00:18 -!- Menche has joined #hardfought 02:06 -!- theRaisse has joined #hardfought 02:11 -!- raisse has joined #hardfought 02:11 -!- mode/#hardfought [+v raisse] by ChanServ 02:15 -!- theRaisse has quit [Ping timeout: 268 seconds] 02:16 -!- raisse has quit [Ping timeout: 268 seconds] 02:20 aosdict: The xNetHack changelog link in git repo README doesn't work. 02:24 -!- MiseryMyra has quit [Ping timeout: 245 seconds] 02:33 -!- MiseryMyra has joined #hardfought 02:49 -!- raisse has joined #hardfought 02:49 -!- mode/#hardfought [+v raisse] by ChanServ 02:57 -!- raisse has quit [Ping timeout: 268 seconds] 03:45 *wves* 03:48 -!- elenmirie has quit [Quit: Going offline, see ya! (www.adiirc.com)] 03:51 -!- bug_sniper has joined #hardfought 03:51 -!- raisse has joined #hardfought 03:51 -!- mode/#hardfought [+v raisse] by ChanServ 03:52 yani: add a keyboard command that switches between warning level and monster symbol 03:52 <@Khor> I'm missing a way to single this channel from the BILLION OTHER CHANNELS in this discord server 03:54 Khor: Use IRC not discord! *hides* 03:55 -!- MisteryMyra has joined #hardfought 03:56 -!- MiseryMyra has quit [Ping timeout: 276 seconds] 03:58 -!- MisteryMyra is now known as MiseryMyra 04:05 -!- raisse has quit [Ping timeout: 260 seconds] 04:10 -!- bug_sniper has quit [Ping timeout: 264 seconds] 04:18 @Khor mute all other channels and hide them 04:24 El-FIQo! 04:24 FIQ: W and P are synonymous in FIQhack? 04:24 (And R/T) 04:25 yes, NH4 04:26 ok, fantastic, I didn't know if they'd be any weird consequences of merging them (they are identical as far as I can tell) 04:26 <@Khor> Yeah but then I have to mute like 04:26 <@Khor> a billion channels 04:26 <@Khor> I thought of that idea but didnt sound too appealing 04:26 khoR: can you mute the server and unmute just the channel? 04:26 <@Khor> nope 04:27 <@Khor> tried that too 04:27 <@Khor> I muted all the other categories this channel is not in 04:27 <@Khor> but that still leaves like 04:27 <@Khor> 20-30 active channels 04:33 khoR: I think I've found something 04:33 let me test 04:34 Yeah, go to the server settings 04:35 khoR: https://imgur.com/a/aB6SV 04:37 * mcw momentarily looks like a regular discord wizard 04:39 FIQ: question: Is there any legal reason to have two pointers to the same item? 04:39 I handle equipped/owned items separately already. 04:41 mcw: can you elaborate? 04:42 having 2 pointers to something isn't bad by definition and you shouldn't try to prevent it from happening by accident as a side effect of what you're doing 04:43 FIQ: Right, but if it is true that only one thing can own an item at any given time, I can use a std::unique_ptr to enforce that. I'm not saying it's bad if I have to share the pointer, but I'm wondering when that would actually happen. 04:43 FIQ: e.g. monsters are referenced by maps, tiles, etc, and that's A-OK 04:43 I haven't done C++, std::unique_ptr isn't something I recognize 04:44 It means that only one thing is valid, so if you assign it to someone else, you lose your ownership of it. 04:44 Also, you should cache monster positions by in addition to having them in a linked list, also having them be reachable via map querying 04:44 FIQ: yes, they are reachable by map querying 04:45 So the reason for me to use std::unique_ptr is that it catches stupid mcw-errors, in that I can't give the item to something/one else and maintain a dangling reference that I shouldn't. 04:45 So I'd fail-fast if I try to access it after I've given it away. 04:46 Are you saying that if you have a pointer to an object, say, obj 04:46 which is on your person 04:46 and then the object is dropped onto the floor 04:46 which in NetHack code would be obj_extract_self(obj) and then dropy(obj) 04:46 obj would become NULL? 04:47 Yes. 04:47 I see 04:47 Yes, that seems useful 04:47 But what if you need to track it still? 