00:21:33 Unstable branch on cbro.berotato.org updated to: 0.33-a0-636-g92776e1539 (34) 00:55:58 Monster database of master branch on crawl.develz.org updated to: 0.33-a0-636-g92776e1539 04:33:49 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5249-g4a8afe7061 05:16:36 Unstable branch on crawl.akrasiac.org updated to: 0.33-a0-636-g92776e1 (34) 09:37:59 <08nicolae> possibly stupid idea i gotta get out of my system. a few days ago r-i was talkin' about how some branches don't really have much in the way of interesting terrain. and i thought about Capital-V Vaults. vaults are associated with big doors... and there's already a door themed monster (vault wardens). what if there was some special kind of vault doors in Vaults. like they take a while to open or something so you're getting chased but your 09:38:00 only exit is a big metal door and it's not opening in time D: 09:41:49 <08wormsofcan> doors that automatically open and close... 09:50:31 <08nicolae> yes, exactly. now you're thinking with portals 10:56:38 <09hellmonk> Concerned that it would make pre-opening or pre-closing doors matter too much 10:58:26 <08nicolae> there is that, yes 10:59:36 <08nicolae> this is why i post my stupid ideas. someone points out an issue with them and then they stop being stuck in my brain. 11:03:49 <06wensley> I dunno, as someone who remembers old Vaults, I still think modern Vaults is super cool 11:11:50 New branch created: pull/4203 (1 commit) 13https://github.com/crawl/crawl/pull/4203 11:11:51 03Flugkiller02 07https://github.com/crawl/crawl/pull/4203 * 0.33-a0-637-gb6cdccc287: fix: Make spells that never miss unaffected by blindness 10(39 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/b6cdccc28794 11:11:55 Hi, everyone. I had some ideas about a randomized god, who reveals the abilities at each level of piety. The abilities would be grouped so that you always receive direct damage, buff, escape at different piety levels, but you can't be sure what you get beforehand. 11:12:45 And the abilities would build on each other. For example, at 1* you get a single-target damage ability, and at 4* you get an AOE, but applies the 1* ability on all targets. 11:12:57 I would like to hear comments about this. 11:18:02 (Some more context: I posted this on Discord; index thought it would be impossible to balance; Implojin suggested that I should really bring my ideas to this IRC channel as well) 11:22:33 I feel like that was tried long ago but I don't see a branch for it 11:27:43 Can I interprete that as "there might be interest, but lacks implementation"? 11:28:57 BTW, on the balance issue, I simply feel that I am not good enough a player to judge balance at all. 11:30:07 not sure it's actually interpretable, actually, since if it did happen it was so long ago that pretty much everything has changed since then 🙂 11:30:24 but it does at least suggest there's been some interest and some discussion 11:34:20 <08nicolae> yes, i worked with dpeg on a random gods proposal approximately nine thousand millennia ago, so the idea does exist, it would indeed be hard to balance but i suppose if that can be gotten past 11:35:04 <08nicolae> "guaranteed access to random means of damage, buff, escape, etc." seems kind of close to nemelex 11:35:36 Yes. 11:35:52 Also, randomized abilities are already a thing in Makhleb. 11:37:39 <08nicolae> my feeling is that it would be better to have a selection of random abilities but you can at least see them before signing up, although that might encourage scumming? 11:37:49 <12ge0ff> are there any notes about it? maybe in the dev wiki? 11:38:26 <08nicolae> on the ancient dev wiki, probably 11:39:02 I know I've seen it somewhere, so one of the dev wikis seems likely, yes 11:39:07 <12ge0ff> https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:god:propose:start too many proposals 11:39:33 <08nicolae> https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:god:concept:random_god 11:39:39 <12ge0ff> nice 11:40:58 <08nicolae> it is very old but nonetheless it exists 11:41:15 Theme-wise, I also feel I can reuse Pakellas (I assume an evocation-based Pakellas is no longer a consideration?). He's the god of devices and/or inventions, so very naturally he can "reveal" a machine piece by piece. 11:41:56 So perhaps using charges as power costs. 11:42:13 (all the more similar to Nemelex...) 11:43:14 evocation god's too powerful pretty much by definition, no? 11:43:26 Not something I want to revive either. 