00:35:47 Unstable branch on cbro.berotato.org updated to: 0.34-a0-1101-gb63a3c07f7 (34) 00:46:22 Monster database of master branch on crawl.develz.org updated to: 0.34-a0-1062-g464970fc77 01:34:01 03WizardIke02 07* 0.34-a0-1102-g6a4db99097: Skip setting useless skills in fsim (pooka109) 10(32 minutes ago, 1 file, 9+ 3-) 13https://github.com/crawl/crawl/commit/6a4db990970b 01:58:59 -!- indigaz24 is now known as indigaz2 02:54:25 03Implojin02 07* 0.34-a0-1103-g6529974544: Remove actor ground_level methods 10(4 minutes ago, 15 files, 17+ 33-) 13https://github.com/crawl/crawl/commit/652997454408 04:33:20 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 05:35:41 Unstable branch on crawl.akrasiac.org updated to: 0.34-a0-1103-g6529974 (34) 06:29:26 03Implojin02 07* 0.34-a0-1104-g1034eacce1: Remove a netted monster blink check 10(62 minutes ago, 1 file, 0+ 15-) 13https://github.com/crawl/crawl/commit/1034eacce1de 06:55:11 <06dolorous_84348> @gammafunk crawl.develz.org unstable builds are now properly keeping up with trunk, but its monster database builds are still stuck at commit 1062 for some reason? (Also, crawl-0.33 builds on crawl.dcss.io still appear to be on commit 8 of that branch instead of 9?) 08:49:17 <09gammafunk> %git 08:49:18 <04Cerebot> Implojin * 0.34-a0-1104-g1034eacce1: Remove a netted monster blink check (3 hours ago, 1 file, 0+ 15-) https://github.com/crawl/crawl/commit/1034eacce1de 08:49:32 <09gammafunk> Oh 08:53:55 <09gammafunk> %git stone_soup-0.33 08:53:55 <04Cerebot> dolorous * 0.33.1-9-gc92837dd87: Fix delay info leak on artefact heavy weapons. (4 days ago, 1 file, 4+ 0-) https://github.com/crawl/crawl/commit/c92837dd8700 08:57:17 <09gammafunk> hrm, I actually don't understand why cdi isn't building stable, will need to look more into that 08:59:53 <09gammafunk> random warning from cdo: > mapmark.cc:867:2: warning: extra ‘;’ [-Wpedantic] > }; 09:00:17 <09gammafunk> not related to monster builds but might be worth a look 09:00:17 Monster database of master branch on crawl.develz.org updated to: 0.34-a0-1062-g464970fc77 09:00:25 <09gammafunk> hmmm 09:00:36 <09gammafunk> *??-version 09:00:36 <04Cerebot> ansi Monster stats Crawl version: 0.34-a0-1062-g464970fc77 09:01:16 <09gammafunk> interesting 09:01:42 <08nicolae> hello. in addition to thinking about r-i's vault requests, i recently realized that i can probably put more options into the one-of-a-kind rings, (which will help dilute the pool of the ridiculous stuff). i was thinking school spell enhancers now that rings of ice and fire aren't real anymore. but maaaaybe harm and rampage also? i would like some input on what is probably okay as a rare ring drop and what i should avoid even then 09:08:08 03Implojin02 07* 0.34-a0-1105-g467a078616: Remove an extraneous semicolon (gammafunk) 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/467a078616aa 09:10:06 <03implojin> harm rings sound fun. rampage probably also fine. i'd been hoping not to dilute rampage too much since it's getting pretty common nowadays, but one-of-a-kind sounds fine, and we always have the option to let rampage stack to increase the value of seeing it more than once in a game. 09:10:53 <09hellmonk> give all the rings *contam then you can add whatever you want 09:11:11 <08nicolae> ^fragile 09:11:48 <08nicolae> i suppose i could hard-code the really good artprops to also include *contam or ^fragile or something 09:12:28 <08nicolae> what do we think about Regen or Archmagi generating 09:13:41 <09hellmonk> does regen attunement include artifact properties 09:13:49 <08nicolae> i have no idea tbh 09:14:56 <08nicolae> by default regen only shows up on stuff that's set to slow-swap 09:34:43 <08nicolae> my current thoughts are: Regen Harm Rampage Archmagi +Inv rMut Clar RMsl SH:5 HP: MP: Conj Hexes Summ Necro Tloc Fire Ice Air Earth Alch Forge with possibly some kind of secondary list or other weighting option so that the spell enhancers don't take up half of the weight 09:50:39 <08nicolae> anyway, not urgent, i can put that off for a while 09:50:46 <08nicolae> more importantly: where the fuck is my graph paper notebook 10:17:05 <09gammafunk> I stole it. I'll give it back when I'm done adding all the graffiti messages. 