--- Log opened Sun Sep 23 00:00:45 2018 20180923 00:23:24< irker381> wesnoth/wesnoth:1.14 josteph 3f3ee23093 manpage: Fix option name typos and updat AppVeyor: All builds passed 20180923 00:50:34-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] 20180923 02:33:47< irker381> wesnoth/wesnoth:master mattsc 81ef9ea390 Merge pull request #3506 from mattsc/lua AppVeyor: All builds passed 20180923 04:28:19<+wesdiscordbot> celticminstrel: It definitely sounds interesting. Unless I'm completely misremembering, it was been mentioned for 2.0 that it might be possible/worthwhile to keep WML for defining more structural things/data as well as keeping lua available for UMC, so it sounds like reducing the use of ActionWML would be beneficial whether or not Wesnoth ends up migrating to a new engine. 20180923 05:34:17-!- irker381 [~irker@wesnoth/umc-dev/bot/shikadibot] has quit [Quit: transmission timeout] 20180923 05:35:23< celticminstrel> I'll try to finish the post tomorrow, then (I've got it drafted and saved in the forum). 20180923 05:39:01<+wesdiscordbot> @Pentarctagon possible, but likely not probable, and definitely not preferable 20180923 05:39:37< celticminstrel> ??? 20180923 05:39:55< celticminstrel> Definitely preferable. 20180923 05:40:14< celticminstrel> And if it doesn't happen, I really don't think your prototype will end up doing very well. 20180923 05:40:32<+wesdiscordbot> Not preferable because, as I said, the design of WML is intrinsically tied to the preprocessor and text substitution macros. 20180923 05:40:38< celticminstrel> So what? 20180923 05:41:08<+wesdiscordbot> Not to mention the whole design of WML macros is very awkward. 20180923 05:41:14< celticminstrel> How do you mean? 20180923 05:42:06<+wesdiscordbot> Look at this macro mess I had to write, for example https://github.com/wesnoth/wesnoth/blob/master/data/gui/widget/button_default.cfg 20180923 05:42:33< celticminstrel> I have no idea what you're getting at. 20180923 05:44:54<+wesdiscordbot> It got even more of a mess when I tried to expand it to 25x25 buttons. Macros are very rigid for more than simple text block substitution and can quickly get messier and messier as you try to add more arguments that have to fall down to 2, 3 nested layers.... Granted, without ActionWML we eliminate most of the need for macros, but then there's little that we're left with to use WML for and we'd do better using Lua 20180923 05:44:55<+wesdiscordbot> directly or some config format whatever engine we use accepts rather than reimplementing WML. 20180923 05:45:36< celticminstrel> Except we wouldn't be compatible, which is going to lose a ton of addon devs. 20180923 05:45:54<+wesdiscordbot> And if the way WML works is changed significantly (such as eliminating the preprocessor), I don't see how it would make it any easier on people to learn that X once did Y but now does Z. 20180923 05:46:11< celticminstrel> What are you even talking about now? 20180923 05:46:22< celticminstrel> "X once did Y but now does Z" Huh? 20180923 05:46:53< celticminstrel> Also why would we eliminate the preprocessor? Designing things so that it's not needed as much is one thing, but there's no need to eliminate it altogether. 20180923 05:47:00<+wesdiscordbot> What if something done with the same WML syntax functioned differently in 1.x than 2.x 20180923 05:47:10< celticminstrel> Well, obviously we wouldn't do that. 20180923 05:47:19<+wesdiscordbot> Then we can't eliminate macros. 20180923 05:47:21<+wesdiscordbot> Macros are baaaad. 20180923 05:47:30< celticminstrel> Correct. We can't eliminate macros. 20180923 05:47:41< celticminstrel> We can, however, engineer things so that people have less need for them. 20180923 05:47:45<+wesdiscordbot> We should eliminate text macros 20180923 05:47:55< celticminstrel> For example, the new way of defining unit abilities that I proposed awhile back. 20180923 05:48:03< celticminstrel> We can't eliminate text macros. 20180923 05:48:10<+wesdiscordbot> We can if we drop WML 20180923 05:48:12<+wesdiscordbot> You see 😛 20180923 05:48:15< celticminstrel> But we can't drop WML. 20180923 05:48:20<+wesdiscordbot> We can drop WML. 20180923 05:48:37< celticminstrel> I'm not going to get into a can we can't we thing with you here. 20180923 05:48:38<+wesdiscordbot> Again, what in WML is there left to configure without actionWML? 