04:49 Bah, yes, you're right, std::shared_ptr it is, but I can still use move semantics to get the obj automatically becoming null when you drop it. 04:50 So if you do dropy(std::move(obj_extract_self)); it still nulls the self copy of object 04:50 Just needs more discipline. 04:50 mhm 04:50 I really need to try out C++ some day... 04:51 Some of it is nice, some of it is goddamn awful. 04:51 My C++ "experience" goes as far as cout<<"Hello world!" several years ago 04:52 To be honest the main annoyances coming from C (I've only in the past year or so started to use C++ more) is working out exactly how references work, and trying to avoid templates. 04:53 You can do some neat stuff though - https://git.tubhack.com/mcw/tubhack/src/master/src/monsters.cpp Monsters are always returned as MonsterPtr (which is a std::shared_ptr), and they are allocated from a pool, and as soon as all references to the monster disappear, it automatically gets returned back to the pool. 04:55 what git thing is that 04:55 the site I mean 04:55 gogs, just a super no-config git hosting thing 04:55 ah 04:55 I will definitely move to github once I'm ready 04:56 was heading for the "Raw" button expecting github (didn't read the link carefully) and was a bit lost at first :P 04:56 But it's nice to be able to point trusted people at things in the meantime 04:56 ah 04:57 I feel there is pressure on me once it is fully public, I'm not sure why. 04:57 -!- ProzacElf has quit [Ping timeout: 256 seconds] 04:59 mcw: one thing Qud does a lot in its data is inheritance from a "higher" type so you don't have to specify all the data for similar kinds of monsters 04:59 I suppose this could be useful 04:59 FIQ: inheritance can be useful, but sometimes less readable, I'm honestly on the fence. 05:00 -!- ProzacElf has joined #hardfought 05:00 <@Khor> mcw: but that would also supress notifications from this channel 05:00 In nethack it's nice that if you want to see how something is handled, it (mostly) in one or two places, but if the behaviour is specified in the monster, or the weapon, or wherever, it's more of a hunt to find it. 05:00 khoR: You can add an exception, look at the second image. 05:01 as an example you could inherit from a base humanoid that specifies the generic properties of humans, so you don't have to do .speed = 12, .ac = 10, .resistances = 0, .resistancesConveyed = 0 05:01 on all the human-type monsters 05:01 <@Khor> oh, the thumbnail didnt show more 05:01 But maybe you find that better 05:01 <@Khor> thanks 👍 05:02 FIQ: ah I see what you mean, yeah, I'm probably going to make the MonsterSpec look more like the ObjectSpec (with a Builder), then you can do something like .InheritProperties(OtherMonsterSpec) 05:02 Rather than actual C++ inheritance 05:02 That makes a lot of sense, thanks *implements* 05:03 mcw: btw 05:03 the goblin struct is nice and readable 05:04 but how well does it scale 05:04 In terms of defining more, or something else? 05:04 defining more 05:04 I mean the raw size of the file 05:05 Ah yeah 05:05 if you did that style and added all nethack monsters 05:05 lemme whip something else up 05:05 the file would be 15k lines :P 05:05 FIQ: yes it would be unreadble 05:05 Well, just really big 05:05 Not saying the current way is any better... 05:05 FIQ: imagine slex specified like this *shudder* 05:05 FIQ: [Demo] had the idea to specify stuff like this in .json file(s) 05:05 emacs already has trouble reading vanilla monst sometimes lol 05:05 some lagspikes at times 05:06 Splitting over multiple files is bad I think 05:06 yes 05:06 It's so nice to be able to open one file and look around in it 05:06 I agree 05:06 gimme 10mins, I will show something 05:07 out of curiously, I looked at the other roguelike's internals I've looked at lately's size 05:07 [801][fiq@fiq-desktop ~/CoQ]$ cat ObjectBlueprints.xml | wc -l 05:07 11110 05:07 but that thing has even more definitions than NetHack 05:09 mcw: .ch = {A_NORMAL, {L'o', 0}, 0, 0}, 05:09 what are those out of curiousity? 