10:21:40 Monster database of master branch on crawl.develz.org updated to: 0.34-a0-1062-g464970fc77 10:22:59 <06regret-⸸nde※> (Regeneration attunement does seem to work properly for ring randarts. Enough of these properties are messy enough in terms of incentivizing swaps per-spell or for flattening rare given threats that I'd also be fine with the expansive list mostly if it's also deploying ^Fragile / ^Contam / Bane as a default, yes.) 10:24:35 <09gammafunk> %git 10:24:36 <04Cerebot> Implojin * 0.34-a0-1105-g467a078616: Remove an extraneous semicolon (gammafunk) (78 minutes ago, 1 file, 1+ 1-) https://github.com/crawl/crawl/commit/467a078616aa 10:25:26 <09gammafunk> @napkin0815 it looks like the git repo clone at git.develz.org/crawl/git is not updating. I can fetch from there at ~crawl/source/crawl-monster and it succeeds, but it's about 40 commits behind the repo on github 11:37:10 <12napkin0815> ok, checking 11:43:39 <12napkin0815> crawl@crash:~/source/crawl-monster> git remote update Fetching origin remote: Counting objects: 423, done. remote: Compressing objects: 100% (100/100), done. remote: Total 423 (delta 366), reused 379 (delta 323) Receiving objects: 100% (423/423), 54.87 KiB | 0 bytes/s, done. Resolving deltas: 100% (366/366), completed with 157 local objects. From git://git.develz.org/crawl 464970fc77..467a078616 master -> origin/master 11:43:40 * [new branch] lua5.4 -> origin/lua5.4 8f697b7bc7..c92837dd87 stone_soup-0.33 -> origin/stone_soup-0.33 Fetching submodule crawl-ref/source/contrib/lua From https://github.com/crawl/crawl-lua * [new branch] lua5.4 -> origin/lua5.4 f2279c1..1706b0f upstream -> origin/upstream 11:43:47 <12napkin0815> some git receive-pack got stuck 12:03:34 <08nicolae> thank you for returning it, you spelled my name wrong though 12:04:38 <08nicolae> yeah, that's a fair point, i might stick with Fragile and Bane just to continue the theme of stuff that can't normally generate on a ring. which props in particular do you think need nerfing? regen, probably rmut even though it's already been around for a minute, what else? 12:05:46 <08nicolae> well, muts have gotten a lil nerfed lately so maybe rMut is less urgent to dial back on 12:16:15 <08nicolae> anyway, it's not like the shop isn't fine the way it is, i just like endlessly tweaking things, because of how i am 12:16:37 <08nicolae> now for the real question, how many shambling mangroves should be in the alderking's retinue 12:33:26 <06regret-⸸nde※> RMsl and Rampage both are big effects that would be regular swaps otherwise. 12:40:52 <06regret-⸸nde※> Shambling mangroves are somewhat redundant with the entangling root spells of Alderkings themselves (they're explicitly themed as fae, though Depths currently only has spriggans to represent such), though large collections of monsters can function on theme regardless of redundancies, weighed against how much reward and warning a given vault presents as usual. 12:49:47 <06regret-⸸nde※> @dolorous_84348 re: c2427cd, I removed the : set_feature_name("crystal_wall", "wall of pink crystal") line mostly because we're not specifying the default colour of crystal walls by default, and therefore it's conspicuously inconsistent to highlight the most common crystal walls (in two overflow altars) over the nearly two dozen other cases. If this is meant to cover accessibility's sake, I'd much rather hardcode it on the C 12:49:47 side of things to specify in descriptions crystal colours based on their tile / console colours rather than have to make sure every single vault with differently coloured crystal then includes lines describing their wall colours. 12:49:54 <09gammafunk> Thanks! 13:04:28 <08nicolae> yeah, i'm thinking an advanced version of the ancient Phyte Club concept, just a bunch of murderous plants. possibly an eleionoma. dryad would be nice but it's weak and i think supposed to be player only. there's spriggan druids but that feels a little like stepping on the enchantress's stuff in arcadia.des 13:06:17 <06regret-⸸nde※> The only spell-only monster allowed to generate in vaults are the eldritch tentacles in the one wizlab. (I should write more for docs/develop/levels/reserved_monsters.txt....) 13:06:31 <08nicolae> that's what i figured, yeah 13:07:27 <02Darby> oh huh, the one vault dryad I thought I vaguely remembered is gone 13:07:41 <08nicolae> "now, what if you made a special vault-defined alderking that had the normal spells but also had awaken forest and awaken vines and put it in a vault full of trees" i could do that but gammafunk would shoot me 13:08:02 <08nicolae> i wonder if it was one of mine, that sounds like something i would do 13:16:36 <06regret-⸸nde※> It was mainiacjoe_swamp_entry_angry_trees, which... I plausibly need to nerf another time? 13:31:44 <06regret-⸸nde※> ...In checking over Lair vault kill numbers, I, uh, am somewhat confused as to why querying Cerebot privately with !lg * current br=Lair s=kmap seems to be providing unorganized output? 