20180923 05:48:44-!- boucman [~rosen@wesnoth/developer/boucman] has joined #wesnoth-dev 20180923 05:48:45< celticminstrel> Uhh... everything? 20180923 05:48:53< celticminstrel> Like, literally everything? 20180923 05:49:01< celticminstrel> Units, terrain types. 20180923 05:49:07< celticminstrel> Game config parameters. 20180923 05:49:16< celticminstrel> Hotkeys, preference entries. 20180923 05:49:19< celticminstrel> Help entries. 20180923 05:49:23< celticminstrel> I've probably missed something. 20180923 05:49:33< celticminstrel> Oh, scenarios of course. 20180923 05:49:56< celticminstrel> And other addon tags - campaigns, modifications, etc. 20180923 05:50:30<+wesdiscordbot> And these couldn't as well be written in JSON or any other config language? 20180923 05:50:38<+wesdiscordbot> Why are you people so attached to the clunkiest markup language ever? 20180923 05:50:47<+wesdiscordbot> Even XML is easier to write than WML. 20180923 05:51:15< celticminstrel> You think so? 20180923 05:51:20<+wesdiscordbot> Yes. 20180923 05:51:28< celticminstrel> I don't think I agree. 20180923 05:51:40< celticminstrel> I'd probably rate them about the same, to be honest. 20180923 05:51:44<+wesdiscordbot> 20180923 05:52:07< celticminstrel> Sure. 20180923 05:52:08<+wesdiscordbot> [some-tag] attribute1=value1 attribute2=value2 [/some-tag] 20180923 05:52:15< celticminstrel> XML has some nice features, I guess. 20180923 05:52:25<+wesdiscordbot> I actually talked about this ages ago. 20180923 05:52:35<+wesdiscordbot> The only thing WML has going for it over WML is built-in i18n. 20180923 05:52:39<+wesdiscordbot> *over XML 20180923 05:52:50<+wesdiscordbot> Let's not forget that WML also lacks nice things like specific data types. Ie, in JSON you can have key: [ a, b, c ]. In wml, that's [key] value=a [/key] [key] value=b [/key] [key] value=c [/key] 20180923 05:53:16<+wesdiscordbot> I'm going to add something to that before celticminstrel jumps to try to shoot it down using a biased argument. 20180923 05:53:19< celticminstrel> @Vultraz To answer your question, yeah, sure, they could be written in JSON. But that requires you to rewrite them in JSON, which is the problem here. 20180923 05:54:07<+wesdiscordbot> Back in the day, a certain person went to great lengths to add some flimsy data type checks to wmlscope, a tool which we were supposed to use to inspect mainline WML on a regular basis before releases but ended up unofficially abandoning due to its inherent flimsiness. 20180923 05:54:09< celticminstrel> ie, the problem is needing to rewrite things, which ideally we want to avoid, because needing to rewrite things will drive addon devs off. 20180923 05:54:11<+wesdiscordbot> How did he achieve this you ask? 20180923 05:54:26<+wesdiscordbot> By attempting to enforce a naming convention on WML macro arguments. 20180923 05:54:45<+wesdiscordbot> Emphasis on attempting, because no-one has time for that crap. 20180923 05:55:28< celticminstrel> Also @Vultraz - WML could totally be extended to have built-in support for key=a,b,c 20180923 05:55:44<+wesdiscordbot> I think the fact that wmllint and wmlscope have utterly failed their mission statement over the years even before their maintainer decided to move onto greener pastures (pattern???) should speak oodles about how intrinsically broken the language is. 20180923 05:55:55<+wesdiscordbot> ok, then how is key=a,b,c any different from the string value "a,b,c"? 20180923 05:55:59< celticminstrel> Which is a convention already used all over the place in the Wesnoth codebase. 20180923 05:56:06< celticminstrel> Well, that's certainly a good question. 20180923 05:56:21< celticminstrel> And the answer would probably have to be - it isn't different. The two are convertible to each other. 20180923 05:56:52<+wesdiscordbot> there's also the crap that has to do with actually accessing data from, say, Lua. All that child_range crap? Other languages don't have to deal with that. 20180923 05:56:56<+wesdiscordbot> Ok I should know by now not to attempt to participate in these discussions since people just ignore me lol 20180923 05:57:12< celticminstrel> Sorry shadowm. 20180923 05:57:22< celticminstrel> I just didn't have anything to say in response to that. 20180923 05:57:35<+wesdiscordbot> honestly, we should listen when shadowm says something is bad. she has more experience than either of us with this mess. 20180923 05:57:42< celticminstrel> @Vultraz I don't see what's wrong with child_range. 