11:43:54 So there's the excuse to reuse. 11:46:15 Also, do you think I should write down more concrete examples? 11:46:24 Like a design doc. 11:47:41 SayItsName (L17 DrRe) ASSERT(false) in 'god-companions.cc' at line 866 failed. (Snake:3) 11:56:59 <08nicolae> it couldn't hurt 11:59:22 Where is the preferred place to write those? Tavern 2.0 or somewhere else? 12:00:51 <12ge0ff> yeah, the new Tavern is an appropriate place for this 12:04:10 SayItsName (L18 DrRe) ASSERT(false) in 'god-companions.cc' at line 866 failed. (Shoals:4) 12:05:56 <12ge0ff> ^--- that's the "should be impossible to reach here" assert in _get_saved_apostle() 12:18:29 SayItsName (L20 DrRe) ASSERT(false) in 'god-companions.cc' at line 866 failed. (Vaults:4) 12:41:15 <12ebering> fully automatic doors: you can't open them, just . and pray 12:47:30 <04dracoomega> Ugh... I know I've fixed various ways to reach that point in past, but I guess more remain. They're such a pain >.> 12:47:57 <04dracoomega> Hard to diagnose since the problem usually happens long before the actual assert 12:49:22 <04dracoomega> Guess I will have to take another stab in the near future anyway 13:01:16 <03implojin> ahyangyi: "I want to do a complicated random-ability god as one of my first crawl contributions, but I admit I'm not a great player and I can't judge balance" honestly sounds like a recipe for disaster, the contributor who created the wu jian council started from basically that position and became so frustrated that they left crawl development entirely 13:04:16 <12ge0ff> this sounds more as an encouragement than a warning, since we do have WJC in-game now 13:04:22 <03implojin> (if you want to do it anyway, we'll surely try to help, but something smaller would be an easier first contrib) 13:04:48 <03implojin> yeah, I'm not trying to scare anyone off from doing crawl dev! if you have an idea you're excited about, that's great; just trying to set expectations 13:05:07 <04dracoomega> Okay, now that I'm awake again, I've been typing up a response of my own ^^; 13:05:51 <04dracoomega> As other people have already said, Crawl's codebase doesn't make 'modular' abilities very easy for a number of technical reasons. But 'this is hard' does not in itself have to be a reason never to do something, if that something is sufficiently worthwhile. So, what do you think this god would bring to the table that distinguishes itself? My general opinion on things like 'modular randomized abilities' is that they're often just 13:05:51 less interesting than 'randomized choice of curated, non-modular abilities'. That by 'assembling' things out of simple pieces, while you technically end up with a far larger number of possibilities, that each possibility is that much less distinct from other ones. You can see this a bunch in some other games that generate 'random monsters' rather than 'choosing a random subset of handmade monsters', where everything starts to blend vaguely together and 13:05:52 feel indistinct from each other. This is not to say that this has to happen with any abilities that are modular (Makhleb's destruction flavors are arguably a 'modular' tweak to Annihilation/Atrocity/Carnage, but it's deliberately a limited one), but the more modular one makes it, the more I feel this almost inherently happens. There are so many gods in the game that there's already a large amount of 'variety' available per run in terms of god choice 13:05:52 (and how those choices combine with what the rest of your character is doing.) so merely offering 'variety' in a god is not itself a compelling argument. So what do you think this experience of gaining 'approximately anything, over time' will do that is interesting or rich in itself? Which could not be accomplished more easily (and perhaps with more interest) via a smaller amount of curated content? Certainly, I think there's room for gods that have 13:05:53 some randomized aspects of themselves or how you interact with them - that is a very broad space - but I'm not sure about that one, if that makes sense? 13:06:59 BTW, my "first contrib" was probably this: https://github.com/crawl/crawl/commit/576a475cf8a21c5a125139297ff8a2133ad89e43 . I guess (meh) graphics is pretty irrelevant to the discussion here though. 13:07:22 <12ge0ff> woah, 2010 13:09:01 <12ge0ff> which is 2 years earlier than DracoOmega's first patch 😛 (ducks and covers) 13:13:41 I agree. Joining a god expect "it could be anything" just isn't going to work. (and we already have Xom for that). But I think a more structuralized randomized god might have value. The "draconian/demonspawn-random", not the Xom-random. 