--- Log opened Thu May 10 00:00:40 2018 20180510 00:18:22< celticminstrel> "celmin for constantly phrasing my own arguments better than I ever can." o.o 20180510 00:29:56-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] 20180510 00:59:38< irker663> wesnoth/wesnoth:1.14 Jyrki Vesterinen 3c344e8da5 Disallow units with negative HP AppVeyor: All builds passed 20180510 01:08:17-!- gfgtdf_ [~chatzilla@x4e363a75.dyn.telefonica.de] has joined #wesnoth-dev 20180510 01:10:33-!- gfgtdf [~chatzilla@x4e32b7af.dyn.telefonica.de] has quit [Ping timeout: 264 seconds] 20180510 01:10:42-!- gfgtdf_ is now known as gfgtdf 20180510 01:11:30-!- gfgt [~androirc@x4e32b7af.dyn.telefonica.de] has quit [Ping timeout: 256 seconds] 20180510 01:31:34<+discordbot4> Is there a general consensus on further integrating (or avoiding integration of) steam-related features to Wesnoth in future versions? (workshop, overlay, join MP game via friends list, etc) 20180510 01:31:54< celticminstrel> Overlay should just work. 20180510 01:32:05< celticminstrel> I don't know why it doesn't, but it probably will by the next major release. 20180510 01:32:39< celticminstrel> If by "join MP game via friends list" you mean "login to MP server with your Steam ID", I at least want to see that happen (and there's an open issue for it). 20180510 01:32:51< celticminstrel> Workshop support will almost certainly never happen. 20180510 01:32:57< celticminstrel> Achievements may happen at some point. 20180510 01:34:23-!- DeFender [~DeFender1@89-138-79-9.bb.netvision.net.il] has joined #wesnoth-dev 20180510 01:34:48<+discordbot4> šŸ‘ to all. By "join MP game via friends list", I mean that steam has an (invite to game) / (join game) system accessible from its friends menu. My understanding is that it doesn't work unless specifically implemented. 20180510 01:36:03< celticminstrel> Ah. No idea if that's gonna happen but I think the likelyhood is low. 20180510 01:36:07< celticminstrel> ^likelihood 20180510 01:36:12<+discordbot4> šŸ‘ thanks 20180510 01:36:46-!- DeFender1031 [~DeFender1@89-138-218-118.bb.netvision.net.il] has quit [Ping timeout: 265 seconds] 20180510 01:37:26< mattsc> gfgtdf, anybody: I am trying to properly port my AI development add-on to 1.14 and 1.15, so I am running it with --log-debug=deprecation 20180510 01:37:27< mattsc> I understand most of what I am seeing (incl. the fact the some of the messages come from mainline code, rather than my add-on), but there are two that I donā€™t get: 20180510 01:37:28< mattsc> > info deprecation: needs_select has been deprecated indefinitely. 20180510 01:37:30< mattsc> > warning unit: Unknown attribute 'previous_recruits' discarded. 20180510 01:37:31< mattsc> I grep-ed through the entire add-ons folder, and the mainline data folder, and those strings donā€™t seem to appear anywhere. 20180510 01:37:32< mattsc> In fact, I grep-ed through the full Wesnoth folder, and they only seem to appear in src/, nowhere else. 20180510 01:37:33< mattsc> Any insights? 20180510 01:54:58< gfgtdf> "that I donā€™t get" doe that men that you expected those messages but they dont appear, or that those mesage appear but you didnt understand them ? 20180510 01:55:24< mattsc> The latter. Sorry. 20180510 01:55:45< mattsc> I donā€™t understand what causes them. 20180510 01:56:11< gfgtdf> well "previous_recruits" were afaik soemthing used internally by the carryover code, was quite longh ago though 20180510 01:56:28< gfgtdf> "needs_select " is a key in [set_menu_item] (its deprecated) 20180510 01:58:04< mattsc> Right. It (needs_select) 20180510 01:58:20< mattsc> ā€¦ is nowhere in my add-on. Or in mainline. So where does the message come from. 20180510 01:58:46< mattsc> Same goes for previous_recruits. 20180510 01:58:56< gfgtdf> maybe try disabling all other addons 20180510 01:59:15< gfgtdf> actualyl i cannot find any deprecation message for previous_recruits in src actually 20180510 01:59:51< mattsc> Well, the rpevious_recruits message is not a deprecation message. 20180510 02:00:11< mattsc> I also checked through the entire add-ons folder, doesnā€™t appear in there either. 20180510 02:00:47< gfgtdf> well then that mesage is probably an engine bug, i assume it doesnt happen in the fist scenario but only in later scenario (when there is carryover) ? 20180510 02:01:10< gfgtdf> i assume that last mssage was about 'previous_recruits'? 20180510 02:03:20< mattsc> I checked for both of them, and canā€™t find them anywhere 20180510 02:05:05< gfgtdf> maybe you are loading an old savefile that has such menu items in them = 20180510 02:05:15< mattsc> Huh, however, when I go into the saved game, needs_select is in there. 20180510 02:05:24< mattsc> grep didnā€™t find that because itā€™s a .gz 20180510 02:05:39< mattsc> Nope, not an old save. 20180510 02:05:54< mattsc> produced 2 min ago 20180510 02:07:18< gfgtdf> aw wait i think i see whats going on 20180510 02:08:05< mattsc> And now previous_recruits is in the other save also. 20180510 02:08:05< gfgtdf> the depcrecation cehck just chetcks wheterh that key is present, and the savgame code writes needs_select=no (whihc is the default value), so the loadgame code will give you that message 20180510 02:08:11< gfgtdf> ye you can probabyl ignore it 20180510 02:08:35< mattsc> Okay, sounds good then. 20180510 02:08:43< mattsc> I was ignoring it anyway. 20180510 02:09:03< mattsc> Should I file an issue about this? I assume this should be fixed at some point. 20180510 02:09:10< gfgtdf> ye woudl be nice 20180510 02:09:20< mattsc> okay, will do 20180510 02:13:00< mattsc> BTW, there is mainline Lua code that produces deprecation warnings; e.g. by using wesnoth.get_variable or helper.distance_between. 20180510 02:13:08< mattsc> Should that be fixed in 1.14, or master only? 20180510 02:13:38-!- louis94 [~~louis94@91.176.171.238] has quit [Ping timeout: 246 seconds] 20180510 02:28:16< celticminstrel> Not sure, maybe start with master and then decide whether to backport it? 20180510 02:28:58< mattsc> Well, it wonā€™t break anything throughout 1.14, so not sure if thereā€™s really a point. 20180510 02:29:11< mattsc> On the other hand, I can test 1.14 right now, master not yet. 20180510 02:29:12< celticminstrel> A lot of deprecation warnings were squashed before 1.14.0, but I think that was without --log-debug=deprecation 20180510 02:29:33< mattsc> yeah. 20180510 02:29:58< celticminstrel> I didn't consider it a priority to remove things that wouldn't be seen without unusual settings. 20180510 02:30:22< mattsc> Makes sense. 20180510 02:31:04< mattsc> Iā€™m just happening to notice them right now, because Iā€™m trying to clean up that horrendous mess of AI code in my add-on. :P 20180510 02:31:19<+discordbot4> mainline should contain NO deprecated code 20180510 02:31:29<+discordbot4> such fixes go to 1.14 too 20180510 02:32:17< celticminstrel> Well I guess that answers your question. 20180510 02:32:35< celticminstrel> If they're definitely going into 1.14 then they can be fixed there first and forward-ported. 20180510 02:33:24<+discordbot4> Because a lot of people rely on mainline for reference 20180510 02:33:29< mattsc> Sure. I can take a shot at some of them at least. 20180510 02:33:31<+discordbot4> and if it has deprecated code people will use it 20180510 02:36:03< mattsc> Yep. I certainly did so back in the day. And still do, sometimes. 20180510 02:42:40< mattsc> celticminstrel: Is there a difference between helper.modify_unit and wesnoth.wml_actions.modify_unit? 20180510 02:42:53< mattsc> I guess the Lua syntax is slightly simpler for the former ... 20180510 02:43:07< celticminstrel> I'm not actually sure? 20180510 02:43:20< celticminstrel> But both are written in Lua so you could look at them to answer that question. 20180510 02:43:37< mattsc> Okay. 20180510 02:47:00< mattsc> Looks like the helper version can only handle direct keys. 20180510 02:49:38< mattsc> Heh: ```Note: This appears to be less powerful than the [modify_unit] tag and may be removed at some point in the future.``` 20180510 02:51:05-!- gfgtdf [~chatzilla@x4e363a75.dyn.telefonica.de] has quit [Quit: ChatZilla 0.9.93 [Firefox 52.7.4/20180427222832]] 20180510 03:36:40-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] 20180510 03:37:25-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev 20180510 03:38:27-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Client Quit] 20180510 03:39:11-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev 20180510 03:47:10-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: And lo! The computer falls into a deep sleep, to awake again some other day!] 20180510 03:47:56-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev 20180510 03:48:01-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Client Quit] 20180510 03:48:10< irker663> wesnoth/wesnoth:master Iris Morelle c517abb122 pofix: Update with 1.14.0 -> 1.14.1 rule AppVeyor: All builds passed 20180510 03:48:46-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev 20180510 03:48:59-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Client Quit] 20180510 03:49:43-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev 20180510 03:55:22-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has quit [Quit: celticminstrel] 20180510 04:58:21<+discordbot4> tfw you change your password and then immediately forget it šŸ˜ 20180510 04:59:02<+discordbot4> fml 20180510 05:00:16< irker663> wesnoth/wesnoth:master Iris Morelle fc002f49a9 pofix: Update with 1.14.0 -> 1.14.1 file AppVeyor: All builds passed 20180510 05:01:19<+discordbot4> Use a password manager. šŸ˜› 20180510 05:04:49<+discordbot4> resets again 20180510 05:33:06<+discordbot4> @youthfulIdealism celticminstrel: IIRC the steam overlay somehow hooks into directx/opengl automatically, so the transition to opengl in 1.15 should automatically cover the steam overlay. 20180510 05:37:48-!