00:41:24 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-5014-gf9597c7f93 00:49:26 Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.03-2234-g30d508e18d 03:22:53 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5014-gf9597c7f93 05:06:31 Unstable branch on crawl.akrasiac.org updated to: 0.31-a0-640-g39d97e5 (34) 08:02:17 <05icoson> I think the stack may be misleading (clua stack at the bottom looks buggy?); that is an ASSERT_RANGE crash so is happening in C++ 08:02:39 <05icoson> possibly it's directly in an l-*.cc file so we don't see much 12:38:31 <09gammafunk> Hmm, OK. Maybe it will trigger again and I can get a seed via wizard mode. But I just paused qw right before it engaged those gnolls, manually moved forward a bit, resumed qw, and then no further crashing. Situation with gnolls didn't seem special in any particular way. A group of them at end of corridor opening into a room with unobstructed LOF to closest one. Qw just seemed to land a shot and then the crash. 12:53:34 the most likely candidate for a range check error with max 70 is GYM, so some sort of out of bounds issue; historically it looks like this has previously resulted from ranged combat which that char had (`cell_is_solid` triggers a range check on env.grid for example) 12:53:55 I guess it is possible that some lua code itself is triggering an unprotected env.grid access 14:02:09 -!- daethoven83 is now known as daethoven8 14:09:29 <04dracoomega> So... I have been giving things some thought. 14:09:37 <04dracoomega> Concerns were raised about Sigil of Binding being too good for continuous kiting of things (ie: bind, back up, shoot a couple times, cast again and rebind, etc.) and while I feel like in practice it's actually a fair bit trickier and more space-requiring to pull this off than was stated (at least in my own multiple earlygame playthroughs), I was thinking of a way to specifically counter this while still being able to use it for a 14:09:38 temporary lock-down, space opener, or safe inner flame detonator that also requires usual deliberate player movement that I rather liked. 14:09:38 <04dracoomega> What if the sigil essentially stole future momentum from an enemy. And when the binding broke off, the enemy would become swift for a while afterward (maybe proportional to how long it was bound, but slightly longer) and could not be rebound while swift. Then you could use it for a single bind, but it would be overall bad in the long run if you were trying to 'abuse' it. 14:09:53 <04dracoomega> Also, having messed up the spawn chance for jinxbite obviously means that the 'fixed' version is a lot stronger than was planned. But sort of glomming feedback together, what if we got rid of the jinxsprite spawning, just had the direct damage/debuff on WL check per attack, and maybe moved the debuff to just plain dazing something? The slow/weak coinflip was admittedly an attempt to nerf it from just slow, but I can see the point 14:09:53 of there being a lot of complexity in one spell. I sort of wish we had more minor debuffs that were okay to throw around en mass at low level. There's very few, really. (I considered daze, but daze often does nothing at all. Maybe daze and a mini-stun on first application and leave it at that? The hope was that the damage also mattered earlygame) Another thing I've considered that might be a more interesting downside to jinxbite is... what if it gave 14:09:54 will 1/2 to the player for a while on casting? This is a downside that, due to formulas, ought to have significantly more impact on enemy spell chances late instead of early, but doesn't have quite the same 'rarely matters and then unpredictably matters a lot' that a few other downsides tried did. I realize this isn't a visible downside in SOME battles, of course, but... other things can always show up that makes it matter. What do people think? 15:25:08 Unstable branch on underhound.eu updated to: 0.31-a0-640-g39d97e502b (34) 16:55:47 <06pleasingfungus> re jinxbite: daze is just slow with worse ui; i’m planning to cut it at some point 16:56:16 <06pleasingfungus> would be interesting to brainstorm other debuffs 16:56:39 <06pleasingfungus> agree with cutting the sprites and applying a debuff directly 16:57:45 <06pleasingfungus> will/2 is an extremely situationally relevant effect; it’s mostly irrelevant 16:58:01 <06pleasingfungus> think it’s fun but not useful for limiting jinx’s long term utiltiy 17:01:14 <06pleasingfungus> re minor debuffs, i suggest drain 17:01:48 <06pleasingfungus> that weakens enemies in a few way (accuracy, spellpower, etc), and also scales poorly with xl 17:02:05 <06pleasingfungus> just weak would also be fine, imo 17:05:15 <06pleasingfungus> re sigil of binding - i like this line of thought. sort of weird if you can give yourself swiftness with binding, though - probably not exploitable but probably a bit complex. alternate idea - what if, when bind sigil elapses, the victim is blinked to a random space adj to the caster (if possible)? would avoid the need for any status tracking - maybe too harsh…? 