13:34:53 <06regret-⸸nde※> (Possibly something to drag over to #crawl-sequell , but, it clearly jumps back up from a descending order at 137x gammafunk_worm_tunnel, and then again at 188x wormcave.) 13:34:54 <06regret-⸸nde※> https://cdn.discordapp.com/attachments/747522859361894521/1436091492547559455/image.png?ex=690e57ed&is=690d066d&hm=7ce84b7225c9df9df235bfc68b2e368b5f9668a40e59fc4853efa9e42043b83d& 13:36:40 <08o____0> It looks like sequell dms the messages in the correct order in IRC at least 13:36:46 <09gammafunk> I rewrote the irc code to use the asyncio-enabled latest version of the irc module so multipart queries are bugged that way 13:37:01 <09gammafunk> I'll be getting to fix that soonish 13:37:14 <06regret-⸸nde※> I can wait. 13:37:20 <09gammafunk> just something odd happening in the message processing that's introducing some arbitrary randomness 13:50:20 <06dolorous_84348> Okay. I'll remove it. Sorry for the confusion. 13:51:09 <06regret-⸸nde※> ...I would have been fine with that latter approach (of dynamically labelling the features as such), for the record. 13:51:31 03dolorous02 07* 0.34-a0-1106-gfb9580d432: Revert "Properly use pink crystal wall desc." 10(2 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/fb9580d4320e 13:52:44 <06dolorous_84348> So would I, but hardcoding it in C as you suggested sounds like the better approach to me. 13:54:07 <02Darby> honestly hypothetically doing it for crystal makes me wonder if it would be good to commonly do for other often-recolored features like stone and metal, though I know not if it would interact well with all the vaults that do that 13:57:30 <06dolorous_84348> The wall of orange crystal in the chaos spawn temple also still has a separate description. (What I was going for in that case was the in-game association of orange crystal with "stuff that messes with the player's head.") I suppose that also needs removal/adjustment? 15:34:34 03dolorous02 07* 0.34-a0-1107-g41b0740ec8: Remove the other Nemelex pink crystal wall desc. 10(10 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/41b0740ec835 15:34:34 03dolorous02 07* 0.34-a0-1108-gcd5996bbb5: Remove now-unused pink crystal wall description. 10(8 minutes ago, 1 file, 0+ 4-) 13https://github.com/crawl/crawl/commit/cd5996bbb55d 15:46:24 03dolorous02 07* 0.34-a0-1109-g40b86555ee: Adjust orange crystal wall description. 10(3 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/40b86555ee49 15:49:19 <06dolorous_84348> For now, going with adjustment. 16:06:48 03dolorous02 07* 0.34-a0-1110-g8cfb72bb48: Fix comment typo. 10(16 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/8cfb72bb48b5 16:06:48 03dolorous02 07* 0.34-a0-1111-gbd72e070ad: Add another Xom "send in the clones" message. 10(2 minutes ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/bd72e070ada1 16:28:57 <02Darby> for my part, while I do agree that a 1/15 chance was too minimal to be noticed, I think the second course of action (instead *increasing* the chance while expanding these escape spells to include things beyond blink) would make for better gameplay than normal cast weights, possibly with a targeter note like Aim: [monster] (X% to hit, can escape nets) 16:30:09 <02Darby> I feel that sort of means of reducing xv burden by telegraphing things on sight or on action (a la unusual item highlights or the immune to nets note of amorphous monsters) is good, while all resists, defenses, attacks, and mechanics of any kind incentivize knowing what the monsters can do, without which there's no game left (and monsters that blink are somewhat obvious about it due to blinking a lot, just not quite enough to 16:30:10 consistently escape nets at the right time) 16:31:32 <02Darby> (I'd consider trying to flag those spells and implement such a thing myself, but feat commits are still a little outside my niche of currently tiles and upcoming vaults) 16:38:48 Unstable branch on underhound.eu updated to: 0.34-a0-1109-g40b86555ee (34) 16:51:55 <06regret-⸸nde※> I do... find it kind of baffling in general to be dealing continuously with official, presented-to-the-public arguments of "it's bad for any mechanics to exist where people need to read anything about what they fight" in our extremely text-based game that was heavily streamlined down to strategizing over combat. (The thrust of this has come up before for describing status effects, the existence of see invisibility, and quite 16:51:55 a few other effects in the past few years.) We didn't end up with over five hundred monsters after purging most of the harmless spawns from most branches by accident, and as long as weaker or stronger monsters exist so e.g. fire giants do a lot less damage than orbs of fire, the inherited response on many other fronts for not knowing what monsters can and can't do was "good, die". 