- Stragus [~alexis@modemcable198.134-161-184.mc.videotron.ca] has joined #wesnoth-dev 20180510 05:38:01< irker663> wesnoth: David white wesnoth:1.14 3f06813fcf59 / src/ (actions/attack.cpp random.hpp random_synced.cpp random_synced.hpp): added use_prng preference which adds a new experimental pseudo-RNG for casual ca https://github.com/wesnoth/wesnoth/commit/3f06813fcf597b63271e26c9a3a9aa1480020cca 20180510 05:38:03< irker663> wesnoth: David white wesnoth:1.14 df72d8ced06d / / (436 files in 25 dirs): Merge branch '1.14' of https://github.com/wesnoth/wesnoth into 1.14 https://github.com/wesnoth/wesnoth/commit/df72d8ced06de40badfd37d8a217c00065a3a1b8 20180510 05:38:32<+discordbot4> ... What 20180510 05:39:04<+discordbot4> @Denivarius Was this a pull request? 20180510 05:39:10<+discordbot4> Pushing to a wrong repository? 20180510 05:39:26<+discordbot4> no. I coded and pushed it. 20180510 05:39:39<+discordbot4> I know you haven't been active in almost 10 years, and it shows. 20180510 05:39:46<+discordbot4> it's an experimental mod that I'd love for people to try out. 20180510 05:39:50< Stragus> Hey guys. Just a little bug in "A new land" in "The rise of Wesnoth" campaign, it's very possible to "avoid deaths for as long as possible" indefinitely: http://www.rayforce.net/wesnoth000.png http://www.rayforce.net/wesnoth001.png 20180510 05:40:01<+discordbot4> ponders 20180510 05:40:14<+discordbot4> The naked std::cerr in the diff is probably the most obvious evidence of your unfamiliarity with the codebase. 20180510 05:40:16< Stragus> And so the player can end up with thosuands of gold for the next scenario 20180510 05:40:24<+discordbot4> well 20180510 05:40:31<+discordbot4> But that's not my point. Why didn't you file a pull request for this so we could have a proper discussion of it? 20180510 05:40:34<+discordbot4> there are a lot of places that use std::cerr 20180510 05:40:37<+discordbot4> to be fair 20180510 05:40:44<+discordbot4> I saw some other std::cerr. That isn't a done thing? In any case it only occurs if you have a preference set. 20180510 05:40:47<+discordbot4> Most of them are only in initialization code. 20180510 05:41:25<+discordbot4> In fact I don't see the point of forcefully "displaying" this (for the 1% of the audience who actually see stderr). 20180510 05:41:39< wedge009> Stragus: That's hilarious - maybe report that to the campaign maintainer or at least submit an issue report? 20180510 05:41:59<+discordbot4> Stragus: oh god what 20180510 05:42:06<+discordbot4> the point is because it's an experimental feature and if you set it you probably want to see it's active 20180510 05:42:20<+discordbot4> once the feature isn't experimental then the std::cerr can be removed 20180510 05:42:42<+discordbot4> Fair point 20180510 05:42:44<+discordbot4> We don't ever push experimental features directly into stable. 20180510 05:43:04<+discordbot4> Experimental bug fixes, yes (with a lot of care). 20180510 05:43:20<+discordbot4> We might consider this an exception. 20180510 05:43:24< Stragus> wedge009: Right, that would seem appropriate. Really, the scenario should just end after X turns when no unit has died 20180510 05:43:51<+discordbot4> Who is "we" in this case? I'm assuming you were discussing this in private all along, making this a repeat of the UtBS debacle? 20180510 05:43:56< wedge009> It's up to the campaign maintainers to perhaps set a limit on how many turns can pass. 20180510 05:44:06<+discordbot4> I knew he was working on it. 20180510 05:44:09<+discordbot4> well, sadly it seems like master is kinda broken right now. :-/ But again, it's protected by a preference. Though this is a good time to suggest something: I think we should consider moving to a more modern release system where we push out things regularly 20180510 05:44:15< Stragus> And I didn't even recall the best units to survive indefinitely, so it's not exactly hard to do 20180510 05:44:45< wedge009> master is very much broken. 20180510 05:44:53< wedge009> Intentionally, I believe. 20180510 05:44:57< wedge009> WIP. 20180510 05:45:35<+discordbot4> I did rather want you to push to a separate branch but having it on 1.14 isn't a huge deal breaker. This isn't strictly the type of thing we absolutely cannot allow in stable. 20180510 05:45:46<+discordbot4> And probably a lot of people would appreciate having it 20180510 05:45:52<+discordbot4> Now. 20180510 05:45:56<+discordbot4> Instead of a year from now 20180510 05:45:58<+discordbot4> or three 20180510 05:46:11<+discordbot4> Since it's gated it doesn't affect the existing code 20180510 05:46:16<+discordbot4> so it should be safe 20180510 05:46:59< wedge009> Regular releases implies everything is thoroughly tested... 20180510 05:47:13<+discordbot4> it does need a user-facing preference in the UI though 20180510 05:47:15<+discordbot4> I can add that 20180510 05:47:27<+discordbot4> wedge009: what about regular release for bug fixes 20180510 05:47:28<+discordbot4> šŸ˜› 20180510 05:47:50<+discordbot4> @Vultraz that would be awesome though I just added it like this for now so people could try it and give feedback. 20180510 05:48:02<+discordbot4> not sure what more rapid releases would mean for translations though, depending on the definition of "rapid" 20180510 05:48:57< Stragus> wedge009: The screenshots come from the Steam version, so I assume it's a stable release (I'm on Gentoo and didn't feel like compiling 20 dependencies for Wesnoth) 20180510 05:49:01<+discordbot4> Translation updates can be pushed to Steam as they come in, too 20180510 05:49:06<+discordbot4> so anyway this is the feature: If you set use_prng=yes in your prefs file, and you play a campaign with deterministic seed set, you will get access to this new prng. When a unit has 70% chance to hit it will still hit 70% of the time, but it will use a pseudo-random distribution which is much less streaky and designed to feel more 'fair'. It is recommended for casual players who find the randomness of Wesnoth stressful and 20180510 05:49:07<+discordbot4> unfun and want something that feels more intuitively pleasant to them. 20180510 05:49:19<+discordbot4> would love people to try it out and let me know how they like it. 20180510 05:49:40<+discordbot4> So it's piggybacking on the Deterministic mode option? 20180510 05:50:25<+discordbot4> Also, let me clarify this: I am actually one of the people who've been (silently) supporting the idea of using a PRNG for years. 20180510 05:50:42<+discordbot4> I believe requiring deterministic mode would be very bad for UX. I can imagine players going "I enabled the fair pseudo-RNG but it's still unfair!" 20180510 05:50:48<+discordbot4> The timing is just not ideal. 20180510 05:50:59<+discordbot4> Also, yes, master is broken thanks to vultraz's AR work 20180510 05:51:05-!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has quit [Ping timeout: 268 seconds] 20180510 05:51:10< wedge009> A lot of the negative Steam reviews I see complain about Wesnoth being about 'luck' instead of 'strategy'. I think they miss the point but oh well. 20180510 05:51:31< wedge009> Accelerated Rendering as opposed to Augmented Reality? 20180510 05:51:42<+discordbot4> I've actually been complaining about the implementation of the Deterministic mode option from a UX standpoint since before 1.13.0 was released. 20180510 05:51:43<+discordbot4> of course šŸ˜› 20180510 05:52:11< wedge009> Stragus: Yes, Steam releases are stable - same as the stand-alone platform releases, I should think but don't know for certain. 20180510 05:52:30< wedge009> I didn't have to compile 20 dependencies... 20180510 05:52:38<+discordbot4> Indeed, those reviews are hugely missing the point. You're not intended to completely avert all casualties. This isn't an RPG. 20180510 05:52:49< wedge009> Oh wait, Gentoo compiles everything. 20180510 05:52:56<+discordbot4> And incidentally, the idea of using a PRNG was being discussed around the time I released 1.13.0, years ago. 20180510 05:53:17<+discordbot4> I'd have greatly appreciated to see code back then instead of several years later in a stable release. 20180510 05:53:43<+discordbot4> actually 20180510 05:53:50<+discordbot4> The only dev team member who was in opposition to the idea was iceiceice as I recall. 20180510 05:54:00<+discordbot4> I think dave didn't mean it uses the "deterministic mode" option 20180510 05:54:17<+discordbot4> I think he means random vs random_synced 20180510 05:54:36<+discordbot4> That's not an option people get though? 20180510 05:54:47<+discordbot4> yeah deterministic mode is the way you get random instead of random_synced, right? 20180510 05:55:11<+discordbot4> The RNG is always synced in single-player for generating replays I thought (too bad gfgtdf is asleep right now, he knows this stuff). 20180510 05:55:20< Stragus> wedge009: Luck does play a part, and I understand that can be annoying to some (even if you understand the statistics involved). But it's not like a battle simulation should be deterministic, this isn't Chess... 20180510 05:55:39<+discordbot4> The only thing that Deterministic mode does is to ensure that the sequence doesn't change if you reload. 20180510 05:55:57<+discordbot4> I.e. so you can't save-scum your way out of a tricky situation. 20180510 05:56:08<+discordbot4> Yeah I think maybe we should call it "save random seed" instead of "deterministic mode" 20180510 05:56:34< wedge009> Deterministic random seed? 20180510 05:56:40<+discordbot4> lol 20180510 05:56:48<+discordbot4> "Deterministic" is too big a word for the average player as it is. :p 20180510 05:56:55< wedge009> Yeah, we had that discussion. 20180510 05:56:56<+discordbot4> I can see the questions from that naming now 20180510 05:57:15< wedge009> Fixed randomiser? 20180510 05:57:18<+discordbot4> At least for me, "deterministic" made me think that there isn't any randomness at all. 20180510 05:57:28< wedge009> Randomised fixer? (Okay that one was being silly.) 20180510 05:57:29<+discordbot4> "Save random seed" is more clear. 20180510 05:57:32<+discordbot4> That was what I thought when it first popped up as well. 20180510 05:57:47< wedge009> Only if you know what a seed is supposed to do in RNGs. 20180510 05:58:28< wedge009> Stragus: That's the point, the luck factor is inherent to Wesnoth's design. 