13:14:27 As for modularity, I also agree that too much modularity simply doesn't work (or, just don't actually create interesting abilities / tactical decisions). 13:24:40 The only reason I brought up modularity is that: when there are multiple damaging abilities, I want them to share the damage type and quirks. That seems to be a way to reduce the perceived complexity, and not look like a haphazard ability collection. 13:27:12 <06wensley> Maybe this sort of thing would make more sense if formulated as a species rather than a god? Say, a species that gets a random selection of powers inspired by the existing slate of gods (obviously not as simple as just randomly parceling out existing god abilities, mind). Like demigods, but with actual flavor 13:27:41 we already have Ds 13:28:19 (there iss at least some attempt to make tiers play together) 13:30:59 <06wensley> I think there's room for more species in the vein of demonspawn that don't step on demonspawn's toes (obviously, otherwise draconians wouldn't exist) 13:31:51 <08wormsofcan> I think active abilities are best suited for a god slot 13:33:45 <04dracoomega> (I mean, I think one of the species with active abilities that we have is quite popular and effective at having abilities, these days. I think there's room for something more that does also. Or, at least, I don't think there's any reason to veto that design space ahead of time) 13:35:34 <06wensley> I'm imagining that such a species, like demigods, wouldn't be eligible for god worship. Make it easier to balance and is perfectly natural given the right flavor. What's neat is that this could let us play in the space of existing gods in a way that doesn't necessarily make sense given the existing kit of each god 13:36:30 ( We have so many species with active abilities, that I can't even guess which one you were talking about: Draconian, Formicid, Naga, Vampire, Barachi? ) 13:37:12 <04dracoomega> Okay, yes, this is a fair point. (But I was thinking of draconians at the time) 13:37:24 <04dracoomega> Who have a randomized and flashy button 13:37:54 <06regret-⸸nde※> They also rely on a small fixed pool of flashy buttons. 13:38:06 <08wormsofcan> doesn't this just create a stronger incentive to make it a god package? 13:38:31 I think index's point is modularity is bad. 13:38:42 <03implojin> (quick note for everyone: discord quotes don't display on the irc side of the bridge, i think) 13:38:45 Or, too many moving parts is bad. 13:42:59 <06wensley> Yes, I agree that trying to assemble a given ability via mad-libs tends to result in either blandness or nonsense (although sometimes nonsense is fun) 13:43:52 <06wensley> Although, arguably, this is how we generate randarts, so... 13:48:16 <04dracoomega> I mean, there are contexts in which that sort of thing is fine 13:48:27 <04dracoomega> Note that: most randarts are relatively junk and you don't have to interact with them 13:48:43 <04dracoomega> So you're basically getting many chances to try rolling something that actually 'works' 13:49:03 <04dracoomega> (Besides, the bar for distinctiveness amount individual slot-fillers is much, much less than your god powers) 13:49:53 <08wormsofcan> yeah equipment is intended to fit on a scale of utility while gods are expected to have parity with each other 13:51:04 <06regret-⸸nde※> (Comparative references for modular designs: * Coglin gizmos actually are very consistent in important ways: six properties of the nine or so properties always show up every time. They get away with it half because those majority properties are shared across the entire rest of the game and half because they are entirely passive. If the species package was to be compared to a god, the dual-wielding provided at 0* would be 13:51:04 substantially more emphatic power. * Rift Wizard 2 has boss-type mutators on top of monsters meant to replace the large number of previous uniques and per-monster subclasses Rift Wizard 1 had, and most feedback shows people missed the character and mechanics of those previous fixed arrangements. * Rift Wizard 2 also had randomized spell-stat-boosts staves attuned to given elements, in a game actually fully prepared for it, alongside fixed gimmick staves 13:51:05 in the same equipment slot. This was scrapped for sets of fixed staves, because figuring out what they'd