20180923 05:57:44<+wesdiscordbot> That is something to say in and of itself celticminstrel. 20180923 05:57:48< celticminstrel> Maybe so. 20180923 05:58:25< celticminstrel> child_range is basically an iterator. It's the same as keys or values on a map. 20180923 05:58:40< celticminstrel> So I would say... other languages do kinda deal with that sort of thing? 20180923 05:59:52< celticminstrel> I'll admit that WML is a bit weird as a language, but it does serve it's purpose decently well in my opinion. I wouldn't object to Wesnoth 2.x deprecating it (probably indefinitely) and phasing in a new configuration format, but dropping it should really not be on the table. 20180923 06:00:30< celticminstrel> If you really must drop it, then what you need to do is provide a separate automatic converter program that converts WML to the new format. 20180923 06:00:57< celticminstrel> Though I have a feeling that that could still drive away a few addon devs... but probably not as many. 20180923 06:00:59<+wesdiscordbot> here's the biggest argument against keeping WML in any form: in order to do so, we'd essentially need to reimplement the preprocessor, and the WML tokenizer in a new engine, add all the Lua accessor functions like child_range, figure out how this data is supposed to be managed on the engine side, which means either reimplementing the config class which is a bad idea, or figuring some internal bridge to convert between 20180923 06:01:00<+wesdiscordbot> WML and whatever data objects the engine uses because please god let's not use the config class please. All that is in and of itself a huge amount of work, and if we're gonna drop WML anyway? Not worth it. 20180923 06:01:28<+wesdiscordbot> Reimplementing the config class is not a bad idea in and of itself. 20180923 06:01:32< celticminstrel> I have no idea why you hate the config class so much. 20180923 06:01:44<+wesdiscordbot> It doesn't even need to be reimplemented. 20180923 06:01:47< celticminstrel> TBH it's not all that much different from that json class that you seem to really like. 20180923 06:01:51<+wesdiscordbot> It's the most freestanding piece of work in the entire game. 20180923 06:02:03<+wesdiscordbot> It doesn't depend upon anything else other than helper functions from src/serialization. 20180923 06:02:04< celticminstrel> That's... very likely to be true, actually...? 20180923 06:02:13< celticminstrel> I haven't looked at it in detail. 20180923 06:02:24< celticminstrel> It does depend on t_string which I think pulls in a lot of other stuff. 20180923 06:02:33< celticminstrel> But that dependency could be excised by templatizing it. 20180923 06:02:46< celticminstrel> Or I guess templatizing the attribute value class, to be more specific. 20180923 06:03:09<+wesdiscordbot> I'd regard t_string as an implementation detail (not just of WML even). 20180923 06:03:16<+wesdiscordbot> It can be swapped with whatever else if needed. 20180923 06:03:25< celticminstrel> Right, that's why I said templatize. 20180923 06:03:31<+wesdiscordbot> Or kept since the need for i18n isn't going away. 20180923 06:03:36< celticminstrel> Because it can be swapped with something else that converts to string. 20180923 06:03:49<+wesdiscordbot> It doesn't quite work that way. 20180923 06:03:51<+wesdiscordbot> The way I was thinking of it at least was: 1.15 - 1.16: deprecate ActionWML 1.17 - 1.18: Remove ActionWML -if 2.0- also deprecate all WML and switch to XML/JSON/whatever 2.1: remove WML 20180923 06:03:52< celticminstrel> Though I guess templatizing the config class might not be the only way to deal with that. 20180923 06:04:00<+wesdiscordbot> A lot of the time t_strings from WML as consumed as t_strings instead of strings. 20180923 06:04:04<+wesdiscordbot> *are consumed 20180923 06:04:19<+wesdiscordbot> Until they need to be displayed on a UI or whatever. 20180923 06:04:28< celticminstrel> @Pentarctagon IMO 2.1 is too early to remove WML, and ActionWML probably shouldn't be removed until 2.0. 20180923 06:04:40<+wesdiscordbot> Godot already has a json parser https://godot.readthedocs.io/en/3.0/classes/class_json.html 20180923 06:05:02< celticminstrel> Well, good for Godot. Unreal has one too. 20180923 06:05:27<+wesdiscordbot> I don't feel JSON is a good choice. 