903 15:47:09< tad_> But, hey, at least it's finally setting the player (unix account, for me) name instead of leaving it blank.
20160903 15:48:09-!- horrowind [~Icedove@2a02:810a:8380:10a8:21b:fcff:fee3:c3ff] has quit [Quit: horrowind]
20160903 15:48:12< tad_> Yeah, I'm testing using TSG because it's the smallest I have which I know has an error I can check because I just added it.
20160903 15:49:32< tad_> My forensics makes me think that all was well but someone wanted to shuffle sides and did it poorly then someone wanted to localize the team names and missed the fact you have to duplicate the work because of the way shuffle works.
20160903 15:50:22< celticminstrel> I feel like changing the way shuffle works would be better than duplicating the work...
20160903 15:51:24< tad_> We'll see. I have back-traced the live data back to where it's changed. Now I have to up-trace the bad data back to where it's created. Hopefully it's not far away.
20160903 15:54:15< tad_> My guess is that the easiest course of action will be to fix the error for user_team_name but the wisest course will be a rewrite of the shuffle. But first I need to find the source of the data. Hopefully, it's a copy and not a re-load and re-parse from disk. But, from the names, I fear it's a re-load and re-parse.
20160903 15:55:00< celticminstrel> Wait, what?
20160903 15:56:29< tad_> Nearby function names are pro-active "load_scenario" or some such. Dunno what they do, or if they're involved but as I was headed to bed the impression I hd last night was I was about to discover a complete reload of the campaign so they could shuffle the sides of a scenario ...
20160903 15:57:25< tad_> But, its possible what I'm seeing just started out as a copy-n-paste and the names were NOT changed to protect the innocent.
20160903 15:58:20< celticminstrel> It's also possible that "load_scenario" really means "load the scenario data from the in-memory game config".
20160903 15:59:24< tad_> I'm so deed into "this called that, extracting this from it, as a copy, and passing both as data to something else" it's hard to keep track. So I'm following just the side number, leader id and user_team_name
20160903 16:00:45-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 16:00:46< travis-ci> wesnoth/wesnoth#10699 (master - 2c4d72f : Jyrki Vesterinen): The build passed.
20160903 16:00:46< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157313368
20160903 16:00:46-!- travis-ci [~travis-ci@ec2-54-145-84-169.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 16:01:14< tad_> const&value=nonconst&state --> pass(value,state) --> extract value.key --> pass(key,value,state) --> extract value2 --> pass(value2,state) .. yup, no idea where key and value went .. ah well .. keep on keeping on .. 
20160903 16:01:21< tad_> I am going cross-eyed.
20160903 16:01:48< celticminstrel> \o/ Finally passing build!
20160903 16:01:58< JyrkiVesterinen> :)
20160903 16:01:58< tad_> WOW
20160903 16:03:10< tad_> So, anyway, I'm up, I'm properly cafinated, I've had my bitch about the code .. I s'pose it's time I got back to it.
20160903 16:05:23-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160903 16:05:59< loonycyborg> JyrkiVesterinen: what exactly was wrong with that test harness?
20160903 16:06:18< loonycyborg> it was sometimes working, sometimes not
20160903 16:06:37< loonycyborg> one time my compile warning fix made it work
20160903 16:06:49< JyrkiVesterinen> The commit message explains it in full: https://github.com/wesnoth/wesnoth/commit/2c4d72f544b447d5a225f349b2078923e2f767cf
20160903 16:07:11< JyrkiVesterinen> In short, the test depended on timing.
20160903 16:08:06< loonycyborg> nice ty
20160903 16:09:09< celticminstrel> "info.scenario_name().scenario_name" :|
20160903 16:09:37< JyrkiVesterinen> That's how plugin accessors work, unfortunately.
20160903 16:09:58< JyrkiVesterinen> They can only return configs, which are automatically converted to tables.
20160903 16:10:12< celticminstrel> I had assumed that set_accessor_xyz would make them actually return an xyz. :/
20160903 16:10:45< JyrkiVesterinen> Me too, but I quickly noticed that they return tables instead. :S
20160903 16:27:43-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has joined #wesnoth-dev
20160903 16:27:44< travis-ci> wesnoth/wesnoth#10700 (master - 357cbfe : Jyrki Vesterinen): The build passed.