16:55:11 <06dolorous_84348> This sounds good to me too. Obviously, we don't want to make the monsters too smart, but it seems intuitive that a monster with escape options would, to an extent, actually use them. 16:55:34 <02Darby> yeah, they do have emergency and escape flags for this in other situations 17:00:20 <02Darby> (most instances of monster intelligence being something crawl avoids have other reasons beyond monsters being too smart in the abstract, as I see it: monsters ignoring your resists because the alternative makes the resists feel worse, monsters not fleeing or kiting because it's considered slower or more tedious. this is intelligence that amounts to "effective rNet+", the same way slimes have "actual rNet++") 17:06:49 <06dolorous_84348> Very much agreed. To a point, if they're not reading messages, they're not paying enough attention, and if they're not paying enough attention, their odds of survival should be a lot lower. 17:07:25 <06dolorous_84348> Nothing wrong with showing mercy sometimes, to be clear, but there should be limits to it. 17:09:04 <02Darby> I do think it's possible to reach the point of actual gotchas, a la ToME's reputation for randuniques with one-shot-level abilities, but crawl never goes near that IMO 17:10:20 <02Darby> (I admit that's secondhand and I'm not 100% on the details over there) 17:10:39 <04dracoomega> As a matter of coincidence, I my Refactor So Many Things branch actually had already edited that block of code in the opposite direction - increasing the chance enough to actually matter, while extending it to other variations of blinking instead of just vanilla blink. (This, of course, is some of the problem of working on things that get so sprawling - it can be hard for anyone else to keep tabs on what you're doing until you're 17:10:39 done. I've tried to keep people apprised in general terms of the big things, but lots of small things didn't end up warranting individual mention.) 17:13:13 <04dracoomega> Index made a ridiculous flowchart of how many levels deep I ended up, which is slightly out of date now (having just finished off Sanctuary-related work for now): https://cdn.discordapp.com/attachments/205316046230388737/1434556095640371241/image.png?ex=690e07f9&is=690cb679&hm=cd3125906b3231edecce2283e37d4c0ae0eb54c80aebac26c28f18a407b01db3 17:13:52 <04dracoomega> (Was going to be pushing the whole gray zone seperately of any weapon arts stuff, when I am finally finished with it.) 17:14:24 <02Darby> one of the inherent downsides of large simultaneous drops but I can only imagine the mess of a first impression if you dropped half a weapon art at a time 17:14:30 <08nicolae> draco you're stuck in the flowchart 😰 17:14:39 <04dracoomega> I really have been, I swear 17:15:04 <04dracoomega> Several thousand line diff at this point 17:43:47 <06pleasingfungus> data is beautiful 17:45:18 <02Darby> personally, I like the concept of ****** brand but I'm a bit sketchy on the execution of ********** brand 17:47:53 <06regret-⸸nde※> I still need to make proper additional vampiric graphics for Unleashed Blade.... 17:57:19 03Implojin02 07* 0.34-a0-1112-g21ff419876: feat: Rampage stacking 10(11 minutes ago, 6 files, 71+ 33-) 13https://github.com/crawl/crawl/commit/21ff41987692 18:19:44 <03implojin> Re: 1034eacce the main reason I went the way I did is that we have no existing top-level UI float for things like spflag::escape, and adding such seemed incredibly low return to fixup a 6% check. If someone wants to do proper UI for this, feel free! 18:20:28 <03implojin> Re: Merge conflicts: It's inevitable when you're holding onto branches for weeks, I'm afraid. 18:23:26 <03implojin> I would strongly object to giving players reasons to memorize which spells are flagged as quick_escape without at-a-glance UI support. 18:34:20 <02Monkooky> I feel like for a monster's core gimmick, any burden of reading is fine so long as it's clearly written and easily findable. You can see the monster once, learn what it does by reading, and probably don't need to check again unless specific details come up. Something like this would require an ongoing burden of reading- I can't reasonably memorize e.g. the list of monsters that know blink, and will have to pause to check if a monster 18:34:20 has an escape spell every time I want to net something. 