20180923 06:05:42< celticminstrel> Well, certainly WML wouldn't map that well to JSON. 20180923 06:05:46< celticminstrel> It would map a lot better to XML. 20180923 06:05:49<+wesdiscordbot> I don't care about mapping WML. 20180923 06:05:57<+wesdiscordbot> JSON is just not a good choice for a configuration language, period. 20180923 06:06:05< celticminstrel> Shrug! 20180923 06:06:06<+wesdiscordbot> I would prefer something like XML over JSON. JSON is not as readable to me by a long shot. 20180923 06:06:06<+wesdiscordbot> I'm just going to link to this: https://www.lucidchart.com/techblog/2018/07/16/why-json-isnt-a-good-configuration-language/ 20180923 06:10:22<+wesdiscordbot> celticminstrel: assuming 2.0+ keeps the same odd-even naming for dev-stable releases and that each stable release is ~2 years apart starting from now, WML wouldn't be removed until 2.2 in the ~mid 2020s. That seems like plenty of time to me. 20180923 06:10:38<+wesdiscordbot> well 20180923 06:10:46<+wesdiscordbot> wouldn't be removed for most users 20180923 06:11:13<+wesdiscordbot> given how few use the dev series 20180923 06:14:04< celticminstrel> ... 20180923 06:14:12< celticminstrel> I just realized it's past 2 am. 20180923 06:15:27< celticminstrel> Okay, in that case I'll try to finish my draft post tomorrow. 20180923 06:16:51-!- celticminstrel is now known as celmin|sleep 20180923 06:27:15-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity] 20180923 06:51:57-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180923 06:52:02-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180923 07:21:21< loonycyborg> I think YAML would be better than JSON 20180923 07:23:05<+wesdiscordbot> I'm not really advocating json 20180923 07:27:58< loonycyborg> But what would be a better alternative? Would you script wesnoth scenarios in same format as, say, .travis.yml file is using? 20180923 07:28:48<+wesdiscordbot> I'm trying to figure out how things are done with GDScript 20180923 07:29:49< loonycyborg> GDScript is custom scripting language inspired by python, but it's not a markup language 20180923 07:30:57<+wesdiscordbot> what do they use for serialized data, then? 20180923 07:31:30< loonycyborg> I have no idea atm 20180923 07:32:16<+wesdiscordbot> I think they might use json 20180923 07:32:17<+wesdiscordbot> :thonk: 20180923 07:32:24<+wesdiscordbot> https://godot.readthedocs.io/en/3.0/classes/class_@gdscript.html#class-gdscript-to-json 20180923 07:32:45<+wesdiscordbot> https://godot.readthedocs.io/en/3.0/tutorials/io/saving_games.html 20180923 07:34:25<+wesdiscordbot> so, GDScript + JSON. 20180923 07:47:13< loonycyborg> I think we should use godot from C++, just as replacement for SDL 20180923 07:47:18< loonycyborg> it seems to be possible 20180923 07:47:28< loonycyborg> and restrict its use only for graphics 20180923 07:47:41< loonycyborg> change of scenario scripting can be done as separate project 20180923 07:47:53<+wesdiscordbot> I disagree. 20180923 07:49:29<+wesdiscordbot> It would be foolish not to use Godot's editor. 20180923 07:50:38< loonycyborg> pretty sure we can plug it in somehow for maps only 20180923 07:51:07<+wesdiscordbot> It's not just drawing maps, it also has a whole bunch of other stuff for setting up stuff like, say, animations 20180923 07:51:34< loonycyborg> also some people might still prefer to work with files text directly 20180923 07:52:11< loonycyborg> we could offload map and animation and rest of graphics stuff to godot 20180923 07:52:21< loonycyborg> but still do main logic in wml 20180923 07:52:30< loonycyborg> and migrate to something else later 20180923 07:53:31<+wesdiscordbot> WML is a terrible language for logic... 20180923 07:53:35<+wesdiscordbot> we sure as hell aren't going to code game logic in WML... 20180923 07:54:02< loonycyborg> I mean use it for same stuff as it's used currently 20180923 07:54:23< loonycyborg> need to utilize divide and conquer approach 20180923 07:54:35< loonycyborg> so you have constantly have something working 20180923 07:54:39< loonycyborg> or close to working 20180923 07:54:57<+wesdiscordbot> I've been discussing with jetrel where to start. 20180923 07:55:14<+wesdiscordbot> His opinion is first, to get a map drawing in Godot. 