05:09 first one is clearly attribute 05:09 second is letter 05:09 but the others? 05:10 It's a cchar_t from libuncursed, I think the last two are unused, but I just store the entire struct so I can pass it to mvaddwch 05:10 ah 05:11 Probably lazy, they should have an abstract glyph identifier that gets translated at draw-time 05:11 hmm your engine knows about cchar_t-s? shouldn't that be limited to the display 05:11 FIQ: 100% yes 05:11 heh, great minds think alike :P 05:12 :) Pass 2 will be separating display/engine, it's just much quicker for the initial hacky development to do it lazily at the moment 05:12 heh 05:12 * mcw may pay for this later... 05:12 haha 05:16 @4k?arch lich 05:16 arch lich ~1~ arch-lich (L) | Lvl: 25 | Diff: 0 | Spd: 9 | Res: fire cold sleep shock poison drain | Confers: fire cold | MR: 90 | Generates: gehennom | AC: -6 | Attacks: 5d6 touch cold, 0d0 cast spell | Alignment: -15 | Flags: genocidable, breathless, regenerates, poisonous, undead, covetous 05:18 mcw: hmm I was thinking @inherit stuff 05:18 let's say you would do this in regular C++, i.e. what you have now 05:18 you have a monster, b 05:18 which inherits from a 05:18 a is an unique, you don't want b to be 05:18 !tell LarienTelrunya https://www.reddit.com/r/roguelikedev/comments/8ckscl/feedback_on_randomly_generated_character_tiles/ 05:18 Will do, jonadab! 05:19 how would you even differ from "explicitly defined as non-unique" 05:19 and "implicitly unique from inheriting from a"? 05:19 i.e. is there a difference between .isUnique = false 05:19 and omitting .isUnique alltogether 05:20 -!- MisteryMyra has joined #hardfought 05:21 -!- deadnoob has quit [Read error: Connection reset by peer] 05:23 I think disallowing inheriting from unique is probably fair 05:23 but it makes things unclear I feel 05:24 clearly defining your own bool with true/false/undefined is very dubious here and a terrible idea 05:24 but there has to be some way you could do this 05:24 @4k?iron golem 05:24 iron golem (') | Lvl: 18 | Diff: 0 | Spd: 6 | Res: fire cold sleep shock poison | Confers: nothing | MR: 60 | Generates: gehennom dungeons | AC: 3 | Attacks: 4d10 weapon physical, 4d6 breath drain str | Alignment: 0 | Flags: breathless, mindless, thick hide, poisonous, vegan 05:25 -!- MiseryMyra has quit [Ping timeout: 276 seconds] 05:29 hmm so looking into it a bit more 05:29 seems like plenty of languages actually have that 05:29 nullable booleans/etc 05:29 so I guess if you know what you're doing and is consistent with it, it might not be a bad idea 05:29 oh well 05:30 (it should certainly not be a boolean at least, more like a custom enum with yes/no/inherit if you did it) 05:31 @4k?mumak 05:31 mumak (q) | Lvl: 5 | Diff: 0 | Spd: 9 | Res: none | Confers: nothing | MR: 0 | Generates: gehennom dungeons | AC: 0 | Attacks: 4d12 butt physical, 2d6 bite physical | Alignment: -2 | Flags: genocidable, herbivore, thick hide, infravisible 05:31 @4k?ancient black dragon 05:31 No such monster. 05:31 @4k?elder black dragon 05:31 No such monster. 05:31 update spoiler files 05:31 :xa 05:31 oops 05:32 -!- noty has quit [Quit: WeeChat 2.0.1] 05:32 :xan 05:35 mcw: Why is there trailing {}s in the orc struct? 05:35 You complained about this very thing in NetHack code a while back... 05:35 -!- ProzacElf has quit [Ping timeout: 265 seconds] 05:37 jonadab: I have a keymap configuration for ADOM with hjklyubn 05:37 Including shift+direction for farmove 05:37 And various other NetHack keymaps 05:38 Then the developer broke everything 05:38 and fixing it is apparently "too hard" 05:38 Only recent versions have this issue though 05:38 @mtf hi btw 05:42 nice weather 05:42 how are things going in ADOMland 05:44 -!- MisteryMyra has quit [Ping timeout: 276 seconds] 05:49 " [Demo]: how do you even do such a long, boring, repetitive dungeon without falling asleep in front of the monitor :D" 05:50 Ahh, so slex' yendorian tower 05:50 FIQ: https://git.tubhack.com/mcw/tubhack/src/master/src/monsters.cpp Does this look slightly better? 