18:37:12 <02Darby> of course, you only have so many nets per game 18:37:35 <02Darby> and for many monsters, it's immediately obvious whether they can or can't blink 18:39:03 <02Darby> (still, the UI indicator does seem like it resolves both sides of concern in one) 18:39:27 <02Darby> (since this is already true of netting monsters that can blink, just somewhat less relevant since they don't have an explicit chance) 18:39:36 <03implojin> note that it's not just blink if you want to expand this to use something like spflag::escape; if you keep it as a very limited spellset as it was then it's less problematic of course but when you look at this and think "hey let's fix it up" it would be good to think about how to float this info to players without forcing them to xv every time they go to throw 18:40:14 <02Darby> I'm always for effective telegraphing, at the least 18:40:48 <06dolorous_84348> Same here. 18:49:23 <02Monkooky> while you say this, insubstantial mons were so bad about this that I actually coded anything 19:15:45 03dolorous02 07* 0.34-a0-1113-gf391d15d98: Describe conditions for sniping ego (RypoFalem). 10(4 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/f391d15d9875 20:10:32 New branch created: pull/4882 (1 commit) 13https://github.com/crawl/crawl/pull/4882 20:10:32 03WizardIke02 07https://github.com/crawl/crawl/pull/4882 * 0.34-a0-1114-g85aa045f0f: Fix felids receiving attacks while dead 10(47 minutes ago, 16 files, 46+ 32-) 13https://github.com/crawl/crawl/commit/85aa045f0fce 20:19:09 03WizardIke02 07https://github.com/crawl/crawl/pull/4882 * 0.34-a0-1114-g8f7f020694: Fix felids receiving attacks while dead 10(56 minutes ago, 17 files, 47+ 33-) 13https://github.com/crawl/crawl/commit/8f7f02069456 20:42:26 <06regret-⸸nde※> The list of monsters that know variations of Blink is infinitely more memorizable than the formula for how Willpower checks work versus given different enemies across the whole game, and we based an entire school of magic around that. 20:42:57 <02Monkooky> But we also tell you the odds of success in the targeter for hexes. 20:43:22 <02Monkooky> So you don't need to check will before firing, you get to check will as part of the process of firing 20:43:27 <06regret-⸸nde※> ...Then we can just update the targetter to say "Can blink out"? As was brought up at the very start of this conversation? 20:43:43 <02Monkooky> Ok I think you think I am arguing for something I am not 20:44:23 <02Darby> hexes are one thing, resistible damage spells are another 20:45:06 <02Monkooky> Having 'can blink out' in the targetter is a perfectly acceptable solution that means I don't need to stop and x-v everything before trying to net 20:45:56 <02Darby> rF against e.g. fireball truly are not told to you after the cast, are difficult to memorize on a per-monster basis, and are cast hundreds or thousands of time per run 20:46:23 <04dracoomega> What I find baffling here (okay, one of multiple things I find baffling here) is that monsters can still currently blink out of nets and have always been able to blink out of nets 20:46:41 <04dracoomega> This is not some radical new proposal, but a thing that has always existed 20:47:25 <04dracoomega> (Nor, frankly, do I think it is uneasonable expect a person with any experience to have some idea if a given monster blinks or not, since that's one of the most visibly obvious things monsters do) 20:47:41 <04dracoomega> And relevant for a whole host of reasons 20:48:14 <06regret-⸸nde※> (...Like berserk, actually, when it comes to something far more regularly accessed across a game for a select god.) 20:50:19 <02Monkooky> Sure! the current situation could be better! but there's a stark difference between 'could blink out of a net faster than it would break free, and is irrelevantly more likely to do so' and 'the chance of blinking out of a net is boosted enough to be tactically relevant' 20:52:08 <06regret-⸸nde※> (...I can't lie, I never realized until checking now one of the two most likely to blink monsters are hellephants.) 20:52:29 <06pleasingfungus> oh yeah for sure 20:52:30 <02Darby> re: nets in particular, I feel they're sufficiently uncommonly used per run (something like a single-digit number? I'd have to objstat), and usually on monsters that unambiguously can or can't blink (you don't have to check if an ogre can blink), that if it's best to give the monster a second glance, it's okay. this might come up two or three times a run. 20:54:03 <02Darby> (not that I'm going back on the can escape nets targeter idea. I just feel like it's worth mentioning in the abstract, since the principle's relevant to other things..) 