20180923 07:55:23<+wesdiscordbot> import the tiles and everything 20180923 07:55:28<+wesdiscordbot> well, a subset 20180923 07:55:32<+wesdiscordbot> get it drawing 20180923 07:58:25<+wesdiscordbot> from what I've seen, that should be very easy. 20180923 07:58:33<+wesdiscordbot> (emphasis on should) 20180923 09:13:50-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180923 09:13:56-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180923 10:41:10-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180923 11:01:25-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has joined #wesnoth-dev 20180923 11:07:32<+wesdiscordbot> leaving this here for future reference https://godot.readthedocs.io/en/3.0/tutorials/2d/using_tilemaps.html 20180923 13:20:32-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity] 20180923 13:54:42-!- irker221 [~irker@wesnoth/umc-dev/bot/shikadibot] has joined #wesnoth-dev 20180923 13:54:42< irker221> wesnoth: Jyrki Vesterinen wesnoth:master 79206d92d539 / src/actions/attack.cpp: Fix: Chance to hit wasn't protected from becoming negative https://github.com/wesnoth/wesnoth/commit/79206d92d5390d82a9a79cd6018b575de3aafea0 20180923 13:57:55<+wesdiscordbot> isn't signing an int redundant? 🤔 20180923 13:59:06<+wesdiscordbot> Yes. int is guaranteed to be signed. 20180923 13:59:46<+wesdiscordbot> "There are five standard signed integer types : “signed char”, “short int”, “int”, “long int”, and “long long int”. In this list, each type provides at least as much storage as those preceding it in the list. There may also be implementation-defined extended signed integer types. The standard and extended signed integer types are collectively called signed integer types. Plain ints have the natural 20180923 13:59:46<+wesdiscordbot> size suggested by the architecture of the execution environment; the other signed integer types are provided to meet special needs." 20180923 14:02:04< irker221> wesnoth: josteph wesnoth:1.14 ba8b94111fd7 / data/gui/window/mp_create_game.cfg: MP Create Game: Increase slider step size https://github.com/wesnoth/wesnoth/commit/ba8b94111fd743318d4ef442198ccb2949b9045c 20180923 14:03:02< irker221> wesnoth: josteph wesnoth:master 9abdc20841ac / data/gui/window/mp_create_game.cfg: MP Create Game: Increase slider step size https://github.com/wesnoth/wesnoth/commit/9abdc20841ac12db3bb8cf9338a7f7b4b205bb41 20180923 15:13:12-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180923 15:13:18-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180923 15:26:10< celmin|sleep> ...I think I just spotted a bug in [modify_side]? 20180923 15:26:23< celmin|sleep> Line 52 20180923 15:26:42< celmin|sleep> Pretty sure that should say wesnoth.set_side_id(side.side, cfg.flag or side.flag, cfg.color or side.color) 20180923 15:26:54< celmin|sleep> The difference being the addition of the ors. 20180923 15:26:57-!- celmin|sleep is now known as celticminstrel 20180923 15:29:42< celticminstrel> @Pentarctagon @Vultraz and anyone else interested - I posted in the ideas forum. 20180923 16:11:15<+wesdiscordbot> it would actually be pretty nice to have a scenario that could be completely written in lua 20180923 16:11:53<+wesdiscordbot> thing is right now, there are a lot of things you can't do in lua and end up having to use WML tables for anyway 20180923 16:11:59<+wesdiscordbot> and events in lua are pretty meh 20180923 16:21:20-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has joined #wesnoth-dev 20180923 16:28:17< celticminstrel> Filters are a bit nicer in WML IMO. 20180923 16:34:10< celticminstrel> BTW it would probably be better if you respond in the thread rather than in Discord, so all the responses are in one place for everyone to see. 20180923 17:03:48-!- irker221 [~irker@wesnoth/umc-dev/bot/shikadibot] has quit [Quit: transmission timeout] 20180923 17:28:48< celticminstrel> On an unrelated note, how much work would it be to add a hook into Lua loops that can potentially call into Wesnoth's UI code? 20180923 17:29:53< celticminstrel> I guess it would require patching Lua source, which isn't exactly good, but I think it could be worth it, effectively solving the problem of an infinite loop in Lua bringing the entire program down. 20180923 17:30:39< celticminstrel> I guess another alternative would be to run even event Lua as a coroutine, but that might be a little more complicated... 