20160903 16:27:44< travis-ci> Build details : https://travis-ci.org/wesnoth/wesnoth/builds/157317681
20160903 16:27:44-!- travis-ci [~travis-ci@ec2-54-159-91-81.compute-1.amazonaws.com] has left #wesnoth-dev []
20160903 16:29:02< JyrkiVesterinen> Why did Travis tell us about that build? IIRC, it should be configured to only inform us when build is broken or when someone fixes it.
20160903 16:29:14< JyrkiVesterinen> In other words, it should omit consecutive passing builds.
20160903 16:30:00< celticminstrel> I dunno.
20160903 16:30:15< celticminstrel> It's been like that since before I showed up.
20160903 16:30:58< JyrkiVesterinen> I don't recall it working that way.
20160903 16:40:57-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Quit: wedge009]
20160903 16:43:43-!- mjs-de [~mjs-de@x4e307068.dyn.telefonica.de] has joined #wesnoth-dev
20160903 16:45:18-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 16:46:18-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Client Quit]
20160903 16:48:25-!- DeFender1031 [~DeFender1@46-116-114-128.bb.netvision.net.il] has joined #wesnoth-dev
20160903 16:51:28-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 16:51:55-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160903 16:52:06-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Client Quit]
20160903 16:56:35-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 16:57:52-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Client Quit]
20160903 16:59:36-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Quit: http://www.kvirc.net/ 4.9.2 Aria]
20160903 17:01:38-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 17:05:46-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 250 seconds]
20160903 17:10:23-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 17:10:24< celticminstrel> Got this working, just need to verify once more and then I can push.
20160903 17:17:56-!- hk238 [~kvirc@unaffiliated/hk238] has joined #wesnoth-dev
20160903 17:34:10< celticminstrel> Argh, crash.
20160903 17:37:01-!- exciton [chuck-the-@89.208.170.132] has quit [Read error: Connection reset by peer]
20160903 17:42:15-!- exciton [chuck-the-@89.208.170.132] has joined #wesnoth-dev
20160903 17:51:26-!- hk238|2 [~kvirc@t224.ip7.netikka.fi] has joined #wesnoth-dev
20160903 17:53:06-!- hk238 [~kvirc@unaffiliated/hk238] has quit [Ping timeout: 276 seconds]
20160903 17:56:49-!- hk238|2 [~kvirc@t224.ip7.netikka.fi] has quit [Quit: http://www.kvirc.net/ 4.9.1 Aria]
20160903 18:08:40-!- irker298 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout]
20160903 18:24:33< shadowm> celticminstrel: Haven't properly verified that all the regressions were fixed (and that there aren't more) and people don't seem to care enough to remind me that this is a thing I was supposed to do several months ago.
20160903 18:25:15< celticminstrel> I cared enough to remind you just then.3
20160903 18:25:21< shadowm> I'd need a daily reminder insted.
20160903 18:25:21< shadowm> *ad
20160903 18:25:25< celticminstrel> Okay.
20160903 18:25:33< celticminstrel> I'll remind you again tomorrow then.
20160903 18:26:10< celticminstrel> Just didn't want to annoy you with unsolicited regular reminders.
20160903 18:39:22-!- Kwandulin [~Miranda@p200300760F424143C8F64111927C52C4.dip0.t-ipconnect.de] has quit [Read error: Connection reset by peer]
20160903 18:49:59-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev
20160903 19:22:57-!- Sirp_ [~Sirp@u17402953.onlinehome-server.com] has joined #wesnoth-dev
20160903 19:23:49-!- Sirp [~Sirp@u17402953.onlinehome-server.com] has quit [Ping timeout: 240 seconds]
20160903 19:54:28-!- JyrkiVesterinen [~JyrkiVest@87-100-170-47.bb.dnainternet.fi] has quit [Quit: .]
20160903 20:03:30< shadowm> src/tstring.cpp:297:39: error: dereferencing type-punned pointer will break strict-aliasing rules [-Werror=strict-aliasing]
20160903 20:03:33< shadowm> celticminstrel: ^
20160903 20:03:43< celticminstrel> Right.
20160903 20:05:24-!- trewe [~trewe@2001:8a0:d12f:d901:22a:827a:b1e8:18d0] has joined #wesnoth-dev
20160903 20:07:15< celticminstrel> Fixing now.
20160903 20:08:04-!- mjs-de [~mjs-de@x4e307068.dyn.telefonica.de] has quit [Ping timeout: 260 seconds]
20160903 20:41:29< vultraz> [02:40:16]	celticminstrel	(Speaking of localized images, do we have all the logo text images yet?)