20180510 05:58:38<+discordbot4> also, I assume it should be forward-ported, even if master is currently broken? 20180510 05:58:43< Stragus> All right, bug for surviving indefinitely in "A new land" posted 20180510 05:59:01<+discordbot4> yeah I agree with @jyrkive. It makes it sound like it's not random at all. 20180510 05:59:48< Stragus> wedge009: I actually find luck a little bothersome in a game where save/load is available at any time. It's too easy/tempting to cheat 20180510 06:00:37<+discordbot4> Yes, it should be forward-ported before it gets any harder to keep track of changes in 1.14 that don't land in master. 20180510 06:01:02<+discordbot4> (Like the patch I forward-ported last night because I casually stumbled upon it in a log.) 20180510 06:01:18<+discordbot4> also, IIRC merge commits are not preferred? not sure what saying Showing 436 changed files with 8,857 additions and 6,046 deletions. is useful for. 20180510 06:01:30<+discordbot4> Merge commits are acceptable. 20180510 06:01:47<+discordbot4> Project policy is to avoid merge commits, though. 20180510 06:02:00<+discordbot4> Wait, since when? 20180510 06:02:14<+discordbot4> For a couple of years, IIRC. 20180510 06:02:22<+discordbot4> I remember trying to enforce that until I gave up and started trying out --no-ff myself. 20180510 06:02:42< Stragus> Thinking of it, I think I would enjoy a harder setting level for a campaign where saving is disabled except for the first turn of any scenario. For the rare players who really want a challenge 20180510 06:02:58<+discordbot4> Mostly because it made it easier to keep track of related changesets. 20180510 06:03:41<+discordbot4> Then uh yeah Vultraz came along and for all I know maybe merge commits are supposed to include a message praising our reptilian overlords. 20180510 06:03:52< wedge009> ...that's what the 'deterministic mode' (mis-nomer or not) is supposed to stop. 20180510 06:04:26<+discordbot4> it's supposed to prevent save-scumming by redoing the exact same attack over and over again for better results 20180510 06:04:38< Stragus> wedge009: It doesn't prevent that. If you see the RNG doesn't give you the battle you want for unit X, then you attack with unit Y first, then X 20180510 06:05:11<+discordbot4> though it doesn't really completely do that, since changing the order of attacks can still give a different result 20180510 06:05:25< Stragus> Exactly 20180510 06:05:26< wedge009> You both said the same thing. XD 20180510 06:05:31< wedge009> Fair enough. 20180510 06:06:25< wedge009> I recall older versions (1.4?) tried to disable saves for this reason... I think it was re-enabled, though, for some scenarios like those in UtBS get really long. 20180510 06:06:47< Stragus> What about just an extra difficulty level, hard with no saves? 20180510 06:07:28<+discordbot4> I've seen some games that have an Iron Mode that's similar to that. 20180510 06:07:31<+discordbot4> It's mentioned in the manual. I noticed it when I was updating the manual. 20180510 06:07:33<+discordbot4> "Donā€™t abuse saved games. Long ago, Wesnoth only allowed saving the game at the end of a scenario. Mid-scenario saving was added as a convenience to use if you had to continue the game another day, or to protect against crashes. We do not recommend loading mid-scenario saved games over and over because your White Mage keeps getting killed. Learn to protect your White Mage instead, and balance risks! That is part of the 20180510 06:07:33<+discordbot4> strategy." 20180510 06:08:05<+discordbot4> yeah I think it'd be reasonable to have an iron man mode where saving the game deletes all previous saves you have and quits 20180510 06:08:10< Stragus> Iron Mode, Hardcode mode, whatever you call it. I think there are some people who would like that 20180510 06:08:24<+discordbot4> @jyrkive I'm afraid that goes into the same territory as me avoiding certain websites. 20180510 06:08:38< Stragus> It's very easy to prevent the game from deleting saves... 20180510 06:08:47<+discordbot4> Curiosity eventually overcomes me and I remove the foo.bar.baz 255.255.255.255 rule from my /etc/hosts. 20180510 06:09:10<+discordbot4> It does work for a while better than just making a mental note to avoid foo.bar.baz though. 20180510 06:09:44<+discordbot4> Hardcore mode sounds like a good name to me. 20180510 06:10:11<+discordbot4> I'd like "iron man mode" as the name. 20180510 06:12:01< wedge009> Iron Mauler Mode 20180510 06:12:20< wedge009> That's probably what I was recalling, jyrkive. 20180510 06:12:26< wedge009> The manual. 20180510 06:14:51-!- gallaecio [~quassel@143.red-81-32-24.dynamicip.rima-tde.net] has joined #wesnoth-dev 20180510 06:18:20<+discordbot4> @loonycyborg Can you remind me which are the libraries that are keeping us from providing 64-bit builds for Wesnoth on Windows? 20180510 06:18:37<+discordbot4> (IIRC it's the Gnome family, but I want to make sure.) 20180510 06:25:49<+discordbot4> @Vultraz Also, can you remind me what's the situation with the Discord Rich Presence branch? 20180510 06:26:31<+discordbot4> Very basic integration worked 20180510 06:26:58<+discordbot4> People on discord could see you were at the main menu, playing a game (sp/mp) or in the lobby 20180510 06:27:04<+discordbot4> and what game that was 20180510 06:27:41<+discordbot4> What more integration is needed? 20180510 06:27:48<+discordbot4> I wanted to add player numbers 20180510 06:27:57<+discordbot4> ie, 2/4 slots open 20180510 06:28:11<+discordbot4> and eventually I wanted to add the ability to join a game directly from discord 20180510 06:28:17<+discordbot4> you can do that, but it's a bit complicated 20180510 06:28:34<+discordbot4> I think it's better to use a staggered approach for this kind of thing. 20180510 06:28:51<+discordbot4> If it doesn't all have to go in at once, the rest of the features can wait. 20180510 06:28:57<+discordbot4> but jyrki said I couldn't merge it into 1.14 since the lib required the VS 2015 redist 20180510 06:29:03<+discordbot4> Integration would also require ability to build the game without Discord Rich Presence support. 20180510 06:29:05<+discordbot4> Oh for flip's sake. 20180510 06:29:15<+discordbot4> IIRC, the branch can't be built without it. 20180510 06:29:21<+discordbot4> and that 20180510 06:29:41<+discordbot4> Well, that just requires someone well-versed in CMake and SCons to go over it (cough Pentarctagon cough) 20180510 06:30:16<+discordbot4> No idea about the runtime library DLL issue. Are we allowed to ship those? 20180510 06:30:29<+discordbot4> I remember back when we were moving away from VC++ 6.0 we weren't. 20180510 06:30:51<+discordbot4> looks like there are more build options now 20180510 06:31:01<+discordbot4> (I just said "runtime library dynamic link library", for the record.) 20180510 06:31:02<+discordbot4> "(Windows) Enable to statically link the CRT, avoiding requiring users install the redistributable package. (The prebuilt binaries enable this option)" 20180510 06:31:26<+discordbot4> Is it allowed by the libraries' EULA? 20180510 06:31:53<+discordbot4> Because for instance IIRC you are not allowed to statically link GNU libc or libstdc++ into a non-free program. 20180510 06:32:08<+discordbot4> it's under the MIT license 20180510 06:32:17<+discordbot4> What is? 20180510 06:32:50<+discordbot4> the source code 20180510 06:33:01<+discordbot4> Of...? 20180510 06:33:11<+discordbot4> the RP library 20180510 06:33:21<+discordbot4> API. 20180510 06:33:30<+discordbot4> I'm not asking about that. 20180510 06:33:38<+discordbot4> I'm asking about the VC++ runtime it depends on. 20180510 06:33:58<+discordbot4> uhhhh 20180510 06:34:13<+discordbot4> I have no idea.... 20180510 06:34:16-!- aeth [~Michael@wesnoth/umc-dev/developer/aethaeryn] has joined #wesnoth-dev 20180510 06:34:53<+discordbot4> I'm confused 20180510 06:36:15<+discordbot4> Yeah, same, you said jyrki said... and that's not usually conducive to a fluid conversation. :p 20180510 06:36:36<+discordbot4> I just assumed you'd have look into why that is a problem. 20180510 06:36:48<+discordbot4> the redist problem was we're supposed to be able to build with 2013 20180510 06:36:49<+discordbot4> on 1.14 20180510 06:37:00<+discordbot4> which... 20180510 06:37:00<+discordbot4> Also, hang on. 20180510 06:37:08<+discordbot4> actually doesn't seem relevant? 20180510 06:37:13<+discordbot4> If the library is licensed under MIT then we could technically maintain or build our own. 20180510 06:37:23<+discordbot4> Requiring the VS2015 runtime is a problem because Rich Presence support isn't optional at the moment. 20180510 06:37:25<+discordbot4> One that doesn't require the MSVC++ runtime. 20180510 06:37:40<+discordbot4> Thus, if the branch was merged right now, it would break VS2013 builds. 20180510 06:38:01<+discordbot4> Integrating it in Wesnoth would be helpful for Linux since you'll be hard-pressed to find any Linux distributions shipping it. 20180510 06:38:21<+discordbot4> @jyrkive but there's a static build now 20180510 06:38:24<+discordbot4> @shadowm We could even simply recompile it so that it uses VS2013 runtime in VS2013 builds. 20180510 06:38:59<+discordbot4> The official builds do not use the VC++ runtime though. 20180510 06:39:20<+discordbot4> They are linked against the infamous msvcrt.dll and use GNU libstdc++ otherwise. 20180510 06:39:30<+discordbot4> I'm so damn confused here 20180510 06:39:35<+discordbot4> Ah, right. Official builds aren't made with Visual Studio. 20180510 06:39:52<+discordbot4> (I still think they should be) 20180510 06:40:27<+discordbot4> We'd still have the option to simply rebuild the Rich precense library, though. It would just use libstdc++ instead. 20180510 06:40:46<+discordbot4> .... 