actually boost across the hundreds of spells that game was quite opaque (and descriptions for randomized abilities will be quite long!), and because a variety of stats didn't actually provide much power for the schools they'd work on versus having more fixed staves (e.g. a summoning-centric school could get damage boosts or vice versus). * As I 13:51:05 understand it, Noita has extremely randomized wand properties of chaining effects and stats, and this works for it mostly because the game's systems and pace are built around expecting sudden accidental death, and not fully parsing a dangerous wand can be a fast way to explode enough things to die to such.) 13:53:14 <08wormsofcan> btw djinn fits pretty well in the space of "species with randomized abilities" 13:54:25 <06wensley> Don't forget demonspawn, which have a whole elaborate system of facets 14:47:30 <06wensley> If anything, crawl already has too many viable options, let's get some more unviable ones. I want a god of self-loathing, who is incensed that anyone would bother worshipping it, and it does nothing except debuff you and buff your enemies in magnitudes that increase with piety 14:52:05 03DracoOmega02 07* 0.33-a0-637-ge182471b36: Add a hopefully-clarifying statement to Searing Ray cast message (ge0FF) 10(2 minutes ago, 1 file, 4+ 0-) 13https://github.com/crawl/crawl/commit/e182471b3690 14:52:05 03DracoOmega02 07* 0.33-a0-638-g5766912e8f: Fix weird Mercury Arrow autotarget issues (Lici) 10(6 minutes ago, 1 file, 5+ 0-) 13https://github.com/crawl/crawl/commit/5766912e8fed 15:03:42 <12ebering> good April 1 idea 16:41:02 Unstable branch on underhound.eu updated to: 0.33-a0-638-g5766912e8f (34) 16:43:26 -!- Changesite1 is now known as Changesite 16:46:06 <04dracoomega> Geez, I've apparently been at this rewrite for over 3 weeks now (glancing over old commit dates) T.T 16:46:50 <04dracoomega> I mean, I guess I did ballpark it as "I could see this taking me a month" somewhat early on, but nonetheless it doesn't feel like it's been 3 weeks >.> 16:50:10 <04dracoomega> (Can I actually finish it in one more week? ....maybe? (I say with low confidence)) 17:43:16 <09gammafunk> my still unmerged qw branch, which I've finally resumed work on after a couple weeks break, still has its starting commit from late july 2024 17:47:41 <04dracoomega> I mean, I've been at this basically every day for the last 3 weeks (except for maybe 1 or 2?) but it's just... a lot 17:53:38 <09hellmonk> wow, I missed a big discussion 19:19:44 <04dracoomega> auto_switch is apparently currently quite broken for coglins. (It can cause you to spend 5 turns without a prompt, performing the switch mid-combat, but if you are already wielding a weapon in inventory slot (a) and (b), it will just spend a turn waiting in place with the message "You're already wielding that!") 19:20:52 <04dracoomega> I was going to say "Does anyone really need this anymore, now that ammo isn't a thing, and you can bump attack with ranged weapons, and breadwielding is going away" but I guess there was a bug report just a few years back involving enhancer staves inscribed with !a >.> 19:24:14 03Implojin02 07* 0.33-a0-639-g4e2cdc3fdc: Fix stash resist inscriptions for dragon scales (Lici) 10(18 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/4e2cdc3fdc88 19:31:48 <04dracoomega> @implojin Hmm... probably the problem here is that basenames aren't supposed to have plusses. I know I had to go around and patch up a few places where dbnames and such were also getting it (and they never should). To be honest, I'm not quite sure the intended difference between DESC_BASENAME, DESC_QUALNAME, and DESC_DBNAME 19:32:40 <04dracoomega> (Is there some significant reason to have all 3, for instance? I sure don't know.) 19:34:26 <03implojin> I'd assumed that was related to this bit and was somehow intentional: One minor loss is that item_def::name can no longer be passed ISFLAG_KNOWN_PLUSSES to hide printing of plusses on items, but on reviewing all current uses of this, approximately the only change I can see is that item plusses will be used for the message about manifold assault being incompatible with your current weapon, and the name of armour echoes (which is 19:34:26 arguably an improvement for them, now that they're not longer +0 all the time). DESC_DBNAME no longer includes plusses by default (it never made sense there). if that behavior was unintentionally broken, then sure 19:36:25 <03implojin> this isn't