20160903 20:41:30< vultraz> uhhh
20160903 20:41:39< vultraz> we have a good number
20160903 20:41:48< vultraz> but i dont think all
20160903 20:43:33< vultraz> celticminstrel: what are your thoughts on this line graph for Attack Prediction? https://forums.wesnoth.org/viewtopic.php?f=10&p=600231
20160903 20:44:32< vultraz> celticminstrel: also, I was able to fix the filter issues in mp create, except now if no games show and you switch to Campaigns, generator.cpp:382 asserts. it displays fine if yoi comment that out. Can we remove that assert>
20160903 20:58:44< celticminstrel> So what that means is that, for some reason, it wasn't able to fill the space alotted to it.
20160903 20:59:35< celticminstrel> If you want to remove the assert, I'd suggest replacing it with a log message and applying the same change in thorizontal_list::place.
20160903 21:01:49< celticminstrel> As for the graph, I'll check that... probably after I push.
20160903 21:19:46< vultraz> It's possible the graph is a better representation of the data
20160903 21:20:01< vultraz> hp/percentage change as the two axises 
20160903 21:20:52< vultraz> (line graph, that is)
20160903 21:21:09< vultraz> (right now we have a bar graph)
20160903 21:22:19< celticminstrel> A bar graph seems more logical to me, since it's a discrete data set.
20160903 21:22:32< celticminstrel> You won't ever end up with 3,2145 HP.
20160903 21:22:40< celticminstrel> Dot not comma
20160903 21:23:34< celticminstrel> I'm getting a really annoying crash in the loading screen while connecting to the server.
20160903 21:23:48< celticminstrel> It's annoying because the stack trace appears to be an outright lie.
20160903 21:32:25-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Quit: wedge009]
20160903 21:32:41-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 21:33:31< celticminstrel> Does that suggest the stack is corrupted?
20160903 21:34:12-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 21:34:22-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Client Quit]
20160903 21:37:02-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 250 seconds]
20160903 21:39:44-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 21:40:17-!- celmin [~celticmin@unaffiliated/celticminstrel] has joined #wesnoth-dev
20160903 21:40:30-!- fabi [~fabi@176.0.22.31] has quit [Ping timeout: 244 seconds]
20160903 21:41:21-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 21:43:58-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 255 seconds]
20160903 21:44:25-!- Appleman1234 [~Appleman1@KD119104108200.au-net.ne.jp] has quit [Ping timeout: 255 seconds]
20160903 21:44:52< celmin> vultraz: I posted in that thread.
20160903 21:46:08-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 21:48:16-!- wedge011 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 21:48:32< vultraz> Invasion of the Wedge
20160903 21:48:37-!- wedge010 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 265 seconds]
20160903 21:49:47< celmin> Heh...
20160903 21:50:23-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 258 seconds]
20160903 21:51:05-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has joined #wesnoth-dev
20160903 21:52:57-!- wedge011 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has quit [Ping timeout: 276 seconds]
20160903 21:57:25-!- wedge009 [~Thunderbi@60-241-236-92.static.tpgi.com.au] has joined #wesnoth-dev
20160903 21:58:38< tad_> celmin: PR 766 is my least-change fix for user_team_name .. "works for me" .. I'm betting it breaks something for MP games, most likely when randomizing sides, and then, only if sides have differing user_team_names.
20160903 21:58:55< celmin> We should get gfgtdf to review it then, probably.
20160903 22:01:19< tad_> Well, sure, someone who knows what's actually expected on the MP side would be good. It was a mess finding this .. track the data down the call chain, find the bad ref, track back up, find where it changed, track back down to the change site.  sheesh.
20160903 22:01:40< vultraz> Welcome to Wesnoth development :P
20160903 22:02:07< vultraz> Ya know, I did want to change the keys from team_name and user_team_name to team_id and team_name at one point
20160903 22:02:11< vultraz> but that had to be reverted
20160903 22:02:24< tad_> All this because it's old-style C code and all this C++ we see is a bad attempt to hide the fact it's ALL one HUGE shared/mod object.
20160903 22:02:51< vultraz> did you say... C code? D:
20160903 22:03:15< celmin> So my guess that it holds a reference to the scenario definition on the game config was correct?