20180510 06:40:47<+discordbot4> I... 20180510 06:40:53<+discordbot4> I really have no idea what's going on anymore 20180510 06:41:36<+discordbot4> hopes jyrkive will explain this for Vultraz 20180510 06:41:45<+discordbot4> @Vultraz To clarify, which C++ standard library implementation is used is a property of the compiled binary, not the source code. 20180510 06:42:12<+discordbot4> It can be the source code's too if it makes use of non-standard functionality available in specific implementations though. 20180510 06:42:17<+discordbot4> ok, so if users don't need to install the redist to run the static version why is this a problem 20180510 06:42:40<+discordbot4> Like for example MSVC++ used to carry a hash_map (or so) class long before C++11 came around with unordered_mpa. 20180510 06:43:20<+discordbot4> With static linking, there is the licensing question: are we allowed to distribute binary code of Microsoft Visual C++'s standard library? 20180510 06:43:48<+discordbot4> I would assume so or you could never use it 20180510 06:43:55<+discordbot4> Even if it's statically linked, it just moves the binary code into our own .dll (or .exe), instead of having it in a separate .dll. 20180510 06:44:00<+discordbot4> @Vultraz You'd be surprised. 20180510 06:44:14<+discordbot4> At least back in the day, you couldn't. Apps were required to ship with the redist installer instead. 20180510 06:44:38<+discordbot4> So your installer would have to run MS' installer first. 20180510 06:45:03<+discordbot4> Also, the MSVC++ runtime is proprietary. 20180510 06:45:43<+discordbot4> At least with Linux (the kernel) this is actually a constant issue with proprietary drivers -- they are not on solid legal grounds because of the GNU GPL clauses about y linking non-free code. 20180510 06:46:11<+discordbot4> šŸ˜’ 20180510 06:46:18<+discordbot4> It's a corollary of the kernel's monolithic architecture. By loading a module you're not just running a program like when you launch a shell -- you're statically linking it into the kernel at runtime. 20180510 06:46:21-!- discordbot4 is now known as discordbot 20180510 06:46:38<+discordbot> So it effectively becomes part of the kernel, unlike most Windows drivers. 20180510 06:47:43<+discordbot> On top of that, Linux's kernel API is not stable. Even from the engineering perspective, it's expensive to maintain a proprietary driver because the underlying API keeps changing. 20180510 06:48:16<+discordbot> They do that on purpose to combat out-of-tree drivers, really. 20180510 06:48:18<+discordbot> It's not a problem for in-tree drivers (i.e. ones which have the source code in Linux kernel tree), because the maintainers keep them up-to-date as they change the API. 20180510 06:48:46<+discordbot> Indeed, it gives companies a strong incentive to make their drivers open source to get them into the tree. 20180510 06:49:37<+discordbot> Only the largest companies can afford to keep their drivers closed source. 20180510 06:50:16<+discordbot> You can see here what happens when things go wrong in this regard and the devs are forced to yield to external pressure: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux-stable.git/commit/?id=98f42e3f849412895171d096b901bfa3ff0fe004 20180510 06:50:33<+discordbot> reminds me of http://www.kroah.com/log/linux/stable_api_nonsense.html 20180510 06:50:46<+discordbot> "Simple, get your kernel driver into the main kernel tree (remember we are talking about GPL released drivers here, if your code doesn't fall under this category, good luck, you are on your own here, you leech .)" 20180510 06:55:59<+discordbot> what kind of elitist bullshit is that.... 20180510 06:56:26<+discordbot> The kind that isn't elitism. 20180510 06:57:17<+discordbot> It turned out to be extremely useful. It's thanks to this lisencing and policy that Linux got a huge collection of drivers for various hardware. 20180510 06:57:35<+discordbot> Otherwise it would be almost unused today. 20180510 06:57:50<+discordbot> Who wants to use a kernal that doesn't have drivers for anything? 20180510 06:58:07<+discordbot> Not to mention that most drivers just "work" out of the box if enabled during build and using udev. 20180510 06:58:36<+discordbot> I didn't need to download anything at all to get my new desktop working with Linux. 20180510 06:58:59< irker663> wesnoth/wesnoth:1.14 Charles Dang 805a08b67c Revert "Revert tutorial string changes n AppVeyor: All builds passed 20180510 07:03:38<+discordbot> Hmm. 20180510 07:03:39<+discordbot> https://cdn.discordapp.com/attachments/259976436490829825/444031745608646687/unknown.png 20180510 07:11:36<+discordbot> Odd... 20180510 07:13:13< matthiaskrgr> the terrain transition kind of soX 20180510 07:16:00< galegosimpatico> Hahah, well done. 20180510 07:30:32< Ivanovic> @shadowm I see that you updated pofix 20180510 07:30:44< Ivanovic> but did you also apply that on the files or does it only affect the announcement text? 20180510 07:31:11<+discordbot> There's a -w switch to enable using the website textdomain rules. 20180510 07:31:39<+discordbot> Otherwise they are ignored normally ā€” and conversely, when using -w the mainline textdomain rules are ignored. 20180510 07:31:53< Ivanovic> ahh, okay 20180510 07:32:56< Ivanovic> I was just confused because nothing was changed in the po files 20180510 07:32:58< Ivanovic> :) 20180510 07:33:13< Ivanovic> (at least not in the ones residing in 1.14 itself) 20180510 07:57:03-!- Bhoren [~Bhoren_wh@2a01:e0a:c:2150:c802:fbe3:3dc9:e5fe] has joined #wesnoth-dev 20180510 08:00:10-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has joined #wesnoth-dev 20180510 08:02:14< irker663> wesnoth: Jyrki Vesterinen wesnoth:1.14 9c841d134dd1 / src/actions/undo_recall_action.cpp: Fix #3065: unit halo remains after undoing a recall https://github.com/wesnoth/wesnoth/commit/9c841d134dd12c37474e79482cc48a6d3706aefb 20180510 08:02:16< irker663> wesnoth: Jyrki Vesterinen wesnoth:1.14 250a6e569b9f / changelog.md players_changelog.md: Changelog entry for commit 9c841d134dd12c37474e79482cc48a6d3706aefb https://github.com/wesnoth/wesnoth/commit/250a6e569b9fde2f51f94eae24e6a9e01ae40125 20180510 08:04:20-!- Necrosporus is now known as Guest84030 20180510 08:04:20-!- Guest84030 [~Necrospor@unaffiliated/necrosporus] has quit [Killed (moon.freenode.net (Nickname regained by services))] 20180510 08:04:21-!- Necrosporus [~Necrospor@unaffiliated/necrosporus] has joined #wesnoth-dev 20180510 08:08:54<+discordbot> so... 20180510 08:09:14<+discordbot> when exactly are we NOT in a synced context? 20180510 08:09:35<+discordbot> because as Dave's code stands now, it won't work in SP, since shadowm said it's synched. 20180510 08:10:05< irker663> wesnoth: Jyrki Vesterinen wesnoth:master 8fa3f6a6d4be / src/actions/undo_recall_action.cpp: Fix #3065: unit halo remains after undoing a recall https://github.com/wesnoth/wesnoth/commit/8fa3f6a6d4be0796965f343e44fda7f21c95a453 20180510 08:10:07< irker663> wesnoth: Jyrki Vesterinen wesnoth:master 9db47334a977 / changelog.md players_changelog.md: Changelog entry for commit 8fa3f6a6d4be0796965f343e44fda7f21c95a453 https://github.com/wesnoth/wesnoth/commit/9db47334a9774bcdcab46c3669de8ce3563f989b 20180510 08:10:19<+discordbot> šŸ¤” 20180510 08:12:28<+discordbot> I think his intention was not to have it work in MP 20180510 08:13:24<+discordbot> I'd take what I said with a grain of salt since I'm not familiarized with the RNG code to that degree. 20180510 08:14:09<+discordbot> You can easily test yourself and see if it works though. 20180510 08:14:38<+discordbot> I'm busy frying larger aquatic fauna right now. 20180510 08:15:44<+discordbot> well, I'll try to make sure it's working as expected by the time we're ready for 1.14.2 20180510 08:16:06<+discordbot> Yeah, you'd better decide on a schedule for that too asap. 20180510 08:16:18<+discordbot> And please for the love of god give people a 1-week-long string freeze this time. 20180510 08:20:04<+discordbot> it is so 20180510 08:20:27< zookeeper> Stragus, hmm, yeah being able to survive indefinitely is certainly a bug. how difficult was it to construct that situation in the first place, though? 20180510 08:21:08< zookeeper> i mean, can you easily replicate it by just camping the whole scenario in that specific (or similar) formation? 20180510 08:24:35< zookeeper> Stragus, does the triton in water (below the ship) have the trident, making him unattackable by the wounded archers? 20180510 08:36:20-!- madmax28 [madmax28ma@gateway/shell/matrix.org/x-mkvctodqnrnloidv] has quit [Ping timeout: 260 seconds] 20180510 08:36:23-!- Kawa[m] [kawamatrix@gateway/shell/matrix.org/x-nraijnklfmffkjqp] has quit [Ping timeout: 255 seconds] 20180510 08:36:31< zookeeper> mattsc, i didn't check old discussions (yet), but it looks like the retreat CA was put in place to avoid wounded units (who refuse to attack) blocking healthier units from taking a shot. of course in the screenshots the map is already full so they have no room to retreat, but i wonder if it was working correctly earlier. 20180510 08:36:35-!- ChipmunkV[m] [chipmunkvm@gateway/shell/matrix.org/x-hqbnxterzwlrbnjt] has quit [Ping timeout: 269 seconds] 20180510 08:46:14< zookeeper> the intent of course was that by combining 1) no-suicide attacking, 2) retreating wounded units and 3) an endless supply of enemies, it'd be unrealistic for the player to be able to avoid having their units gradually worn down enough for a kill. 20180510 09:20:17<+discordbot> @Vultraz So are we in string freeze now, or will it start later? 20180510 09:32:52-!- ChipmunkV[m] [chipmunkvm@gateway/shell/matrix.org/x-rsvcdojiudubhhvz] has joined #wesnoth-dev 20180510 09:48:41<+discordbot> @jyrkive later 20180510 09:50:31-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180510 09:59:01-!- madmax28 [madmax28ma@gateway/shell/matrix.org/x-zdnwaxhxfufyvzwj] has joined #wesnoth-dev 20180510 09:59:08-!- Kawa[m] [kawamatrix@gateway/shell/matrix.org/x-vlraqvbydfvlbpqv] has joined #wesnoth-dev 20180510 10:25:35<+discordbot> At start of string freeze I'll need to run pot-update right? 