17:05:49 <06pleasingfungus> a little worries about tweaking it by making it more complex (either way), but down to try things 17:05:50 <04dracoomega> Well, I assumed the player didn't actually get any benefit to sticking themself in place. Can flavor it as some sort of magical feedback, since you were the source of the magic in the irst place. 17:06:15 <06pleasingfungus> so much to explain to players :p 17:06:22 <04dracoomega> Otherwise you could sort of... 'pre-buff' by placing a sigil, standing on it, and waiting to charge up. Not useful mid-battle, but useful before doing some things and feels a little weird 17:06:27 <08wormsofcan> give the player attraction status 🤔 17:06:32 <06pleasingfungus> lol 17:07:23 <06pleasingfungus> also re bsig, wdyt about going from 2 traps to 1 but making em trigger on any moves into adj tiles too? would simplify eg dealing with monster diagonal move choices 17:07:29 <06pleasingfungus> (mostly) 17:08:55 <04dracoomega> Having played with the spell a fair bit myself, I think that: a) Triggering on moves into adjacent tiles will make it much too trivial to trigger. Like, one of the limiting factors of the spell is the chance of things being placed in awkward locations compared to the shape of the current battle, and I feel having a 'range' on it will make it almost always straightforward to trap a thing without needing to do much. b) In situations 17:08:55 of multiple enemies nearby, that sounds like it will end up trapping a completely unpredictable enemy, since it will depend on their internal order 17:09:29 <04dracoomega> Like, whatever ends up moving 'first', which is something opaque to the player and normally doesn't matter 17:10:42 <04dracoomega> (Have you actually played with it yourself, by the way?) 17:12:03 <06pleasingfungus> no sorry, i’ve had approx 2min dev time this week 17:12:12 <06pleasingfungus> nonstop toddler/dayjob mode lately 17:12:19 <06pleasingfungus> only get to chat on mobile in idle moments 17:13:09 <06pleasingfungus> re triviality, it seems like it preserves the behavior of bsig being unsafe in corridors 17:13:20 <06pleasingfungus> no opinion re broader context 17:13:31 <06pleasingfungus> will defer to yr judg 17:15:04 <04dracoomega> I'm curious... if we try the 'give swift on unbind and prevent rebinding while swift' and obviously don't want to give swift to the player, is it more intuitive to bind the player if they step on it and just not give swift, or to do something entirely different about like 'magical feedback' like just removing the sigil and draining some mp or some other 'feedback' style penalty. 17:15:49 <06pleasingfungus> i think it’s probably fine to just give the player swift if we do that 17:15:59 <06pleasingfungus> is it actually abusable? 17:17:38 <06pleasingfungus> another thought re bsig - there’s some world where we radically simplify it and make it just instantly bind one critter in some radius. plausibly this is a more boring world; unclear 17:17:47 <04dracoomega> Well, if you make one in a corridor and step on it, assuming you don't get immediately clobbered, you actually make up more distance than you lost by standing still. Also I can see someone making one at the start of a battle that isn't too scary, but which they might later want to withdraw, and know that they're safe for a few turns and then can get a free swift ready and waiting afterward to withdraw if needed 17:18:10 <04dracoomega> I felt the making of trap and moving monster onto trap was pretty fun, personally. A somewhat unique movement consideration 17:18:42 <06pleasingfungus> re the former case, are you counting -swift? 17:18:55 <04dracoomega> Well, I am assuming monsters don't get that 17:18:56 <06pleasingfungus> re the latter, that sounds extremely unwise… 17:19:03 <06pleasingfungus> so? players do 17:19:12 <06pleasingfungus> that’s already how player swift works 17:19:21 <06pleasingfungus> wouldn’t have to add anything new or explain anything 17:19:39 <04dracoomega> Hmm.... 17:20:40 <04dracoomega> I mean, that's still an asymmetric effect for stepping on it, only a more complicated one: 'it will stop you, then speed you, then slow you' instead of just 'it will stop you' (or some other single penalty). I realize swiftness the spell already works that way, so there's some precedent, but I guess I'm not convinced that this is a smoother asymmetry? 17:20:49 <06pleasingfungus> it would actually be quite a bit of extra work to make bind give swift to players and without -swifting em after 17:21:00 <06pleasingfungus> i mean my feeling is that it’s dangerously complex 17:21:09 <06pleasingfungus> which is why i suggested several alternatives above 17:21:12 <04dracoomega> I mean, I don't want to give swift to players with no downside at all, to be clear! 17:21:16 <06pleasingfungus> sure 17:21:22 <04dracoomega> I just didn't want to give them anything 😛 17:21:36 <06pleasingfungus> sure, so do the blink thing i suggested 17:23:19 <04dracoomega> I'm... I'm not sure how that train of thought goes from there to here. Like 'don't want to let players move fast because of this' to... 