20160903 22:03:39< tad_> Sorta explains why every .hpp is needed almost everywhere.  IF you need "game state" you need almost every .hpp because the 'game state object' has all of it 
20160903 22:05:15< tad_> Nah. It holds the actual data and all the refs and copies and clones eventually reach the same point .. two copies if you randomize .. and the names imply "read" as from file but they mean "read" from original into copy .. *I*THINK*  .. it's a call-me-call-you-and-isn't-bob-your-uncle? code.
20160903 22:05:53< celmin> Heh...
20160903 22:09:13< celmin> The fix is surprisingly small.
20160903 22:09:49< tad_> Anyway. PR 766 fixes it for me. If gfgtdf doesn't approve of it I'm going to push for a rewrite because SP campaigns are gonna want it fixed by 1.14 ...
20160903 22:10:02< tad_> Like my comment on the one-line fix?
20160903 22:10:09< celmin> Yeah. :P
20160903 22:10:22< celmin> It's unsigned, but I guess the commit author gives that.
20160903 22:10:41< celmin> Which would require git blame, but whatever.
20160903 22:10:50< tad_> Wife is glad I'm done with it.  Pens flying. Shoes thunking. Random grumbles and grunts.
20160903 22:10:59< celmin> What. XD
20160903 22:12:26< tad_> My behavior every time I found another indirection and had to track back up and back down .. "Ha! Found it in this function!" Nope. Function takes data from global and mods other data from global.  YUCK! Throw pens and shoes and start in again ...
20160903 22:12:45< celmin> XD
20160903 22:17:09< tad_> vultraz: changing team_name to team_id was probably wise.  but then changing user_team_name to team_name was a mistake. To much chance of error since you changed "team_name" from an id to a translatable.
20160903 22:17:43< celmin> I think wmllint handling was a major reason why it was reverted.
20160903 22:17:47< vultraz> tad_: I couldn't think of a suitable name otherwise 
20160903 22:17:58< vultraz> tad_: and wmllint had issues, so we reverted
20160903 22:18:03< celmin> I think his point was that you didn't need to change bouth.
20160903 22:18:05< celmin> ^both
20160903 22:18:18< tad_> team_id and user_team_name work for me and wmllint should have been fixable
20160903 22:18:24< celmin> The crash seems to be in different places in std::string each time...
20160903 22:19:22< vultraz> that's reasonable, I guess
20160903 22:20:06< tad_> To me "user_" means "the user see's this, so be sure you want to do that"
20160903 22:20:51< tad_> The problem on teams is it really needs a [team] tag for the shared stuff.
20160903 22:20:58< celmin> If I recall correctly, isn't team_id actually a comma-separated list?
20160903 22:21:10< celmin> team_name, sorry
20160903 22:21:18< tad_> And the problem with randomized sides is it should have been a mapping instead of re-writing the config file on-the-fly like that
20160903 22:21:30< celmin> I think it's possible to have non-reciprocal teams somehow.
20160903 22:21:46< zookeeper> what shared stuff?
20160903 22:21:47< celmin> eg, side 1 treats side 2 as an ally, but side 2 treats side 1 as an enemy.
20160903 22:22:02< celmin> I forget how that's done.
20160903 22:22:25< zookeeper> easily by changing the teams on side turn, i'd imagine
20160903 22:22:40< tad_> zookeeper: A team can have a name, no problem. But if the side is setting a name, that's a side thing .. unfortunately it's called "user_team_name"
20160903 22:23:03< vultraz> tad_: oh I see what you mean
20160903 22:23:08< vultraz> utn is like a side_name
20160903 22:23:16< vultraz> whereas team_name/id allies sides
20160903 22:23:24-!- EliDupree [~quassel@idupree.com] has quit [Ping timeout: 250 seconds]
20160903 22:23:26< zookeeper> tad_, i don't see what the problem with that is
20160903 22:23:29< tad_> vultraz: At least as it's used in SP and on your Status Table
20160903 22:24:02< vultraz> tad_: yeah, that's a good use for it, anyway
20160903 22:24:43< tad_> The problem is with randomized sides for MP it you (1) randomize then (2) take the user_team_name from the first [side] (after randomizing) which has a given team_name and (3) copy that user_team_name to all sides having that team_name
20160903 22:24:54< vultraz> tad_: what about this: team = id, display_name = translatable name
20160903 22:25:18< tad_> So if each side has a different user_team_name, every play on randomized MP should get a different name .. or at least that's how I read it
20160903 22:25:59< zookeeper> tad_, does not compute. i don't know the context of what you're trying to fix.