20180510 10:33:39<+discordbot> yes 20180510 10:33:42<+discordbot> but not now 20180510 10:33:47<+discordbot> because it's not a freeze 20180510 10:50:02<+discordbot> wow, Dave committed something. awesome. 20180510 10:51:22<+discordbot> I was thinking about a different tweak to the PRNG, one that explicitly favors the player against the AI. in other words, a cheating PRNG that caters to the expectation on luck whingers. 20180510 10:51:56<+discordbot> I believe Dave's tweak is much better than that. 20180510 10:53:02<+discordbot> it's similar. my idea was, if a human player has more than one strike in the attack sequence, and a 60% or greater CTH, then one strike is guaranteed to hit. conversely for the AI, 40% or lower, one is guaranteed to miss. 20180510 10:53:53<+discordbot> for attacks with a single strike, it would just do a flat 15-20% bonus/malus 20180510 10:54:01<+discordbot> Promoting as little as 84% to 100%? Sounds way too aggressive to me. 20180510 10:54:02< zookeeper> urgh. 20180510 10:56:44<+discordbot> Right, most players wouldn't enable it. It would be an optional mode or mod for luck whingers. Unfortunately I don't think it CBD with current WML features 20180510 10:57:30<+discordbot> what do you think of that, @Denivarius ? 20180510 11:10:18-!- irker663 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout] 20180510 11:13:14-!- vladimirslavik [vslavik@nat/redhat/x-fzvwmwqylnxkyjuj] has joined #wesnoth-dev 20180510 11:17:08-!- irker723 [~irker@uruz.ai0867.net] has joined #wesnoth-dev 20180510 11:17:08< irker723> wesnoth: loonycyborg wesnoth:1.14 3723c4edfabf / packaging/flatpak/org.wesnoth.Wesnoth.json: Drop flatpak manifest from stable branch since it's maintaned elsewhere https://github.com/wesnoth/wesnoth/commit/3723c4edfabf6dfd35c226c2ff1e59912dcaae47 20180510 11:41:37< irker723> wesnoth/wesnoth:1.14 David white df72d8ced0 Merge branch '1.14' of https://github.co AppVeyor: All builds passed 20180510 12:03:34-!- louis94 [~~louis94@91.176.171.238] has joined #wesnoth-dev 20180510 12:22:20-!- Appleman1234 [~quassel@124x38x163x22.ap124.ftth.ucom.ne.jp] has joined #wesnoth-dev 20180510 12:36:52-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180510 12:36:58-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180510 12:53:06-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc] 20180510 13:04:06-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev 20180510 13:06:00< mattsc> zookeeper: Yeah ā€¦ Iā€™l have a look. I guess there should be an exception to the no-suicide attack when thereā€™s nothing else that can be done, or something along the lines. 20180510 13:07:48< mattsc> Although, from description in the issue posted, it sounds like not being able to retreat was only part of the problem. 20180510 13:08:11< mattsc> Stragus: do you have a replay you could send me? 20180510 13:09:00< mattsc> zookeeper: Either way, ending the scenario after 30 or so turns as suggested by Stragus might not be a bad thing. 20180510 13:09:02-!- gallaecio_ [~quassel@187.red-83-46-126.dynamicip.rima-tde.net] has joined #wesnoth-dev 20180510 13:09:24-!- gallaecio [~quassel@143.red-81-32-24.dynamicip.rima-tde.net] has quit [Ping timeout: 260 seconds] 20180510 14:05:22-!- gfgtdf [~chatzilla@x4e363a75.dyn.telefonica.de] has joined #wesnoth-dev 20180510 14:15:57< irker723> wesnoth/wesnoth:1.14 Jyrki Vesterinen 250a6e569b Changelog entry for commit 9c841d134dd12 AppVeyor: All builds passed 20180510 14:20:52-!- Appleman1234 [~quassel@124x38x163x22.ap124.ftth.ucom.ne.jp] has quit [Remote host closed the connection] 20180510 14:27:09< gfgtdf> is there a lua/formula/wml function to convert a wesnoth color_range (like 'Red' ) to a rgb hex string ? 20180510 14:28:06< gfgtdf> for exampel if i want a string () to appear in a certains side team color ? 20180510 14:30:22< zookeeper> so how does one run wesnoth so that all translations are available? 20180510 14:31:55< zookeeper> ...available for choosing, that is. 20180510 14:33:00<+discordbot> IIRC, the check is hardcoded in at compile time. 20180510 14:33:34<+discordbot> In other words, you'd have to compile your own build to play with an incomplete translation. 20180510 14:33:56< zookeeper> how does Aurinkovuohi on #general post screenshots using the finnish translation, then? 20180510 14:34:10< zookeeper> maybe it's different for the steam version or something, i dunno. 20180510 14:37:08< gfgtdf> can i use ~RC with a color id liek "Red" aswell ? 20180510 14:37:16< gfgtdf> color id like* 20180510 14:40:43< zookeeper> that's the typical use, yes. 20180510 14:41:12< zookeeper> anything else is very rare in comparison. 20180510 14:44:21< gfgtdf> ah right 20180510 14:44:39< gfgtdf> i meant ~BG 20180510 14:49:34-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc] 20180510 15:14:31-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev 20180510 15:34:42< Stragus> zookeeper: It wasn't difficult, I didn't even recall the best units for the job 20180510 15:35:17< Stragus> More precisely, I should have recalled all ranged units. I had too many melee units, so even wounded ranged attackers would keep pummeling them, and I had to rotate them constantly 20180510 15:35:31< Stragus> With ranged units, I think I would have had nothing to do except press "end turn" 20180510 15:36:07< Stragus> And yes, the triton had the ranged trident, so they would be afraid to attack him 20180510 15:36:48< zookeeper> right, okay. do you still have a save/replay? 20180510 15:39:59< Stragus> Yup, give me a moment to post that (not at home, checked IRC remotely) 20180510 15:40:10< zookeeper> no hurry 20180510 15:44:15< Stragus> Here's the replay: http://www.rayforce.net/TRoW-A%20New%20Land%20replay.gz 20180510 15:52:05< mattsc> Stragus: Thanks. I downloaded the replay and will have a look. 20180510 16:01:40< zookeeper> well, there sure are fewer lvl2's than i would have thought 20180510 16:03:22< mattsc> zookeeper: if you have suggestions about how to change the AI (or whether at all), let me know 20180510 16:04:03< mattsc> I can probably have a first look tonight (which is still quite a bit off in my timezone). 20180510 16:09:25< zookeeper> retreating seems to work, so my first thought would be to tweak the recruitment limits or whatever so that there's a few more lvl2's (especially marksmen) in play already in the beginning, and perhaps to remove lvl1 recruitment altogether after some point (at turn 5-10 or so) 20180510 16:12:10< mattsc> Okay, sounds good. Iā€™ll let you take the lead on this then for now. 20180510 16:16:21< Stragus> Retreating wounded units was working until there were so many of them piled up that they couldn't move anywhere 20180510 16:18:30< mattsc> We could take zookeeperā€™s suggestion one step further and have the AI recruit gradually stronger and stronger units, going all the way to L3ā€™s only after a dozen or so turns. 20180510 16:19:26< zookeeper> could also edit a few hexes to make the northeast side of the castle more attackable 20180510 16:19:30< Stragus> Assuming the L3 units can reach the player, and they can rotate out when wounded, that should work 20180510 16:20:04< zookeeper> having more lvl2's would mean less lvl1's, so the surroundings would be less crowded anyway 20180510 16:20:07< Stragus> I don't feel editing the map hexes is the best solution... 20180510 16:20:54< Stragus> (if the castle is too hard to defend, then a good player would just run elsewhere, _anywhere_) 20180510 16:30:38< zookeeper> Stragus, are you willing to replay a slightly tweaked version of the scenario and try to do the same thing? 20180510 16:31:08< Stragus> Sure. But I'll recall even more appropriate units though ;) 20180510 16:33:20< zookeeper> okay. and i guess i should make it so that the AI _always_ prefers recruiting lvl2's, if it's able to... 20180510 16:36:53< mattsc> zookeeper: if needed, I am sure I could make the retreat CA move units out of the way of units in the way of units in ā€¦ 20180510 16:37:31< mattsc> Of course, in the situation as shown in those screenshots, thereā€™s almost nothing that can be done any more ā€¦ 20180510 16:37:55< zookeeper> well, yeah, you'd have to move a long chain of units to make room 20180510 16:37:57< mattsc> You could maybe limit the total number of units on the map to ā€¦ something large but not filling the map? 20180510 16:38:22< zookeeper> oh, wow, here's an idea: i could limit the number of lvl1 units, instead of lvl2 units... :p 20180510 16:38:55< mattsc> indeed 20180510 16:39:22< Stragus> What about capping the income of the AI so they would pay upkeep, and can only afford to maintain so many units? It's a bit pointless to add even more units after a point... 20180510 16:39:30< mattsc> Could we have units spontaneously combust if they are in the way but have no space to retreat to? 20180510 16:43:14<+discordbot> Simplest is probably best, zooks idea of capping l1 units feels like it would work to me. 20180510 16:44:33-!- gallaecio_ [~quassel@187.red-83-46-126.dynamicip.rima-tde.net] has quit [Remote host closed the connection] 20180510 16:45:15< mattsc> @UnwiseOwl I know, I wasnā€™t serious about my suggestion. It would give a new meaning to AI units suiciding though. :P 20180510 16:45:18<+discordbot> @sapient_n3t well, right now my change does that -- if you have more than two strikes and a 60% of greater cth you will hit at least once. 20180510 16:46:05<+discordbot> @sapient_n3t however, I would much rather make a change that I think is 'legitimate' than a "gosh you guys are just a bunch of whiners! Here's something that should shut you up!" 20180510 16:52:37<+discordbot> Whoa, I'd missed that. At the risk of sounding like a little whiner, personally I feel that a change like that is a poor choice. Is this as a mod or as an actual change to the rng for everyone. 20180510 16:53:18-!- iwaim [~iwaim@rasteenie.alib.jp] has quit [Ping timeout: 255 seconds] 20180510 16:53:18<+discordbot> It's disabled by default. 20180510 16:53:26<+discordbot> And also not available in MP. 20180510 16:53:32<+discordbot> Oh joy. Excellent. 20180510 16:57:21-!