'we can't make enemies move fast either and so should blink on top of you?'. In that case, where is the player supposed to be blinking? 17:23:21 <04dracoomega> Did I misunderstand something? 17:31:51 <06pleasingfungus> the rule becomes ‘when the sigil expires, the victim is transported/rebounded to the caster’ 17:32:04 <06pleasingfungus> if the caster is the victim, logically… 17:33:07 <04dracoomega> This is, of course, a lot meaner if trying to use this to escape even a little 17:33:28 <04dracoomega> Since even if you're just trying to flee immediately, it only buys you a few steps before you have the same problem you just had 18:01:45 <06pleasingfungus> sure 18:31:13 <09gammafunk> @icoson Hrm, I guess the crash did trigger again repeatedly on another seed: https://crawl.dcss.io/crawl/morgue/qw/crash-qw-20230827-224255.txt Similar out of bound problem. Probably unrelated to this, but apprently qw is using up more than 64 mb of memory (I've raised its limits on cdi again) sometimes and I need to investigate this a bit more to see if it's some kind of leak in my new retreating code or if that's just how things 18:31:13 are now, but it's somewhat suspicious that I've seen this hitting the new 64mp memory limit twice now in two different games when fighting a single monster. Would think that even though I've not added a coroutine yield in the intensive search loop, that it wouldn't be doing this search for a single monster and would have had opportunity to garbage collect. Makes me worried that the out of bound crash is somehow related to lua-side corruption from 18:31:14 hitting the memory limit and then passing bad data to a C++ function in one of hte l-*.cc files. But maybe it's just some latent bug in one of those functions like you mentioned. 18:33:44 <09gammafunk> @dracoomega not sure if you saw, but Monkooky posted this thread in tavern after some playtesting: https://tavern.dcss.io/t/feedback-hexslinger-rework-experimental-branch/383 In retrospect it was silly of me to not think of a tavern thread as a way to collect feedback since you'd prefer not to get Discord pings. Feel free to respond there if you like, of course, and relive the fun of Tavern discussion all over again, or maybe just 18:33:44 subscribe for posts. 19:19:02 -!- pivo is now known as p1v0 19:39:01 <04dracoomega> Haha, I never once posted in old Tavern, y'know. 😛 19:39:07 <04dracoomega> Still helpful to read in this case, sure 19:39:55 <04dracoomega> (Honestly, old Tavern seemed especially divorced from the segment of the playerbase that most of the devs took seriously. I never really read it very much at the time?) 19:40:16 <04dracoomega> I'd actually seen that bug with bind myself - just enough that this is familiar without particularly knowing why it was happening 19:41:00 <04dracoomega> Like... bind specifically was only blocking monsters choosing to move, and wasn't coded to block displacement of various sorts, which seemed 'fine', but even if something could blink or trample them, I have no idea why a jinxsprite was doing that 19:41:37 <04dracoomega> Not that jinxsprites, as-is, may be around for long. (Though I do think there's room for something else using foxfire's general idea of seeker projectiles, in a different way, even if this isn't it) 20:20:03 <06pleasingfungus> (and different from ioods and cbl…) 20:20:12 <06pleasingfungus> (crawl sure loves its fake monsters!) 20:24:52 <04dracoomega> I mean, I think they can be neat 20:26:30 <06pleasingfungus> they can! 🙂 21:53:49 <09hjklyubn> @dracoomega inner flame is still resistible. and also I can't use inner flame to monsters with infinite will.(maybe this is intended?) 21:53:49 <09hjklyubn> https://cdn.discordapp.com/attachments/747522859361894521/1145581981971787776/000.PNG 21:54:05 <09hjklyubn> also, is it correct that the chance of generating fire cloud depends on spell power? My spell power is only 2%, but the cloud is generating very well. 21:54:06 <09hjklyubn> https://cdn.discordapp.com/attachments/747522859361894521/1145582052452872202/001.PNG 22:09:12 <04dracoomega> It was intended that it still didn't work on things with infinite will, but it shouldn't be resistible by a titan. 22:10:21 <04dracoomega> Also: spellpower affects cloud number / duration on the explosion, but doesn't currently affect getting a minor flame cloud upon hitting things. Do you think it should? 22:11:02 <04dracoomega> (I kind of didn't want to make that too unreliable, since it's already quite a transient thing anyway) 22:15:30 <09hjklyubn> I haven't had a chance to do a proper test play yet and haven't thought about it much, so it's hard to give good feedback now. If I get a chance, I'll play around with it more and give you some feedback about that! 22:16:06 <04dracoomega> Sure 23:45:26 Monster database of master branch on crawl.develz.org updated to: 0.31-a0-640-g39d97e502b