05:51 what is a hobgin 05:51 FIQ: where is a trailing {} by the way? I must hunt and destroy. 05:51 anyway looks ok 05:51 FIQ: :) 05:51 -!- MiseryMyra has joined #hardfought 05:52 mcw: below .form 05:52 mcw: anyway 05:52 for inheriting booleans 05:52 I think maybe it would be ok to do a trilean for it? 05:53 yes/no/inherit, inherit being defined as 0 (being the default) 05:53 FIQ: C++ is shit at defining structs compared to C, so the {} is required (sadly). 05:53 FIQ: When I inherit, I copy all of the data, I don't reference it, so you can inherit then overwrite the bool flags 05:53 and have functions like isUnique() which goes through the inheritance tree (you need to do this regardless for stuff), and if it ends up with "inherit", returns false, otherwise false/true appropriately 05:53 mcw: ah 05:54 Inherit("blah") literally just copies all of the "blah" spec data over the current stuff 05:54 wasn't sure how C++ magic worked 05:54 but yeah, if it works 05:54 then sure 05:54 <[Demo]> take a pOOP 05:54 Ah Inherit is not actual C++ inheritance. 05:54 mcw: yeah I figured 05:54 but wasn't sure how flexible you could be 05:54 as said 05:54 not worked in C++ 05:54 :P 05:54 https://git.tubhack.com/mcw/tubhack/src/master/include/monster_builder.h#L82 this line copies the other stuff 05:54 mcw: OBTW 05:55 there is a wonderful problem you will eventually have to figure out how to deal with 05:55 And yeah the worst part of C++ is that reading it when you are less familiar with it is *much* harder than reading C. 05:55 you see 05:55 in various places in the source 05:55 there is foo == PM_MEDUSA 05:55 (as an example) 05:55 your quest 05:55 is to figure out 05:55 how to generate PM_MEDUSA 05:55 and friends 05:56 here is how NetHack does it: 05:56 it has a file, called makedefs 05:56 I love strings so I'm likely to use 'if (foo->spec.name == "medusa")' 05:56 which will walk through the mons array 05:56 and write out a generated header file 05:56 with all the defines 05:57 mcw: ah, you see, there is one issue with that 05:57 which may or may not be a problem depending on your opinion 05:57 No compiler safety? 05:57 <[Demo]> efficiency? 05:57 exactly 05:57 @ mcw 05:57 what if you typed meduas 05:57 whoops 05:57 FIQ: Yeah I definitely see it as an issue. 05:57 honestly 05:57 it might be worth it 05:57 <[Demo]> ah yeah it is nice to know i fucked up sooner 05:57 the nathack way is dumb 05:58 for level files I want to make level file parsing a runtime thing 05:58 YESSSSSSSSSSSSS 05:58 this gives the exact same issue, but level files *already* need major testing 05:58 <[Demo]> make everything JSON 05:58 so I think the disadvantage isn't as big 05:59 and I could probably write a tool to parse it as part of my compile script if I wanted 05:59 FIQ: I think what I'll do if I go the string route, is something like 'grep spec.name ==' -> generate a test file that ensures that name exists. 05:59 And then run that as part of the build. 06:00 FIQ: ah yes, exactly the same as what you justsaid re: tool part of compile script 06:00 Typos like hoglin will also be caught there too 06:01 Oh yeah, FIQ [Demo] what do you think about tests like https://git.tubhack.com/mcw/tubhack/src/master/tests/path.cpp more effor than it's worth, or good to build up a massive test-suite? 