20160903 22:26:26-!- EliDupree [~quassel@idupree.com] has joined #wesnoth-dev
20160903 22:26:27< tad_> vultraz: Personally, I don't care what they're called. But if you want to rename, don't re-use existing names (too confusing) and either make the old-new mapping obvious to well documented and easy-to-find
20160903 22:26:43< zookeeper> best option: just don't rename.
20160903 22:26:49< vultraz> team and display_name aren't used
20160903 22:26:59< vultraz> and of course zookeeper loves not renaming things.
20160903 22:27:01< vultraz> :)
20160903 22:27:16< celmin> I thought there was a way to do it without dynamically changing team name.
20160903 22:27:19< tad_> zookeeper: On Menu|Status Table you see the team name. But if you play  from End Scenario what you see changes from what you see if your Load the next scenario
20160903 22:27:31< celmin> Maybe it was something slightly different that I was thinking of.
20160903 22:27:58< zookeeper> tad_, oh, right
20160903 22:29:34< vultraz> celmin: well, one can have sides a and b fight yet c is allied with both
20160903 22:29:44< celmin> That's what I was thinking of, probably.
20160903 22:30:08-!- Duthlet [~Duthlet@dslb-188-104-247-024.188.104.pools.vodafone-ip.de] has quit [Quit: leaving]
20160903 22:30:24< tad_> zookeeper: My concern is the mess with user_team_name appears (from my reading) to have arisen from adding the ability to randomize sides for MP and the desire to keep allied sides together and give them a display/translatable name
20160903 22:31:01< zookeeper> yeah, sounds likely
20160903 22:31:08< tad_> zookeeper: So it's possible my 'fix' for SP will break something for MP or randonized-sides MP.
20160903 22:31:27< tad_> I don't do MP so I dunno
20160903 22:36:40< tad_> OK. I'm off to cook dinner.
20160903 22:36:42 * tad_ waves
20160903 22:36:46-!- tad_ [add94167@gateway/web/freenode/ip.173.217.65.103] has quit [Quit: Page closed]
20160903 22:41:20< celmin> Now it seems to be stuck on "Connecting to server"...
20160903 22:41:41-!- Appleman1234 [~Appleman1@KD119104107133.au-net.ne.jp] has joined #wesnoth-dev
20160903 22:42:52-!- edgrey [~edgrey@178.204.108.159] has quit [Ping timeout: 240 seconds]
20160903 22:43:04< celmin> Well, it moved on, but still seems stuck...
20160903 22:43:16< celmin> Stopped animating.
20160903 22:46:18 * celmin kills it and starts over.
20160903 22:52:27< celmin> :|
20160903 23:05:31-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Quit: Konversation terminated!]
20160903 23:11:28 * celmin wonders if vultraz is around
20160903 23:11:33< vultraz> yes
20160903 23:11:57< celmin> Mind connecting and hosting a game?
20160903 23:13:22< celmin> Aww, crashed.
20160903 23:13:58< vultraz> still shows you as in-game
20160903 23:15:15-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity]
20160903 23:15:49< celmin> It's probably a "something is corrupted" crash - bad free in vector destructor.
20160903 23:16:38< celmin> I guess I'll try again in case it was a fluke.
20160903 23:16:54< celmin> I have no idea how to fix it either way, though.
20160903 23:17:52< celmin> Using the loading screen for all the "connecting" bits works, but I can't seem to prevent it from switching to "Loading" between stages.
20160903 23:18:16< celmin> Guessing it's doing that every time a new window instance is opened.
20160903 23:18:29-!- ChipmunkV [~vova@static-89-94-113-91.axione.abo.bbox.fr] has quit [Quit: ChipmunkV]
20160903 23:20:23< vultraz> celmin: it still showed you as in-game and in-lobby
20160903 23:20:27< vultraz> if you were gone, that's a bug
20160903 23:20:44< celmin> Well, I crashed.
20160903 23:20:52< celmin> I assume it stopped showing me by now though?
20160903 23:21:37< vultraz> no
20160903 23:21:52< celmin> I haven't gotten a "choose a new name" popup though...
20160903 23:22:05< celmin> Then again, it might be frozen.
20160903 23:22:45< celmin> Am I shown now?
20160903 23:22:58< vultraz> no
20160903 23:23:28< celmin> That might be a bug… if it continues to show a player even after the connection has been reset… (I'm assuming the system resets the connection when terminating the program)
20160903 23:24:26< celmin> I kind of want to push this and get someone else to fixup. >_>
20160903 23:25:00< celmin> Okay, now I'm connected again finally.