20:54:12 <08nicolae> take away the blinking from hellephants and then give the tloc stuff to a new monster called telephants 20:55:18 <02Monkooky> though actually now that I'm thinking about this I like the fundamental premise less and less I think this behaviour is relevant in two ways- escape spells acting as strong net resistance, which I'm not convinced is interesting given that nets are mostly constrained by how many you have escape spells meaning you can use nets to force mons to reposition, which sounds really neat but also extremely difficult to communicate 20:58:54 <02Darby> despite their limited quantity, nets are extremely powerful—I've seen them win runs in Zot, against hellpan lords, etc 20:59:10 <02Darby> and forcing monsters to reposition seems like an intuitive interaction of the mechanics to me 20:59:21 <02Darby> similar to how escape and emergency flag spells work 20:59:25 <03implojin> i'd like to point out that any reasoning that hinges on "nets are rare, therefore this doesn't matter" applies more strongly to behavior that happens as a percentage chance while mons are netted 21:00:24 <02Monkooky> hmm, do nets die when mons blink out of them 21:00:35 <02Monkooky> they don't, correct? 21:01:24 <04dracoomega> They do not 21:01:28 <04dracoomega> It works the same as the player 21:01:51 <04dracoomega> (Which is to say that the chance of them being destroyed is based on how much damage was done to them prior to escaping) 21:02:08 <04dracoomega> (I think there's like a 1-in-9 chance to be lost regardless of any damage happening?) 21:02:10 <02Darby> so yes, you could get some more use out of the net (but it's worth noting that the use is more marginal if they almost immediately escape, and they can still break the net anyway assuming the chance isn't 1) 21:02:43 <02Monkooky> well, my initial thought was that this was uninteresting because it just reads 'don't net escaping mons except to make them reposition' 21:02:49 <03implojin> as i wrote above, i really don't think this is worth the time to write UI for, but if whomever picks this up wants to expand the behavior, please do the targeting thing instead of forcing players to go through xv. 21:02:50 <09gammafunk> anyone else get these compilation warnings? In member function ‘void circle_def::init_bbox()’, inlined from ‘void circle_def::init(int, circle_type)’ at coord-circle.cc:79:14, inlined from ‘circle_def::circle_def(int, circle_type)’ at coord-circle.cc:48:9: coord-circle.cc:85:54: warning: ‘*this.circle_def::radius’ may be used uninitialized [-Wmaybe-uninitialized] 85 | origin + 21:02:51 coord_def(radius, radius)); | ^ In member function ‘void circle_def::init_bbox()’, inlined from ‘void circle_def::init(int, circle_type)’ at coord-circle.cc:79:14, inlined from ‘circle_def::circle_def(const coord_def&, int, circle_type)’ at coord-circle.cc:55:9: coord-circle.cc:85:54: warning: ‘*this.circle_def::radius’ may be used 21:02:51 uninitialized [-Wmaybe-uninitialized] 85 | origin + coord_def(radius, radius)); 21:02:58 <03implojin> yes i've seen those for ages 21:03:05 <03implojin> haven't bothered to look into it 21:03:57 <02Monkooky> but if you're not spending the net I'm not sure how well that holds up 21:04:04 <03implojin> there's also these: In file included from AppHdr.h:305, from json.cc:24: In function ‘char* sb_finish(SB*)’, inlined from ‘bool parse_string(const char**, char**)’ at json.cc:966:25: debug.h:58:9: warning: ‘sb.SB::cur’ may be used uninitialized [-Wmaybe-uninitialized] 58 | if (!(p)) AssertFailed(#p, __FILE__, __LINE__); \ | ^~ json.cc:112:5: note: in expansion of macro 21:04:05 ‘ASSERT’ 112 | ASSERT(sb->start <= sb->cur); | ^~~~~~ json.cc: In function ‘bool parse_string(const char**, char**)’: json.cc:852:8: note: ‘sb.SB::cur’ was declared here 852 | SB sb; | ^~ json.cc:71:45: warning: ‘sb.SB::end’ may be used uninitialized [-Wmaybe-uninitialized] 71 | if ((sb)->end - (sb)->cur < (need)) \ | 21:04:05 ~~~~~~~~~~^~~~~~~~~~~ json.cc:957:13: note: in expansion of macro ‘sb_need’ 957 | sb_need(&sb, 4); | ^~~~~~~ json.cc:852:8: note: ‘sb.SB::end’ was declared here 852 | SB sb; | ^~ 21:04:18 <09gammafunk> yep, I get those too 21:04:22 <09gammafunk> I assume these are gcc-only and not in clang 21:04:27 <03implojin> gcc, yes 21:04:28 <09gammafunk> but I