- iwaim [~iwaim@rasteenie.alib.jp] has joined #wesnoth-dev 20180510 17:05:31<+discordbot> Alright I have posted this to the Wesnoth forums explaining it: https://forums.wesnoth.org/viewtopic.php?f=6&t=48160 20180510 17:05:37<+discordbot> and yeah there is no way we will foist this on everyone 20180510 17:07:24<+discordbot> Interesting that it needs determinisic mode. 20180510 17:09:24-!- gallaecio [~quassel@188.79.96.255] has joined #wesnoth-dev 20180510 17:09:42< Stragus> With that proposed change, two attacks at 60% would means 1.2 hits, therefore 80% chance of one hit and 20% chance of two hits? 20180510 17:10:02< Stragus> And two attacks at 50% would mean exactly always 1 hit? Not sure I like that 20180510 17:10:54-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] 20180510 17:12:16-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180510 17:12:31<+discordbot> If I'm reading Dave right, it'd only apply to things with high cth and high number of strikes, so those cases wouldn't be affected. 20180510 17:15:43< irker723> wesnoth/wesnoth:master Jyrki Vesterinen 9db47334a9 Changelog entry for commit 8fa3f6a6d4be0 AppVeyor: All builds passed 20180510 17:26:00< gfgtdf> Stragus: yes that what it menas, for example orcish asssins will have 100% chance to poison. 20180510 17:29:33< Stragus> I think such a change very strongly affects the tactics and value of many units and circumstances. Even worse, it affects balance 20180510 17:30:09<+discordbot> When you put it that way (100% chance to poison) it sounds like a wholly different game. 20180510 17:34:37< gfgtdf> to be fair i thought assisins had a 2 attacks when i wrote that, which would be an noticable change 84%->100% but actualyl they have 3 strikes so it just a 94%->100% 'change' 20180510 17:35:19<+discordbot> The fact is that it's still a change from non-100% cth. 20180510 17:38:46< zookeeper> well of course it completely changes the game if you _know_ you will land at least one hit in many (most?) situations. 20180510 18:00:26<+discordbot> this change isn't really designed for the kind of people who are going to pore over the inner workings. It's designed for casual players who "don't want to feel the RNG is screwing them over". I'm not really interested in feedback from anybody who has looked into the code or how it works at this time. 20180510 18:04:18<+discordbot> I tend to consider myself a player when it comes to stuff like this. 20180510 18:05:02<+discordbot> Most of the game logic is a blackbox as far as I'm concerned, but I'll obviously feel some degree of attachment to the current gameplay because, let's face it, I've been playing this game for almost thirteen years. 20180510 18:05:18<+discordbot> That's an absurd amount of time. 20180510 18:09:21<+discordbot> But there's also another facet to this, which is that I'm a content creator who constantly struggles with trying to achieve reasonable balance with the classic RNG when developing my own campaigns. 20180510 18:09:41< Stragus> Personally, I don't think I have been annoyed because one unit didn't score any hit, the impact is low. But I certainly have been annoyed when 3 thunderers all missed, for example 20180510 18:11:11<+discordbot> I'd like to think that players should have a choice with regards to how to play my campaigns, but I can't help but feel that any kind of semi-officially-sanctioned changes devalue the amount of effort I put into testing and tweaking scenarios. 20180510 18:11:32< Stragus> Agreed 20180510 18:12:24<+discordbot> And yes, that also extends to the CTH mods (as in actual WML/Lua mods, not engine changes) that people have developed. I did end up having to deny people the ability to play my campaigns with one of them but that was really for technical reasons. 20180510 18:13:00<+discordbot> (Namely one of the events it injects interferes with several situations where I need to override CTHs to 0% or 100%.) 20180510 18:14:01-!- louis94 [~~louis94@91.176.171.238] has quit [Ping timeout: 256 seconds] 20180510 18:14:57<+discordbot> A casual observer wouldn't realize that the amount of time I spent testing and tweaking IftU's final boss scenario is on the order of days on and off. 20180510 18:15:27<+discordbot> And most of that was precisely to account for the random nature of battle outcomes. 20180510 18:17:00-!- gfgtdf_ [~chatzilla@x4e363a75.dyn.telefonica.de] has joined #wesnoth-dev 20180510 18:18:27-!- gfgtdf [~chatzilla@x4e363a75.dyn.telefonica.de] has quit [Ping timeout: 240 seconds] 20180510 18:18:39-!- gfgtdf_ is now known as gfgtdf 20180510 18:19:04<+discordbot> you're also perfectionistic šŸ˜› 20180510 18:19:38< Stragus> Creating a campaign certainly seems interesting. So far I'm a bit put off by the WML syntax, I guess I was expecting something closer to Starcraft's Galaxy script... which is very much C-like, the universal syntax standard 20180510 18:22:18-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] 20180510 18:22:50-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180510 18:39:52< gfgtdf> Stragus: you can also use lua in a lot of places. I usually prefer that. 20180510 18:47:59-!- timotei_ [~timotei@wesnoth/developer/timotei] has quit [Ping timeout: 246 seconds] 20180510 18:48:54< Stragus> gfgtdf: Ah yes, that looks a lot more usable than this [set_variable] stuff 20180510 18:55:15-!- timotei_ [~timotei@wesnoth/developer/timotei] has joined #wesnoth-dev 20180510 18:55:44< irker723> wesnoth/wesnoth:1.14 loonycyborg 3723c4edfa Drop flatpak manifest from stable branch AppVeyor: All builds passed 20180510 19:06:01-!- vladimirslavik [vslavik@nat/redhat/x-fzvwmwqylnxkyjuj] has quit [Quit: Leaving] 20180510 19:09:46<+discordbot> "I had this problem to, i solved mine by going into Onedrive/Documents/My Games and then making a new folder named "Wesnoth1.14". I don't know if this is for sure solution but it worked for me so i decided to put it here." 20180510 19:09:48<+discordbot> INTERESTING 20180510 19:09:51<+discordbot> šŸ¤” 20180510 19:10:39<+discordbot> Just as a reminder, I told people we were investigating this. 20180510 19:10:54<+discordbot> I have a feeling this is significant in some way but I'm not sure what... 20180510 19:10:55<+discordbot> I can't do any investigation of my own seeing as how I don't use either OneDrive or Windows. 20180510 19:11:42<+discordbot> I have tried to reproduce the problem, but haven't been able to. 20180510 19:12:46<+discordbot> the odd part is cores have nothing to do with the userdata path... 20180510 19:13:12-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: So long and thanks for all the fish.] 20180510 19:13:36<+discordbot> I don't even know if the core load error is the same bug tbh. 20180510 19:13:46<+discordbot> You can probably reproduce that part quite easily, and investigate it in the debugger. 20180510 19:14:13<+discordbot> Just create a regular file (say, Wesnoth1.14.txt) and tell Wesnoth to use that file as the userdata directory. 20180510 19:14:29<+discordbot> Or create a regular file called Wesnoth1.14. :p 20180510 19:14:54<+discordbot> Extensions are optional in NTFS and FAT with LFN. 20180510 19:15:46<+discordbot> I have literally NFI what I'm talking about here but... does this have any relevance? https://msdn.microsoft.com/en-us/library/windows/desktop/bb762181(v=vs.85).aspx A CSIDL value that identifies the folder whose path is to be retrieved. Only real folders are valid. If a virtual folder is specified, this function fails. You can force creation of a folder by combining the folder's CSIDL with CSIDL_FLAG_CREATE. 20180510 19:16:28<+discordbot> cpp HRESULT res = SHGetFolderPathW(nullptr, CSIDL_PERSONAL | CSIDL_FLAG_CREATE, nullptr, SHGFP_TYPE_CURRENT, docs_path); 20180510 19:17:16<+discordbot> Also, that succeeds. 20180510 19:17:36<+discordbot> At least so far I haven't seen any person post a log containing this line that would show up if the action doesn't succeed: 20180510 19:17:48<+discordbot> cpp ERR_FS << "Could not determine path to user's Documents folder! (" << std::hex << "0x" << res << std::dec << ") " << "User config/data directories may be unavailable for " << "this session. Please report this as a bug.\n"; 20180510 19:18:01<+discordbot> Hm 20180510 19:18:01<+discordbot> I see 20180510 19:18:54<+discordbot> (side note: should we deploy SHGetKnownFolderPath since it says SHGetFolderPath is deprecated?) 20180510 19:18:58<+discordbot> No. 20180510 19:19:03<+discordbot> Why? 20180510 19:19:05<+discordbot> That can be done in master but not 1.14. 20180510 19:19:16<+discordbot> Because 1.14 is still "compatible" with versions prior to Windows 7. 20180510 19:19:30<+discordbot> I'd rather not silently break compatibility for a cosmetic change. 20180510 19:21:14<+discordbot> Also, we already know that the issue is that the paths being created exist/come into existence as something other than folders. 20180510 19:22:05<+discordbot> Changing the API calls isn't going to automagically solve the issue where an external party is apparently intercepting our requests and turning them into "make something that isn't a folder at this path". 20180510 19:22:50<+discordbot> Especially considering that the API call that needs to be changed is actually this: 20180510 19:22:55<+discordbot> cpp bool created = bfs::create_directory(dirpath, ec); 20180510 19:23:22<+discordbot> @Denivarius I realize I may have sounded a little cynical and sarcastic with my idea, but it was a serious proposal. We already let players spawn Fire Dragons with debug node. And there are plenty who find enjoyment in playing with handicap a.k.a. cheats. 20180510 19:23:38<+discordbot> That probably resolves to CreateDirectoryW or something at Boost's end. 20180510 19:28:11<+discordbot> @Denivarius I feel that you may be giving them what they're seemingly asking for without giving them what they really and truly want, deep down in their hearts. šŸ˜‰ 20180510 19:29:08<+discordbot> I think what they want would be RNG that matches their intuitive expectations. 