20160903 23:27:00< celmin> vultraz: ^
20160903 23:29:35< celmin> It's becoming a bit unresponsive...
20160903 23:31:38< celticminstrel> Which turns out to be VNC's fault. :/
20160903 23:31:45 * celticminstrel poke vultraz 
20160903 23:35:25< celticminstrel> Another crash. :|
20160903 23:35:49< celticminstrel> Same general idea - a destructor.
20160903 23:36:17< celticminstrel> There's something very wrong somewhere.
20160903 23:36:37< celticminstrel> I have no idea whether it's in my changes specifically. Hope not.
20160903 23:37:20< celticminstrel> Now in swap hell.
20160903 23:38:04< celticminstrel> Connecting.
20160903 23:38:15< celticminstrel> Swap hell didn't last long this time.
20160903 23:38:39< shadowm> celticminstrel: Download more RAM.
20160903 23:38:43< shadowm> ;p
20160903 23:38:52< celticminstrel> Ha ha
20160903 23:39:04< celticminstrel> Crashed, trying again.
20160903 23:40:28< celticminstrel> When XCode says "Running Wesnoth", it doesn't really mean Wesnoth is running.
20160903 23:40:34< celticminstrel> It means gdb/lldb has launched.
20160903 23:41:30< vultraz> for the record you're no longer shown
20160903 23:41:38< celticminstrel> Who here compiles with scons?
20160903 23:42:21< celticminstrel> Is there a way to set the compiler (eg CC=blah) or add extra flags?
20160903 23:42:36< celticminstrel> (Well, CPP=blah I guess.)
20160903 23:43:55< celticminstrel> Crashed again. :|
20160903 23:44:06< loonycyborg> celticminstrel: there are some ways, debeping on what exactly you want to accomplish
20160903 23:44:07< vultraz> and you're gone
20160903 23:44:50< celticminstrel> loonycyborg: I want to build with clang++-mp-3.8 and, if possible, the -fsanitize=address flag.
20160903 23:45:48< celticminstrel> I see there's "cxxtool"...
20160903 23:46:30< loonycyborg> cxxtool=clang++-mp-3.8 sanitize=address
20160903 23:46:47< loonycyborg> "scons cxxtool=clang++-mp-3.8 sanitize=address"
20160903 23:47:05< celticminstrel> Huh? No quotes around that?
20160903 23:47:26< celticminstrel> Is "sanitize" an actual flag understood by the SConstruct?
20160903 23:47:30< loonycyborg> there's a custom sanitize option int the script
20160903 23:47:35< celticminstrel> I see.
20160903 23:47:36< loonycyborg> I've added it recently
20160903 23:48:36< celticminstrel> Does it take a comma-separated list?
20160903 23:48:41< loonycyborg> yes
20160903 23:49:03< loonycyborg> it passes it directly to -fsanitize
20160903 23:49:19< celticminstrel> Might try thread too later, then.
20160903 23:49:53< celticminstrel> The config seems to be cached. Oh... that's probably from when I built translations, I guess.
20160903 23:50:12< celticminstrel> vultraz: You might as well log out, since this is likely to take awhile.
20160903 23:50:23< vultraz> ok
20160903 23:50:26< loonycyborg> yes, options like that persist between invocations
20160903 23:50:54-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 260 seconds]
20160903 23:52:45< celmin> Wait, so, if I say "jobs=3" or whatever, then that becomes the default for subsequent calls?
20160903 23:53:42< loonycyborg> yes
20160903 23:54:03< loonycyborg> while command line switches don't persist
20160903 23:54:15< loonycyborg> like -j3
20160903 23:54:29< loonycyborg> so jobs was added to make a version of -j that persists
20160903 23:54:34< celmin> I see.
20160903 23:54:54< celmin> It can't find tuple.
20160903 23:55:06< celmin> Presumably I need to pass -stdlib=libc++ or something.
20160903 23:56:23< loonycyborg> build/config.log contains logs of test program builds
20160903 23:56:58< celmin> Hm?
20160903 23:57:35< celmin> I mean it's getting a compile error about missing tuple.
20160903 23:57:54< loonycyborg> boost.tuple?
20160903 23:57:59< celmin> std::tuple
20160903 23:58:16< celmin> Which is why I said "presumably I need to pass -stdlib=libc++ or something"
20160903 23:58:27< loonycyborg> hmm ok
--- Log closed Sun Sep 04 00:00:30 2016