haven't teste dthat 21:04:52 Rebuilding vanilla from a `make clean`, LMK if you care if I get it too 21:05:42 <02Darby> I think a net working as a force-reposition tool against some monsters as well as a stab and escape tool against others in a way that clearly follows from their spellsets expands tactical options in an intuitive way 21:06:27 03WizardIke02 07* 0.34-a0-1114-g5523e3846a: Don't leak info about polar vortex casters 10(22 hours ago, 10 files, 74+ 39-) 13https://github.com/crawl/crawl/commit/5523e3846a8f 21:07:37 <02Monkooky> I think anything which relies on monster cast rates to this degree is rather distasteful 21:08:19 <02Darby> of course, the cast rates are irrelevant if they're increased while in a net—it only relies on cast rates in the current status quo 21:08:20 <02Monkooky> and I'm not sure how you'd concisely communicate how effective netting a mon is for making them blink 21:08:53 <02Darby> (although I would broadly eventually prefer cast rate transparency, that's further off in the distance) 21:12:33 <06dolorous_84348> Maybe the difference between "will sometimes blink out" and "will likely blink out"? 21:13:23 <06dolorous_84348> Since we're not displaying actual numbers? 21:14:33 <03implojin> or, and hear me out, we could leave the chance to cast spells at whatever it's normally at, instead of changing it while monsters are netted, and that way we don't have to bother explaining how it changes in the ui 21:25:57 <02Darby> I think past arguments about monster intelligence (the aforementioned dynamics like not fleeing, not kiting the player, not reacting to situational resists) have given an impression that monster intelligence itself is the problem, as opposed to each of those examples leading to bad gameplay 21:28:19 <04dracoomega> How likely monsters are to cast a variety of things is already based on multiple properties of the battle situation in ways that are complex to specifically articulate to the player and - I would argue - also unnecessary for them to know with that level of detail. The chance of monsters sidestepping or keeping distance around corners, whether they will zap a spell through a resistant ally (whose chance is effectively randomized!), 21:28:19 whether a variety of spells are eligible to be cast at all, for numerous reasons. I don't think any part of the experience of playing this game - even at a fairly high level - depends on the player having this fine-grained detail, and yet this behavioral code still serves a valuable purpose in creating the overall experience of how things play out; it isn't a mistake that you can't reduce monster behavior to a very simple deterministic function. 21:29:04 <02Darby> monsters reacting moderately intelligently is good, when it makes for good gameplay and comes with low design costs—I think existing foe_ratio code, emergency and escape spells, and other modifications to cast chance fall in this boat 21:30:19 <02Darby> this feels like another case of "making the obvious decision in a way that makes the tactical layer a little more nuanced, while not being any more annoying than any other case of resisting a tool" 21:31:02 <03implojin> let us say that darby is right and players throw ~10 nets per game. with the behavioral code that was in place for the last 11 years, said players have witnessed this intentional netblink behavior approximately zero times per game. this particular thing isn't important. 21:31:30 <03implojin> but you're of course free to spend your time implementing what you want 21:31:32 <02Darby> the 1 in 15 chance was way too low 21:31:40 <02Darby> I think, like, that's one thing everyone here agrees on 21:38:39 <06dolorous_84348> Yes. And I've actually seen the behavior once in recent memory: I've been trying to get OnGl off the ground, and a few days ago I caught Dowan in a net, and he blinked out of it. 21:39:10 <06dolorous_84348> (I'm still better at coding for this game than actually playing it.) 21:59:47 <06pleasingfungus> https://www.joewintergreen.com/design-thoughts-let-things-be-rare/ contemplating 22:00:07 <06pleasingfungus> (not staking out an opinion here, just floating a thought) 22:00:13 03WizardIke02 07* 0.34-a0-1115-g7087b09c79: Fix warning about redefining integer limits in MSVC 10(12 minutes ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/7087b09c79c0 22:00:31 <04dracoomega> In case I've been unclear about this, the specifics of this net behavior are of relatively minor importance. I agree that the behavior which existed was pointlessly unlikely. I thought