06:09 Could be useful but assumes a consistentengine 06:11 Also you appear to have given orcs temporary cloning technology 06:11 :) that was to force the test to fail 06:12 So it generates https://pastebin.com/raw/qduYQpLW 06:12 I should rename the test "test_failure" actually 06:13 you shouldhave something like !assert 06:13 FIQ: ah yes, good point 06:14 ASSERT_MAP_NOT or whatever 06:14 mmh 06:14 mhm 06:14 but as said 06:14 assumes consistent engine 06:14 I think for a large number of actions it should be consistent 06:15 What it may need is to be able to do something like AssertMessageShown("You see a door open"), etc 06:16 And for things that are inconsistent, like what mon_spell a monster chooses, I can at least do a ASSERT_OR(all|the|possible|spells) 06:17 thing s 06:17 is 06:17 it tends to be the weird corner cases 06:17 that break 06:18 Aye, but does that reduce the usefulness of "trivial" tests to below the effort to write/maintain? 06:18 I agree they are not the be-all and end-all, or even close to 06:19 https://www.reddit.com/r/roguelikedev/comments/55mekm/unit_testing_roguelike_development_and_you/?st=jg23686k&sh=79d433cf 06:22 FIQ: one of the main reasons I made it in the first place was I can run with -i (for interactive mode), so I can set up situations that I know are failing and watch what happens 06:22 E.g. Force a goblin to spawn on the other side of a door 06:23 Which is likely only useful as I'm developing the feature 06:23 so a debug mode? 06:24 NH has a rather extensive wizardmode 06:24 FIQ: pretty much, just with the addition that you can specify the initial state 06:24 Instead of having to manually spawn the features each time 06:24 the usefulness of tests grows over time 06:25 bhaak: for regression testing? 06:25 yes. the more automated tests you have, the more likely they find something breaking instead of you or the players 06:25 mcw: NH4 has rewind as wizmode command 06:26 FIQ: ah yes, that would be awesome 06:26 there's a certain threshold after which you can't keep up with automated tests. 06:27 1 or 10 tests you can do manually. 100 you could do but only with much time investement. but not 1000 or more 06:27 bhaak: yeah, I am looking into fuzzing 06:28 mcw: I used afl with unnethack's dummy UI port. 06:29 nh4 has a fuzzer 06:29 random fuzzing is a bit problematic, as there are so many steps you have to go through to reach some code. guided fuzzing helps there, although I haven't checked how the coverage is 06:31 for the reproducible dungen layout, I generated a few millions dungeon layouts with a random seed to check for any crashes 06:31 it actually found a few, but that is usually hit and miss. 06:32 (I have to do that again some time to get good data on the OOD monster problem that is currently a problem in unnethack) 06:32 My only experience of testing professionally has been full integration tests, so we run the driver, generate an image/buffer, compare it to either a golden or a pre-known/calculated result. 06:32 OOD monster problem? 06:33 too powerful monsters appear too early in unnethack (Out Of Depth). most probably because of the vault code but I haven't verified it yet 06:34 Ahh, "fun" 06:35 integration tests are cool, too. but if your code isn't well designed to be unit tested, it is often also hard to be integration tested :) 06:35 Main problem with unit testing for us was you can't really mock up part of the GPU. 06:36 btw mcw 06:36 wht ae your vew on AI stength 06:36 Majority of the driver is setting up command/state buffers, which then get executed all at once, so the only way to verify them is to execute them. 06:36 FIQ: as in should they be super clever and annoying, or toned down so they are more "fun"? 06:36 f you let the I do all the thngs players can 06:36 without ealance 06:37 Ah, depends on the monster, I love the idea but don't want all monsters to do it 06:37 I did the later with the dummy UI port (a line orientend 'UI' for unnethack) but there were still problems with random bits 06:37 you get FIQHack 2016 06:37 monstes use castle wand 06:37 :D 06:37 CM spam everywhere 06:37 FIQ: I think I'm only going to let player character monsters do that. 06:37 etc 06:38 I really want each monster to have its own unique(ish) spell list. 06:38 spell/ability list, I should say 06:38 I see 06:38 To make monsters more varied 06:38 Which is an intentional direction, neither wrong nor right, just different. 