20180923 17:30:53< celticminstrel> And still wouldn't fully solve it. 20180923 17:35:57<+wesdiscordbot> you mean so after a loop has been going for X time or iterations a popup would appear? 20180923 18:18:46< celticminstrel> That wasn't quite what I was thinking. I was thinking to allow people to use the menu and quit the game in the middle of a loop. 20180923 18:30:29-!- ToBeCloud [uid51591@wikimedia/ToBeFree] has quit [Quit: Connection closed for inactivity] 20180923 18:30:45< celticminstrel> Would be nice if someone other than gfgtdf responded. 20180923 18:30:54< celticminstrel> In the thread I mean, not the Lua question. 20180923 18:37:44-!- gfgtdf [~Daniel@x4e341d40.dyn.telefonica.de] has joined #wesnoth-dev 20180923 19:11:03-!- gfgtdf [~Daniel@x4e341d40.dyn.telefonica.de] has quit [Quit: Leaving] 20180923 19:22:05-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev 20180923 19:48:32-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180923 19:48:39-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180923 20:22:13-!- gfgtdf [~Daniel@x4e341d40.dyn.telefonica.de] has joined #wesnoth-dev 20180923 20:33:22-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180923 20:33:28-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180923 20:50:06-!- irker370 [~irker@wesnoth/umc-dev/bot/shikadibot] has joined #wesnoth-dev 20180923 20:50:06< irker370> wesnoth/wesnoth:1.14 josteph ba8b94111f MP Create Game: Increase slider step siz AppVeyor: All builds passed 20180923 21:01:49<+wesdiscordbot> Celmin: that’s the kind of thing that should not be co sidered ight now 20180923 21:06:25<+wesdiscordbot> A UI Lua thing, that is 20180923 21:58:48-!- boucman [~rosen@wesnoth/developer/boucman] has quit [Remote host closed the connection] 20180923 22:13:23< celticminstrel> Why not? 20180923 22:13:46< celticminstrel> It's not like it actually depends on the implementation of that UI, it'd just be a call to a function much like what's already in place in the AI code. 20180923 22:19:20< gfgtdf> i think i'll write my own proposal about future of lua/wml. Not sure yet wehther i'll do seperate topic though. 20180923 22:21:19< gfgtdf> celticminstrel: i currrently don't think its worth, mosrly because you don't know how often you should call the ui, if you do it too often it'll be slow, if you do it not often eough, you'll still ge freezes if the loop content takes very long time by itself. 20180923 22:22:02-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 244 seconds] 20180923 22:24:13< irker370> wesnoth/wesnoth:master josteph 9abdc20841 MP Create Game: Increase slider step siz AppVeyor: All builds passed 20180923 22:24:27< gfgtdf> celticminstrel, if you want to 'complteley fix' this issue i think your only option is to use threading and to run the umc in a seperate thread seperated from the ui basuc ui. Which is also rather a lot work and probably not worth it currently. 20180923 22:27:14-!- TheJJ [~rofl@ipbcc06208.dynamic.kabel-deutschland.de] has quit [Ping timeout: 272 seconds] 20180923 22:28:25-!- TheJJ [~rofl@ipbcc06208.dynamic.kabel-deutschland.de] has joined #wesnoth-dev 20180923 22:36:01<+wesdiscordbot> celticminstrel: it's a large project. We should avoid such large projects right now, and that includes your ActionWML proposals 20180923 22:36:23<+wesdiscordbot> I glanced at your post, and they're things we could revisit later, but right now let's not. 20180923 22:36:34<+wesdiscordbot> Especially since you haven't finished your existing big projects 20180923 22:36:36<+wesdiscordbot> like 20180923 22:36:39<+wesdiscordbot> schema 20180923 22:43:06-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180923 22:43:12-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180923 23:00:17< gfgtdf> i think we shodul move away from wml for wesnoth data representation (in particular to get rid of all the wml.tag. codes in lua) and use soemthing that matches the lua object format better, schema will be very helpful for that because its needed for an automatic wml-> lua object conversion by the engine to get compatability. (does [tag] mean an 1-element array to just a normal rable) 20180923 23:01:09<+wesdiscordbot> well... yes? my idea is indeed to ditch WML entirely if we do get wesnoth running in godot or whatever. 20180923 23:01:33< celticminstrel> @Vultraz Uhh... I don't think that project is as large as you seem to think. 