it would be more interesting if the chance was raised relevantly high instead (and would give some monsters a chance to look 'smart' in ways that are reasonably intuitive and potentially interesting). I asked a couple other people at the time, and they 22:00:32 agreed with me. It ultimately isn't that important one way or another. What is important is the philosophy underpinning this entire argument (and many others like it). What monsters do or don't do about nets is ultimately small in the grand scheme of Crawl. But the idea that it is bad for a monster's capabilities to matter in a way that actually influences the player's tactics is such unacceptable nonsense and that is practically what you wrote 22:00:32 directly in your commit removing this net thing. And, in one form or another, it has been restated again tonight, as it is sometimes restated elsewhere. That is what I have the biggest issue with. 22:00:34 <04dracoomega> If asking a player to care about what spells a monster has is 'bad UX', then what are we even doing here? What is the point of making anything if we want it to be inconsequential? The fact that monster behavior is influenced by factors that cannot be neatly boiled down into X chance of performing Y action each turn is sometimes spoken of as some kind of intrinsic design flaw to be eliminated, but Crawl has never been that kind of 22:00:34 game. It has always tried to make monsters seem like something other than robots, who play by rules that at least mimic the player's own on a surface level, that move as semi-intelligent groups that at least can convey the fiction of internal agency. There are so many behaviors that they have under the hood. We give players the tools to know the most important things they need to be aware of and react to, but the game has never asked nor expected 22:00:34 that anyone knows precisely what motivates every action a monster takes and this is not a flaw. It is part of what makes a 'living' environment to many players who will never be good enough for the exact chance of moving left or right that turn to matter. 22:01:42 <03implojin> that isn't what I wrote at all? 22:02:18 <03implojin> > but the end result of that would have been that > players with nets would need to be much more aware of the specifics of the > spellsets their net targets. This expected UX of incentivizing players > to press xv before huckin' nets seems like it would make for worse > gameplay, so I'm not doing that. 22:03:12 <03implojin> What I wrote was that the UX of expecting players to either memorize an escape spellset that as far as I know isn't enumerated anywhere in-game didn't sound like it would make for fun gameplay, and I stand by that? 22:08:15 <02Darby> (for what it's worth, I looked through all the monster spells and as far as I can tell only the ones with "blink" in the name can escape nets in an organic way. a few repositioners like bombard or vhi's can unreliably break out as a consequence but I assume those wouldn't be on the list.) 22:11:37 <02Darby> (it helps in this case that so many monsters that do natural repositioning things like hopping still present it as some form of Blink in xv) 22:58:13 03dolorous02 07* 0.34-a0-1116-gf51bd3897e: Move glowing colours before colour patterns. 10(24 minutes ago, 1 file, 39+ 39-) 13https://github.com/crawl/crawl/commit/f51bd3897ef0 22:58:13 03dolorous02 07* 0.34-a0-1117-g8fd88fbf24: Add coloured writing as a colour pattern option. 10(22 minutes ago, 1 file, 57+ 0-) 13https://github.com/crawl/crawl/commit/8fd88fbf2406 22:58:13 03dolorous02 07* 0.34-a0-1118-g74d9ffc8be: Adjust a sparkling fountain decor message. 10(6 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/74d9ffc8be40 23:00:23 03Implojin02 07* 0.34-a0-1119-g5c53252f6f: Fix an ldoc lua api comment (Omniraptor) 10(5 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/5c53252f6f6d 23:12:39 04Build failed for 08master @ 74d9ffc8 06https://github.com/crawl/crawl/actions/runs/19159663024 23:35:39 Unstable branch on crawl.develz.org updated to: 0.34-a0-1119-g5c53252f6f (34) 23:55:07 03dolorous02 07* 0.34-a0-1120-g30376c0ee6: Move colour patterns before coloured writing. 10(21 minutes ago, 1 file, 54+ 54-) 13https://github.com/crawl/crawl/commit/30376c0ee6fa 23:55:07 03dolorous02 07* 0.34-a0-1121-g4f141f716d: Add a few more writing adjectives and nouns. 10(13 minutes ago, 1 file, 8+ 0-) 13https://github.com/crawl/crawl/commit/4f141f716dc7 23:59:07 Windows builds of master branch on crawl.develz.org updated to: 0.34-a0-1119-g5c53252f6f