06:38 It's why having all the variants is a beautiful thing, you can choose (on any given day) to play what you feel :) 06:39 heh 06:39 ah, yes, if the most important code runs on a specific hardware, you have a problem mocking that up. I had a similar problem once where a lib was just a thin wrapper around an Oracle DB and most of the logic was coming from an external client. you can do a few tests there, even mocking with a bit of ugly code, but it won't help that much as most bugs are beyond your reach 06:39 bhaak: aye, then you just have to live with integration tests and hope and pray. 06:39 Though last time I checked, the Vulkan driver I wrote was running about 300k regression tests per commit. 06:39 but any tests at all is better than none :) 06:39 Amen! 06:42 FIQ: but yes, coming across a room of player monsters is hopefully going to be a big shock. 06:42 "The Ronin drops a scroll named FOOBIE BLETCH. You run in terror." 06:42 "The Alchemist begins to pray. You start to attack but draw back" 06:43 Chaos! 06:43 FIQ: in fiqhack 2016 did your pets also act like madmen? 06:43 hah 06:43 player monster eing able to pray 06:43 is something I want to do 07:20 FIQ: If it prays to the same god as you worship, it resets _your_ prayer counter :) 07:20 <[Demo]> make monsters pray yeah 07:22 <@Khor> aww, mcw's thing doesn't work :feels: 07:36 [hdf-us] [slex] ttc1401 (Ele Ang Mal Law), 1072 points, T:500, quit 07:41 [hdf-us] [xnh] MiseryMyra (Val Hum Fem Law) killed Lord Surtur, on T:15385 07:41 [hdf-us] [xnh] MiseryMyra (Val Hum Fem Law) acquired the Bell of Opening, on T:15385 07:42 khoR: the discord? :( 08:17 [hdf-us] [nd] Eleven (Wiz Hum Fem Neu) rejected atheism with a prayer, on T:4036 08:17 [hdf-us] [nd] Eleven (Wiz Hum Fem Neu), 4852 points, T:4037, killed by a giant beetle, while praying 08:20 morning morning 08:28 mcw: what do you mean? 08:29 @ pets 08:36 I'm really hurting for a co-aligned altar 08:37 I'm about 20 levels in, and the only altar I've found is a cross-aligned one in Minetown 08:37 -!- bug_sniper has joined #hardfought 08:37 https://github.com/microsoft/windows/ :-) 08:38 oops, wrong channel 08:38 :P 08:39 I have so much stuff I wanna bless :c 08:40 <@Tone> MiseryMyra I have a game saved right now where I'm probably about to used confused blessed remove curse to make holy water 08:40 <@Tone> That's always an option 08:40 I might have to do that 08:45 -!- bug_sniper has quit [Ping timeout: 264 seconds] 08:47 @Tone Yesterday I posted my proposal for overhauling alignment record and several religion systems, which I think you had expressed some interest in? Would love some more feedback. 08:48 !players 08:48 K2: [hdf-us] ttc1401 [slex] 08:48 K2: [hdf-eu] No current players 08:48 aosdict whats the link again? 08:48 for your proposal 08:49 https://nethackwiki.com/wiki/User:Phol_ende_wodan/FavorProposal 08:50 <@Tone> aosdict Cool, I'll check it out 08:51 [hdf-us] [slex] ttc1401 (Lib Ang Mal Law), 3809 points, T:893, quit 08:52 FIQ: as in, if you had pets did they spam CM? 08:55 aosdict any plans to add favor for chaotic players doing chaotic/evil things? 08:56 K2: Ideally, chaotic players should be chaotic good... but what did you have in mind? 08:56 The main balance concern is that there should be no farmable or renewable source of favor gains. 09:00 mcw: ah, no 09:02 aosdict: chaotics in nethack arent viewed as chaotic good. they're viewed as evil (hence being able to sac your own race on an altar, etc) 09:02 so 09:02 K2: Not to the extent of, say, ADOM though. 09:02 doing things like helping a monster out of a trap should give you a penalty 09:02 but murdering the watch guard would give you favor 09:03 because evil 09:03 never played ADOm so I cant compare 09:04 You also don't want neutral players to end up in a dead zone of favor because all the bonuses and penalties people could think up only make sense if you apply them along good/evil lines. 