20180510 19:29:39<+discordbot> So that, for example, very high probabilities like 98 % would be rounded up to full certainty. 20180510 19:30:20<+discordbot> Rounding 84% to 100% would be way too much. In fact, it would probably bleed through to the extent that it would start to feel wrong in the opposite way. 20180510 19:30:21<+discordbot> did anyone report it going against their expectations, when the luck was in their favor? 20180510 19:31:08<+discordbot> Food for thought 20180510 19:31:24<+discordbot> I'm not sure what your point is... 20180510 19:32:27<+discordbot> sorry I was referring to my earlier proposal 20180510 19:32:34< Stragus> What about his? The game tracks the past "luck" of the player, and slightly adjust the next rolls towards the expected average 20180510 19:32:49< Stragus> So that several poor rolls in a row become unlikely 20180510 19:34:24<+discordbot> I guess a lot of people have different ideas for the optimum PRNG. what's needed is more WML support so that people can put their mods out there and let others test it. 20180510 19:34:29<+discordbot> The point is, humans are very bad at correctly perceiving risk (especially percentages). there are some social studies (which I cannot link right now, but I had that in class a few years ago) that said most people just categorize in around 5 categories (nearly impossible/unlikely/roughly even/likely/Nearly certain). Also, humans tend to overlook instances, whre things conform to expectations and remember exceptional cases. (SO, if 20180510 19:34:30<+discordbot> you have an attack with a high probability of hitting and it hits as expected on average, it doesn't get remembered after a few minutes. If you have an attack, that should hit with high probabilities and you loose 5 rolls in a row, it gets remembered. 20180510 19:35:21< Stragus> Indeed. The luck tracking I propose could work out 20180510 19:36:21< Ravana_> I have wanted to get access to luck statistics values from in scenario, but did not find any way 20180510 19:38:10<+discordbot> Stragus I'm a bit skeptical that your idea would satisfy the luck whingers... there is a similar one called Better Odds and people were still complaining when they had it turned on. https://forums.wesnoth.org/viewtopic.php?f=6&t=48149#p627689 20180510 19:39:13< Stragus> I guess some people want to play Wesnoth like a game of Chess 20180510 19:40:16<+discordbot> @sapient_n3t so I actually do think having a "Casual" mode which advantages the player in certain places could be interesting. One pretty interesting thing about it is that it could make it easier to develop campaigns. You don't have to make an 'easy' level for your campaign since players can just play on normal difficulty in casual mode and that is effectively easy, and realistically I think it might be more fun for most 20180510 19:40:17<+discordbot> players of easy levels. The main downside would be that it wouldn't well prepare them for normal mode or for multiplayer. 20180510 19:46:19<+discordbot> I daresay there are some of us who will never be well prepared for multiplayer 20180510 19:47:07-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev 20180510 19:49:38-!- louis94 [~~louis94@91.176.171.238] has joined #wesnoth-dev 20180510 19:55:09< zookeeper> Stragus, ok, i have a simple tweak that essentially makes the AI recruit half lvl1's and half lvl2's (whereas previously it mostly just recruited lvl1's)... i haven't balance-tested it myself, you want to have a go? 20180510 19:57:52< Stragus> Sure. Care to tell me what file(s) to replace exactly? 20180510 19:58:47< Stragus> I'm skeptical if that change will be enough though 20180510 19:59:19< Stragus> It also needs to stop piling up units, give them enough room to breath, move out and get healed 20180510 20:00:13< zookeeper> yeah, replace your data/campaigns/The_Rise_Of_Wesnoth/scenarios/15_A_New_Land.cfg with this: https://gist.github.com/ln-zookeeper/bee5b37da80869d9fd1351c352ab52c9 20180510 20:02:06< zookeeper> i'm hopeful that the decreased number of total units gives them enough space to retreat when needed 20180510 20:02:06< mattsc> zookeeper: I have three more reports highlighted that may or may not merit some changes by either you or me: 20180510 20:02:20< Stragus> Got it. And I assume I load some save just before that scenario 20180510 20:02:20< mattsc> https://github.com/wesnoth/wesnoth/issues/3073 20180510 20:02:45< mattsc> https://forums.wesnoth.org/viewtopic.php?f=43&t=20285&start=45#p627721 20180510 20:02:52< zookeeper> Stragus, you can just load the beginning-of-scenario save directly (that is, the one before a turn 1 autosave) 20180510 20:02:58< mattsc> https://forums.wesnoth.org/viewtopic.php?f=42&t=10281&start=30#p627230 20180510 20:04:15< Stragus> The "scenario start" save, all right 20180510 20:06:06< zookeeper> matssc, 1) i know next to nothing of NR so i didn't check that one out when i saw it earlier, 2) only one possible route per guard layout (of which there are 3) doesn't seem like a problem to me and 3) still way too easy, huh..? 20180510 20:10:36< mattsc> zookeeper: I really donā€™t like NR, but I could have a look at that scenario and see what happens. The {EXPERIMENTAL_AI} has a castle switching CA, so it might be as simple as switching that in. 20180510 20:11:15< mattsc> On 2), do I understand that right that we (I?) look into laying out those patrol routes? 20180510 20:11:16< zookeeper> well, i'd think an AI should never be trusted to know when to switch castles 20180510 20:11:59< mattsc> This oneā€™s actually reasonably sophisticated (I didnā€™t write it, Alarantalara did) 20180510 20:12:07< mattsc> which does not mean that it should be trusted 20180510 20:12:23< mattsc> But I could check out whether it does better in that scenario than the default AI. 20180510 20:13:01< mattsc> And on 3), Iā€™ll let you have that one. ;) 20180510 20:15:35< zookeeper> on 2), i think the person just isn't aware that there's 3 possible guard layouts (some/all with only one possible route), and thus thinks the map will always be the same. designing even more guard layouts isn't hard, just a bit tedious, so we could do that. 20180510 20:16:47< mattsc> zookeeper: Oh, okay. Well, let me know what you think is a reasonable approach. We donā€™t need to do everything that is suggested. 20180510 20:17:21< mattsc> Btw, did you see the comments on the AI ally instructions in Liberty? 20180510 20:18:58< zookeeper> for the record, the guard layouts used to be completely randomized, but i was never able to get the spacing to work perfectly, so sometimes one could get an "unsolvable" maze, and ensuring that it couldn't happen meant that the guards were always so far apart that it became a boring and too easy maze. 20180510 20:19:29< zookeeper> so now there's 3 hand-crafted guard layouts that can have tricky but always solvable sections 20180510 20:19:54< zookeeper> i did, and i'd kind of like to get rid of the whole AI controller thing. -.- 20180510 20:20:30< mattsc> Yeah ā€¦ 20180510 20:20:39< mattsc> You wouldnā€™t hear any objections from me. 20180510 20:24:22< zookeeper> there's some annoying bug causing the scrolling direction to get stuck from time to time (or, rather, very often), using 1.14.0. 20180510 20:26:04<+discordbot> Known. https://github.com/wesnoth/wesnoth/issues/2971 20180510 20:27:03< mattsc> Oh, huh, I bet I know what happened in that NR scenario. And it doesnā€™t look like that intended. 20180510 20:27:13< mattsc> *that's 20180510 20:29:02< mattsc> https://imgur.com/a/1h3N9ST 20180510 20:30:06< mattsc> Weā€™re talking about the Death Knight at 23,11. His starting keep is the one at 23,4, so he is supposed to have a big keep to recruit from. 20180510 20:30:43< mattsc> But he probably took one of the southern villages on the first turn, and then proceeded to the single-recruiting-hex castle in the south. 20180510 20:30:47< mattsc> Funny, really. 20180510 20:31:02< mattsc> I think that qualifies as a bug that should be fixed. 20180510 20:35:41< mattsc> zookeeper: ^ 20180510 20:36:06< zookeeper> sure 20180510 20:36:24< zookeeper> is there any need for that mini-castle to even exist within such a tempting range for the AI? 20180510 20:37:03< zookeeper> because that's kiiind of what i meant earlier; shouldn't trust the AI to pick the right castle, better to leave it with no choice. 20180510 20:37:20< mattsc> Indeed. 20180510 20:37:29< mattsc> If you mean the default AI, I fully agree. 20180510 20:38:14< mattsc> It might be to let the player recruit as he comes through there (in case the bigger one right south of it is taken)? Although I doubt that youā€™ll have much gold left at that point in that scenario. No idea. 20180510 20:39:17< mattsc> Actually, the save posted with that issue still has 183 gold and positive income at Turn 6. 20180510 20:39:23< mattsc> So Iā€™d rather not touch that. 20180510 20:40:11< mattsc> The simplest way of doing this would be to give the death knight side the two villages at 21,8 and 23,7 from the start. 20180510 20:40:33< mattsc> That way the side leader will grab one of those in the north, and that has negligible effect on the balance. 20180510 20:44:01< mattsc> Yep - thatā€™s exactly whatā€™s happening... 20180510 20:44:11< mattsc> I assume there are no objections if I simply go ahead and do that? 20180510 20:44:12< zookeeper> isn't there still risk of the leader attacking a lone scout or something, then deciding to take the mini castle? 20180510 20:44:43< mattsc> Yeah, I guess that is true. 20180510 20:44:43< zookeeper> well, i guess by that time it's not so relevant anymore 20180510 20:44:54< mattsc> That would be my guess. 20180510 20:45:12< zookeeper> but yes, go ahead 20180510 20:45:21< mattsc> But I donā€™t know for sure, and I am SAH not going to play that scenario to find out ā€¦ ;) 20180510 20:45:40< mattsc> And doing it the way with the villages is fine too? 20180510 20:45:52< mattsc> We could also simply hold the leader on the keep the first two or three turns. 