the first crawl ui fix related to insane string matching i've run into, so when i saw that i was like "yeah i'm done" 19:36:30 <03implojin> %git eb88551f38 19:36:31 <04Cerebot> Implojin * 0.31-a0-953-geb88551f38: fix: webtiles EV colour while Fragile (1 year, 2 months ago, 1 file, 1+ 1-) https://github.com/crawl/crawl/commit/eb88551f38de 19:37:26 <04dracoomega> Nah, that was a separate thing. (Or rather, there were places those flags were additionally used to hide specifically that for item names that normally used it, outside of 'lookup-style names', but some code that always hide it for particular types of names was accidentally and unrelatedly lost in the same refactoring and possibly only incompletely restored) 19:38:40 <04dracoomega> I mean, if this fixes the bug for now, that's great. But it makes me think there may be some other bugs possibly lurking related to this (or maybe not??). It would probably make sense to review where these different types of lookup-style names are used and see what they expect (and if some can be folded together) 19:39:04 <04dracoomega> Not saying you ought to do that now or anything, but I'll probably put it on my list >.>; 19:39:05 <03implojin> there probably are other dbname/qualname bugs lurking somewhere, in that case 19:39:34 <04dracoomega> I should have fixed it for dbname at least, but I didn't even think about basename at the time 19:39:54 <03implojin> // Partial item names. DESC_BASENAME, // Base name of item subtype DESC_QUALNAME, // Name without articles, quantities, // enchantments. DESC_DBNAME, // Name with which to look up item // description in the db. 19:40:06 <03implojin> presumably this is how they're intended to be used 19:40:26 <04dracoomega> I mean, I've seen how they're described there, but how is basename and qualname actually supposed to be different from each other? 19:40:40 <04dracoomega> Since it feels like each comment kind of implies the other, to me 19:41:40 <03implojin> honestly i can't think of any use for no articles, qualities, enchantments that wouldn't imply the same name used for desc lookups 19:41:47 <04dracoomega> Yeh 19:41:48 <03implojin> probably i'm missing something 19:42:05 <04dracoomega> And we have a bunch of code like if (descrip != DESC_DBNAME && descrip != DESC_BASENAME && !qualname) 19:42:36 I can come up with potential things but have no idea if they're applicable. like, a vault rename or a special unrand might change basename but you still need to look it up with dbname? which might also imply qualname since "the autumn katana" has a different article from its base type 19:43:26 <04dracoomega> Given that some parts of the code for naming these things has been changed recently (and potentially incorrectly), I think the only to be thorough is to actually do a case-by-case check of when each is used, and try to determine if there are any meaningful differences between them 19:43:29 <04dracoomega> Which sounds like a pain ^^; 19:50:16 <04dracoomega> (Like, I probably will at least take a look at some point, but let me tell you that after 3 weeks at this, I long to do something that isn't refactoring for a while afterward :P) 19:50:37 it's refactors all the way down >.> 19:51:12 <03implojin> 🐢 19:52:17 <04dracoomega> (I mean, I say '3 weeks at this', but the ID flag stuff was immediately before this and there was _find_square and then all the summon timer stuff before that and and... >.>) 19:54:42 <03implojin> actually that makefile bug was an overeager pattern too 20:20:22 03dolorous02 07* 0.33-a0-640-g717b6cd007: Fix typo. 10(50 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/717b6cd0070c 20:49:21 <06pleasingfungus> re the discussion from earlier today re procgen and randomness, think the 'procedural oatmeal' problem (https://galaxykate0.tumblr.com/post/139774965871/so-you-want-to-build-a-generator) is worth reading for folx who haven't > So your algorithm may generate 18,446,744,073,709,551,616 planets. They may each be subtly different, but as they player is exploring them rapidly, will they be perceived as different? I like to call this 20:49:22 problem the 10,000 Bowls of Oatmeal problem. I can easily generate 10,000 bowls of plain oatmeal, with each oat being in a different position and different orientation, and mathematically speaking they will all be completely unique. But the user will likely just see a lot of oatmeal. Perceptual uniqueness is the real metric, and it’s darn tough. 