20180923 23:01:36<+wesdiscordbot> unfortunately it seems the built-in serialization option in godot is json 20180923 23:01:50< celticminstrel> So don't use the built-in serialization, right? :P 20180923 23:02:05<+wesdiscordbot> well 20180923 23:02:16<+wesdiscordbot> we could use it for stuff the game writes like saves 20180923 23:02:24< celticminstrel> True. 20180923 23:02:35<+wesdiscordbot> saves don't need comments, et al 20180923 23:03:18< celticminstrel> That said... 20180923 23:03:49< celticminstrel> The current setup is nice in some ways because the scenario config is in fact a subset of the saved game format. 20180923 23:04:03<+wesdiscordbot> true 20180923 23:04:51<+wesdiscordbot> we'll figure something out 20180923 23:05:13< celticminstrel> gfgtdf: If you have a genuine alternative to my proposal I suppose it can go in the same thread... 20180923 23:05:44<+wesdiscordbot> celmin: that said, you should still finish the schema and your other projects before delving into new ones 20180923 23:06:08<+wesdiscordbot> the schema is a solid 1.x series improvement 20180923 23:06:09< celticminstrel> Well, I haven't delved into this yet. I just made a proposal and opened a discussion. 20180923 23:06:42<+wesdiscordbot> true 20180923 23:06:43< celticminstrel> It's not at a stage where I'd feel comfortable starting on the programming of it. 20180923 23:06:50< celticminstrel> And yes I really need to finish schema. 20180923 23:07:03< celticminstrel> Not that you get to tell me what to do, but I was thinking that anyway so it's fine this time. 20180923 23:07:06< gfgtdf> @vultraz: serilisation for what exactly i wonder? most objects in wensoth will still be wesnoth object not godot graphics. objects 20180923 23:07:35<+wesdiscordbot> godot is more than just graphics 20180923 23:07:38<+wesdiscordbot> i mean 20180923 23:07:45<+wesdiscordbot> it has capabilities for more than just graphics 20180923 23:08:05< celticminstrel> It's an engine, so duj 20180923 23:08:09< celticminstrel> duh 20180923 23:08:22<+wesdiscordbot> yes 20180923 23:08:25< gfgtdf> celticminstrel, not exactly alternive for events well it not even finshed yet. 20180923 23:08:45<+wesdiscordbot> which is why i'm confused why gfgtdf says the only thing godot would serialize is graphics opjects 20180923 23:10:29< gfgtdf> hmm maybe its better to delay this until we ahve somethign working in godot. 20180923 23:10:38< celticminstrel> Anyway Vultraz, you don't have any opinions on your own about how to phase out ActionWML? 20180923 23:10:51< celticminstrel> Delay what now? 20180923 23:11:22< gfgtdf> this discussion about whetehr we will use json for savegmes.. 20180923 23:11:27< celticminstrel> Ah. 20180923 23:11:31< celticminstrel> Yeah that seems fair. 20180923 23:12:20<+wesdiscordbot> I honestly don't see the point in phasing out actionwml in 1.x. I say it's better to leave actionWML in 1.x and just ditch it in 2.x. We'll never achieve the same result by a phase-out than a complete, ground-up replacement, so doing that just means two conversions for people to do: 1.14 -> next stable, to ditch the deprecated action wml, then next stable -> 2.0 to switch to whatever new system we come up with 20180923 23:12:38< celticminstrel> I still cannot accept this approach. 20180923 23:13:04< celticminstrel> I could maybe accept it if an alternative to it were added in 1.x and WML as a data language were retained in 2.x. 20180923 23:13:30< celticminstrel> (Which would mean there's still only one conversion for people to make.) 20180923 23:14:06<+wesdiscordbot> At this point I won’t convince you that wml needs to be ditched completely 20180923 23:14:15< celticminstrel> You're correct. 20180923 23:14:39< celticminstrel> Unfortunately it seems that I also won't convince you that it needs to be retained for compatibility purposes. 20180923 23:15:02< celticminstrel> So feel free to make your prototype, but I doubt it's going to be much use at this rate. 20180923 23:15:44<+wesdiscordbot> why? 20180923 23:15:58< celticminstrel> Because lack of compatibility. 20180923 23:16:05<+wesdiscordbot> ¯_(ツ)_/¯ 20180923 23:16:15<+wesdiscordbot> we shall see 20180923 23:16:39< celticminstrel> You're trying to make people switch from one ecosystem to a new one. It doesn't work that way. 20180923 23:17:24<+wesdiscordbot> we shall see 20180923 23:17:26< gfgtdf> i think that whatever data language we have in 2.0 , we can still mostlikeley add a compatability layer automatic conversion afterwards. 