20180510 20:46:07< mattsc> By removing his moves. 20180510 20:46:48< zookeeper> either works for me 20180510 20:47:46< mattsc> Iā€™ll do the latter then, as that does not change the balance at all (not that having one more village for one extra turn would matter) 20180510 20:49:43< zookeeper> or maybe he can simply be a guardian, i don't know 20180510 20:50:09< zookeeper> err, what happens if you mix a guardian and a leader? i don't even know. 20180510 20:52:16< mattsc> Not sure either. Would he return to the keep when no enemy is left in range? Would he stay out as long as enemies are in range? Good question... 20180510 21:00:38-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] 20180510 21:03:22-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180510 21:10:43-!- gallaecio [~quassel@188.79.96.255] has quit [Remote host closed the connection] 20180510 21:20:13< mattsc> I just had a great idea! 20180510 21:20:30< mattsc> I think we should make the Experimental AI the default in 1.15/1.16. 20180510 21:27:26-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] 20180510 21:28:26-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180510 21:29:17< mattsc> No takers? Thatā€™s disappointing. ;) 20180510 21:29:19< zookeeper> mattsc, is there a comprehensive list of specific differences up somewhere? 20180510 21:29:19< Stragus> zookeeper: With your changes, I'm very quickly running out of units to rotate, at turn 16 or so. Still... with more starting gold, for a few extra units, I think one could still hold the fort until units pile up and fail to heal 20180510 21:29:37< mattsc> Ooo, simul-posting. 20180510 21:29:43< Stragus> So it's probably a good idea to end the scenario at some point either way, maybe around turn 30? 20180510 21:30:21< mattsc> zookeeper: yes, there is: https://wiki.wesnoth.org/Experimental_AI 20180510 21:30:54< mattsc> If we did this, weā€™d only have to rebalance every single mainline and UMC scenario ā€¦ 20180510 21:31:11<+discordbot> Meh, most of them aren't that balanced anyway :p 20180510 21:31:39< mattsc> Oh, if you want to become more hated than the RNG, I dare you suggest this. ;) 20180510 21:31:41< zookeeper> Stragus, i suppose an invisible limit like that could be added after a point which should truly be impossible to reach 20180510 21:31:54<+discordbot> You started it! 20180510 21:31:55< zookeeper> Stragus, but apparently the tweaks at least make it considerably harder to do 20180510 21:31:56< Stragus> zookeeper: Agreed 20180510 21:32:02< Stragus> Yes, this is _hard_ 20180510 21:32:08< mattsc> Outside the dev channel, I mean. 20180510 21:32:27< mattsc> It might be an option for MP though. 20180510 21:32:41< mattsc> As in, making it the default. Itā€™s already available by default. 20180510 21:33:09< mattsc> Of course, I still hope that my other AI experiments will actually lead to some usable product in the end. 20180510 21:33:12<+discordbot> At this point it's better than the default in every situation anyway. 20180510 21:34:28< zookeeper> mattsc, i'd suggest we pick a medium-length campaign, do playthroughs with current AI, then playthroughs with the EAI. laborious, but i suppose kind of necessary to get an idea of how much it affects difficulty. 20180510 21:35:00< mattsc> zookeeper: yes, if we wanted to do that, thatā€™s what it would take. 20180510 21:36:09< mattsc> At this time, Iā€™d rather put my effort into working on my current project (Fred) to see if I can get an AI that is truly different, but can be used in a more general setting. 20180510 21:38:09-!- louis94 [~~louis94@91.176.171.238] has quit [Ping timeout: 264 seconds] 20180510 21:40:26< zookeeper> well, i'll try to commit the ANL tweaking tomorrow and come up with something to make the TSG scenario harder, but off to bed for now... -> 20180510 21:40:46< Stragus> Good night, and thanks for the campaign fix 20180510 21:40:59< mattsc> Iā€™ll commit the NR fix later tonight, probably. 20180510 21:41:01< mattsc> GN. 20180510 21:46:57-!- zookeeper [~lmsnie@wesnoth/developer/zookeeper] has quit [Ping timeout: 240 seconds] 20180510 21:52:58-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has quit [Remote host closed the connection] 20180510 21:53:04-!- janebot [~Gambot@unaffiliated/gambit/bot/gambot] has joined #wesnoth-dev 20180510 21:55:56-!- irker723 [~irker@uruz.ai0867.net] has quit [Quit: transmission timeout] 20180510 21:57:53-!- irker328 [~irker@uruz.ai0867.net] has joined #wesnoth-dev 20180510 21:57:53< irker328> wesnoth/wesnoth:1.14 mattsc 672440ce3d Prevent definition of wml.variables to c AppVeyor: All builds passed 20180510 22:01:14-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180510 22:01:18-!- stikonas [~gentoo@wesnoth/translator/stikonas] has quit [Read error: Connection reset by peer] 20180510 22:16:42-!- Bhoren [~Bhoren_wh@2a01:e0a:c:2150:c802:fbe3:3dc9:e5fe] has quit [Quit: Leaving] 20180510 22:21:02< mattsc> @Vultraz Whatā€™s a proper AI? 20180510 22:21:19<+discordbot> You said something about the experimental one 20180510 22:22:05< mattsc> That doesnā€™t guarantee that it will fix that problem (it might, might not, will require testing); and are you willing to rebalance NR if we do so? 20180510 22:23:03<+discordbot> NR is not balanced 20180510 22:23:46<+discordbot> I don't even think I really playtested it much after my chages. 20180510 22:24:13< mattsc> The point is, I am not going to throw a new AI into a scenario, esp. not in the stable series, without a lot of testing - and thereā€™s no way I am going to test NR. 20180510 22:24:20<+discordbot> It was already a mess, it's slightly less of a mess now, and you give its "balance" too much credit. 20180510 22:24:58< mattsc> Might be, but that does not change my previous statement. 20180510 22:25:10< mattsc> Esp. since it is not even clear that it will help the problem at hand. 20180510 22:26:25< mattsc> IF NR will go through a rebalancing cycle in 1.15, we could throw the ExpAI into all scenarios and see how it does. But again, I am not going to do that myself. 20180510 22:26:32< mattsc> The balancing, I mean. 20180510 22:26:43<+discordbot> No one wants to do that šŸ˜› 20180510 22:26:48<+discordbot> balance it. 20180510 22:26:59< mattsc> Yep. 20180510 22:27:17< mattsc> I rebalanced one mainline campaign, one that I actually likie. Thatā€™s enough for one lifetime. ;) 20180510 22:38:06-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc] 20180510 22:51:17-!- stikonas_ [~gentoo@wesnoth/translator/stikonas] has quit [Remote host closed the connection] 20180510 22:52:12-!- stikonas [~gentoo@wesnoth/translator/stikonas] has joined #wesnoth-dev 20180510 23:03:16-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev 20180510 23:06:39< mattsc> I guess I could get going with replacing 124 instances of wesnoth.set_variable(). 20180510 23:08:25< mattsc> I just hope that this will port forward without too many conflicts ā€¦ 20180510 23:10:11<+discordbot> I think this is right 20180510 23:10:14<+discordbot> cpp PWSTR docs_path = nullptr; HRESULT res = SHGetKnownFolderPath(FOLDERID_Documents, KF_FLAG_CREATE, nullptr, &docs_path); 20180510 23:16:53<+discordbot> What are you trying to do? 20180510 23:17:35<+discordbot> Use the non-deprecated function on master. 20180510 23:17:41<+discordbot> But not 1.14 20180510 23:17:46<+discordbot> You do realize that this is effectively a cosmetic change? 20180510 23:17:54<+discordbot> It doesn't fix anything at all. 20180510 23:18:16<+discordbot> Yes. 20180510 23:18:25<+discordbot> The function isn't going anywhere any time soon. 20180510 23:18:39<+discordbot> Millions of proprietary applications out there that can't be "updated" rely on it. 20180510 23:18:58<+discordbot> There are honestly better uses of your time. 20180510 23:19:49< EliDupree> When I run wesnoth from the command line on Linux, I get a bunch of commandline output (including my EoHS errors, which is useful to me). When I run it on Windows through cmd.exe, I don't see any output. Anyone here know how to make it display the output? 20180510 23:20:21<+discordbot> Use --wconsole on Windows or the cwesnoth.cmd wrapper that ships with it. 20180510 23:20:36<+discordbot> There's also a dedicated shortcut for the latter these days. 20180510 23:20:38< EliDupree> thanks! 20180510 23:22:13<+discordbot> @Vultraz If you're going to insist on this at least bump _WIN32_WINNT in SCons and others. 20180510 23:22:32<+discordbot> also, I'm going to demote the warning about duplicate cores to a log message 20180510 23:22:38<+discordbot> instead of an error message 20180510 23:29:37< EliDupree> Hmmm, that worked, I'm seeing wesnoth's own debug messages, but not my own messages that are created using wesnoth.log("err", ...) 20180510 23:31:04< EliDupree> (1.14.1) 20180510 23:36:20-!- mattsc [~mattsc@wesnoth/developer/mattsc] has quit [Quit: mattsc] 20180510 23:36:49<+discordbot> "Possible loggers are info, debug, warning, error, and wml" 20180510 23:37:08<+discordbot> Per https://wiki.wesnoth.org/LuaWML/Display#wesnoth.log 20180510 23:37:55< EliDupree> Did that change? I have known working code that says "err" for some other Wesnoth version, can't remember which 20180510 23:38:36<+discordbot> I don't think it's changed. However, the WML tag counterpart to this function has/used to have aliases for different log levels. 20180510 23:38:51<+discordbot> So you can/could indeed call "error" just "err" with it. 20180510 23:39:02< EliDupree> Changed to "error", still didn't work 20180510 23:39:23<+discordbot> What is the exact code? 20180510 23:42:09< EliDupree> wesnoth.log ("error", "whatever", false) 20180510 23:46:29<+discordbot> Works for me on Linux. 20180510 23:51:04< irker328> wesnoth/wesnoth:1.14 mattsc e2b9359502 Prevent definition of wml.variables to c AppVeyor: All builds passed 20180510 23:55:54-!- celticminstrel [~celmin@unaffiliated/celticminstrel] has joined #wesnoth-dev 20180510 23:57:49-!- mattsc [~mattsc@wesnoth/developer/mattsc] has joined #wesnoth-dev --- Log closed Fri May 11 00:00:42 2018