20:50:07 <06pleasingfungus> same re procgen gods. you may be able to generate a zillion different combination of god powers, but will they feel different to players? 21:06:08 The total number of theoretical "possibilities" doesn't matter (and not what I care about). What matters is when a player gives the same god 2 or 3 tries, what they see. 21:18:56 <06regret-⸸nde※> Several recent god reworks fulfill broad variety in very different ways; Dith supports each different spell schools in different ways (and shadowslip functions differently depending on the main means one deals damage), Makhleb has both the destruction flavours (1 of 4) and marks (choose 1 of 3 from 9), and Beogh of course has the randunique allies (which are admittedly not the most stable still- wizard apostles have been 21:18:56 struggling compared to the other two roles- but which have extremely large potential spell lists and unrand possibilities). 21:21:47 <06regret-⸸nde※> (Xom's... in a very weird superposition where given effects can be rare to ever see in a game, but the bazaar's not as random as I'd like. Upcoming talisman plans will help a little, and so might figuring out more randart properties.) 21:22:35 <06regret-⸸nde※> (I should unearth my "Xom animates non-standard features" stash to help with on the thing Xom's stronger at.) 21:41:30 <09hellmonk> Think I'm somewhat more positive on "random abilities god" than the rest of the devteam but I agree it's a significant challenge to make something that feels interesting and distinct from run to run while also being reasonably balanced and cohesive enough to have identity. Would want a more specific proposal I think. Fwiw I've mulled over combining some elements of pakellas and wudzu (experimental only) for a similar concept before, 21:41:30 but I haven't gone anywhere with it. 21:45:39 <04dracoomega> I mean, to be clear, I'm not saying that no variation on this theme could never work, but I do feel it's sort of creating an uphill challenge to distinguish 'random modular abilities' 'randomized selection of fixed abilities' 21:49:33 03hellmonk02 07[evokerstacking] * 0.33-a0-436-gec62fc33f8: reimplement 10(77 seconds ago, 6 files, 53+ 24-) 13https://github.com/crawl/crawl/commit/ec62fc33f8f1 21:50:05 <09hellmonk> still need to figure out savecompat for this but I think the new implementation works 21:52:58 <09hellmonk> yeah, my vague wudzupakellas idea was offering the player a choice of 3 components with two distinct effects each (that also locked an equipment slot) drawn from relatively short lists, like 5 or so things per effect set. This already sounded like an insane amount of work so I never did anything with it. 22:04:25 <06pleasingfungus> iirc advil was a long time big fan of this idea 22:05:30 <04dracoomega> I have been actually tempted to put some design work into that "What if monstrous demonspawn were its own species?" thing from the other day for 0.34, since when I thought about it, it actually might dovetail nicely with another plan I already had there. (I mean, there can be more than one thing playing with equip slots and related effects, of course) 22:07:22 <09hellmonk> there was some randgod plan going way back to the old times, maybe dpeg's idea? I think that one was just random abilities that changed every game and would have been visible in the god desc though. 22:07:44 <08wormsofcan> the original proposal had random conducts too I think 22:09:45 <04dracoomega> * @return The level of xp debt the given evoker type has * before it can be used again. I hope this isn't actually what it does ^^; 22:10:04 <04dracoomega> (I realize that's just a copy/paste oversight :P) 22:10:09 <09hellmonk> oh sorry, thought i fixed that 22:10:32 <09hellmonk> will update it tomorrow while I work on the savecompat side 22:11:10 <09hellmonk> (actually unsure if I need to do anything for that besides tag minor the old evoker tracking stuff) 22:33:24 https://tavern.dcss.io/t/proposal-the-new-pakellas-a-god-with-randomized-abilities/1079 22:33:37 I wrote down the concrete examples I have in mind 22:34:33 I am not very emotionally attached to the abilities, or their order, in the post, by the way :) 22:36:28 <04dracoomega> Presumably just tag the unmarshalling of you.generated_misc (to discard it on load, from old save) and remove the marshalling 23:35:11 Unstable branch on crawl.develz.org updated to: 0.33-a0-640-g717b6cd007 (34) 23:57:46 Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-640-g717b6cd007