09:04 yeah thats tricky 09:05 maybe as a neutral, you see a monster trapped in a trap. you dont help it get out but you dont kill it either. you stay indifferent. god grants you favor for being an asshole 09:05 dont know how you'd code that 09:06 overall your proposal is interesting 09:06 good reading 09:06 as far as vanilla alignment record is concerned, neutral seems to be "just like lawful, except sometimes the penalties are less harsh" 09:06 I think neutral is more like, your god wants you to do interesting shit, but for every chaotic deed you do you need to balance it out with a lawful. 09:07 And he gets upset if both you do nothing to change his favour *and* if you veer to far to either side. 09:07 Err, or. 09:07 mcw: That would require a different sort of scale though... 09:07 One that measures your moral standing from an objective point of view, not from your god's. 09:08 Ah yeah, not undoable but a bit more complicated 09:08 ADOM has this. But ADOM also has a strong sense of chaos = unapologetic evil. 09:08 Each action is objectively good/neutral/bad, and your god likes you doing good/neutral/bad things. 09:08 Chaos should be evil! 09:08 And your god should get mad if you don't do evil things. 09:08 And I do want to avoid dipping nethack's alignment axis even further down towards chaotic = evil. 09:09 But that's a whooooole different discussion 09:09 *nods* 09:09 Yeah without that it is hard to do something neutral-but-with-alignmenty-things 09:09 Because then it gets into weirdness along the lines of, Moloch's evil too? But you're evil? 09:09 aosdict: that's already kinda weird though 09:09 Yes. The goal is to make it less weird. 09:09 aosdict: you can summon Jubilex through race sac, but he is in Gehennom and aligned ish to moloch? 09:10 Which can be accomplished by bringing the chaotic end of the axis closer to chaotic good. 09:10 Aye 09:12 I mean, you could probably make it work, flavor wise. Ish. 09:12 You and your god are evil and just as bad as Moloch, the only question is who gets to keep the Amulet of Yendor. 09:13 And so while being effectively the same alignment as Moloch, you oppose him because he has the shiny. 09:17 Another problem with a scale based on objectivity is it makes where you start on the scale weird. Do neutrals get to start at ideal neutrality? Or should they be arbitrarily chosen to be too lawful or too chaotic at game start, and have to work in an arbitrary direction towards pure neutrality? 09:21 Perhaps the problem is, neutral alignment is defined more by not doing stuff than it is by specifically doing or abstaining from stuff. 09:21 Which makes things like balancing out each lawful act with a chaotic act feel clunky. 09:25 Say robbing a shop for example: lawful says "I won't do this, it's unlawful!", chaotic says "I'm gonna do this, it's self-serving!" but neutral will say "I'll do it only because the need for doing it exceeds my friendship with the shopkeeper". 09:25 so neutral behavior is based more on 'I have to do this to accomplish X' rather than " I want to do this because reasons/how I feel' 09:26 where X = survival 09:26 or accomplishing a certain necessary task 09:27 aosdict by the way i'm never playing xnh again so i can keep my 4 streak 09:27 :P 09:38 [hdf-us] [slex] ttc1401 (Wiz Ang Mal Law), 2306 points, T:533, killed by a monster (water elemental) 09:48 DynaHack updated (us and eu) 09:49 what updates were there? 09:51 [hdf-us] [xnh] MiseryMyra (Val Hum Mal Law), 128226 points, T:19368, killed by a priestess of Odin, while dressing up 09:52 FIQ provided a fix that helps prevent issues during ascension 09:52 from my commit - Fix: remove the ability to find statues/corpses of former players, as this feature causes desyncs whenever someone ascends (thanks FIQ). 09:53 since we pull directly from tungtn's repo, i havent been able to push it (no permissions) 09:53