20180923 23:17:36< celticminstrel> MS DOS succeeded because it was backwards compatible. Windows 3.1 succeeded where Windows 1 and 2 did not because it was backwards compatible with DOS. People aren't going to switch just because you say "this is better". If they can't bring their old addons with them, they're not going to switch. 20180923 23:17:49< celticminstrel> If you can add a compatibility layer as gfgtdf says, then it could work. 20180923 23:17:51<+wesdiscordbot> apple begs to differ 20180923 23:18:01< celticminstrel> Not really. 20180923 23:18:27< celticminstrel> Please point to specific cases. 20180923 23:18:51<+wesdiscordbot> hard-dropping 32 bit apps for one? 20180923 23:19:01< celticminstrel> That's not an example of this. 20180923 23:19:11<+wesdiscordbot> deprecating OGL? 20180923 23:19:18< celticminstrel> Also doesn't satisfy. 20180923 23:19:28< celticminstrel> In both cases there's a period of compatibility. 20180923 23:19:36< loonycyborg> also 20180923 23:19:45< celticminstrel> In the latter case that period just began. 20180923 23:19:46<+wesdiscordbot> alright, fine 20180923 23:19:48< loonycyborg> what will be workflow for addon makers? 20180923 23:19:53< loonycyborg> with godot 20180923 23:19:58<+wesdiscordbot> I don't know yet 20180923 23:20:06<+wesdiscordbot> because I only just downloaded the damn thing the other day 20180923 23:20:11< celticminstrel> Heh. 20180923 23:20:32<+wesdiscordbot> once we get more things figured out we can better say 20180923 23:20:53< loonycyborg> well I did read up on it too 20180923 23:20:54<+wesdiscordbot> at this point we're both arguing about amorphous possibilities 20180923 23:20:56< celticminstrel> Out of curiosity, are you doing this alone or do you have other volunteers helping? 20180923 23:21:12< loonycyborg> and it seems it mostly geared for making standalone games 20180923 23:21:19< loonycyborg> and not addons for games 20180923 23:21:35< gfgtdf> so i think everyone agrees on porting all nonbreakign api back to 1.14 and adding something like 'api levels' (wesnoth_min_version in addons)? In the forum thread we it looks like we have everyaones ok except @jyrkive who didnt comment on that part yet. 20180923 23:21:56<+wesdiscordbot> dave and jetrel are willing to help, I think 20180923 23:22:38< celticminstrel> gfgtdf: I agree on the principle but I'd argue it makes more sense to actually call it 1.15 but as a stable series instead of a devel series, because I think that should make the minimum version system easier while still allowing for easy bugfix-only releases. 20180923 23:22:59<+wesdiscordbot> that's probably easier, yes 20180923 23:23:15< loonycyborg> so abolish dev releases altogether? 20180923 23:23:30< loonycyborg> in favor of dedicated feature branches 20180923 23:23:35<+wesdiscordbot> yes. 20180923 23:23:37<+wesdiscordbot> that was the idea 20180923 23:23:53< celticminstrel> Basically. 20180923 23:24:30< loonycyborg> well that solves one of problems with my semver proposal 20180923 23:26:13<+wesdiscordbot> i think we're also basically agreed to my proposed path of action, then? 20180923 23:26:21< gfgtdf> for the 1.x branch i currently vote for somethign like simver with a 1. prefix, that is raise the 1.x for every compatabiltiy break, but allow simple features to be added in 1.x.y 20180923 23:27:02<+wesdiscordbot> ie, avoid major features while switching to a rolling release model for 1.x to allow WML improvements and bugfixes while we explore godot forseries 2? 20180923 23:27:30< gfgtdf> also if it is fulyl compatible i suggest, not raising to 1.15, of all addons etc work as is, there is no reason to make uses have toreset their usadata folder 20180923 23:27:40< celticminstrel> I'm not really agreed to the totality of it, more like resigned to it... but the rolling release side of things I can agree to. 20180923 23:27:53< celticminstrel> gfgtdf: You do have a point there, admittedly... 20180923 23:28:35< celticminstrel> I mean we could bake it in that the addons folder is wesnoth1.14 hearafter even for 1.15+, but that might be weird. :P 20180923 23:35:20-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] --- Log closed Mon Sep 24 00:00:46 2018