03:32:10 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b 04:08:32 <04d​racoomega> I wonder a little at whether these 'more themed' troves could be pre-announced, potentially making otherwise-expensive fees feel more reasonable? Or if that just causes people to ignore them more often (since they know they don't want any of 'X-style' loot?) There's theoretically more agency in paying a price for something you know a little more about than 'It could be anything.', but possibly this makes people write off more 04:08:32 troves as not something they'd want? Not sure. There's definitely an odd dynamic at the moment where a few trove fees are 'close enough to free' that you look into them as soon as you can, but often the opportunity to enter comes so late that it's hard for what's inside them to actually be exciting. You expect to have most of your kit figured out by that point. To some degree, I think that entire 'Crawl generates so many items that eventually you 04:08:33 don't care' thing you mention is somewhat a consequence of Crawl having less 'vertical' scaling of item power than a lot of such games have. ie: you can get endgame-quality items fairly early, if you're lucky (and you get enough rolls to be lucky on aggregate about at least a few slots, quite regularly.) And that is often interesting (can motivate pivots and makes runs feel more different from other runs), but it does make it a bit harder to keep 04:08:33 dropping things and have them regularly feel worth using over other things you already have. 04:09:42 <04d​racoomega> Whereas games with more vertical item progression can expect you to want to replace most of your kit many times, and so can keep dropping slightly higher tier things over a long period and expect them to still be regularly relevant 04:10:18 <04d​racoomega> That's something that isn't easily changed (and isn't even strictly better to have changed, of course) 04:14:03 <04d​racoomega> I can definitely see some logic in using banes as entry fees. Honestly, the fees that arguably work the best at present are the ones that are 'challenges' more than costs (ie: lose piety, go to the Abyss or Hell) and it might be possible to explore more of this space. I understand Index's concerns, but there's no reason this has to even be 'fully random bane', but any sort of specific medium-term penalty we want (which can be 04:14:04 preannounced). (LIke, there's two aspects to the bane framework - one of which is 'status effect that builds up and can give you a random curated medium-term penalty eventually', but another is just the ability to give any medium-term penalty we want, and have it have its own XP cost, so that can be used flexibly.) 04:14:21 <04d​racoomega> The trove could even pre-announce it 04:16:42 <04d​racoomega> Though there's possibly some risk with that that the different penalties end up somewhat character-dependent, such that the same thing is untakeable for some and trivial for others. 04:17:41 <04d​racoomega> (Although that isn't so different from how trove fees kind of currently degenerate. Some are close enough to free and others aren't worth considering. It's especially silly when they, say, ask for more potions of haste than you find in the whole 3 rune game ^^; ) 04:23:58 <04d​racoomega> Honestly, this does make me think that 'banes' which forbid specific consumables (or groups of consumables) for a while may be just plain more interesting than asking for them as a fee, because we know the player is able to afford that (and also that it can almost never be 'free'). Wereas I've definitely had trove fees where it wasn't just asking me to survive on less resources, but rather that even if I gave up on using that item 04:23:58 for the whole game, I wasn't guaranteed I'd ever be able to enter. (And then, once in a blue moon, the RNG gives you so many of them that paying the fee is eventually barely a cost) 04:24:50 <04d​racoomega> Also 'accumulate a whole bunch of X' sort of automatically pushes them late into the game, most of the time. 04:25:52 <04d​racoomega> Whereas, say, 'no ?tele or ?blinking for X floors' is scarier early on, but that's also when the trove reward could be the most impactful, making it more tempting 04:38:03 <06m​umra> Half-formed idea but ... generate as a branch entry vault, to guarantee it will be seen, and present an offer like "you can be granted access to this trove but will apply to you until you have obtained the rune from this branch" (or alternately, defeat X inhabitants of this branch) 04:39:31 <04d​racoomega> When does that offer become invalid? If the player doesn't pick it immediately? 04:40:30 <06m​umra> I think it'd have to be (at least which you're in LOS of the portal) 04:41:04 <02M​onkooky> Could put it outside the entrance; need to accept before entering 04:41:24 <06m​umra> The point is the player has to do something now, not come back later when it's easier 04:41:50 <06m​umra> If they can't complete the quest and have to duck out of the branch to go somewhere else, well they can do, but they have to live with whatever status they picked up 04:48:26 <04d​racoomega> In this schema, I think I probably prefer 'immediately outside the entrance' so that we can at least prompt the player when entering, rather than it being LoS based (and more possible to mess up by accident, if a fight breaks out near it at the time) 05:03:40 <06m​umra> story wise i'm thinking along such lines: "A mysterious old woman appears and recounts how a dishonest Naga once tricked her out of her most precious possession. If you agree to take revenge she will in return grant you access to her {library|wardrobe|workshop|armoury|other thematic name}; to ensure you keep up your side of the bargain you will suffer a curse of {downside} until the task is completed." Ok maybe a bit wordy but you get 05:03:41 the drift 05:09:08 <06m​umra> (Maybe rather than the rune, there's a monster somewhere in the branch who's the target, so it's properly more of a sidequest) 05:14:45 <12g​e0ff> > having it [gold pull] active always, but only while resting There are a few problems with picking up gold, which would be great to solve one way or another: * It's extremely tedious if you play a turn-count/collect-all-the-gems game. For example, you explore Orc and you xv each pile of gold to see its size, so you can decide if it's worth spending 10 turns to get it * even if you autoexplore, there's still gold that lands on traps, 05:14:45 which is mega-annoying in Tomb, and gold near Slime walls 05:18:57 <12g​e0ff> (also picking up gold in Abyss is problematic: you need to do it manually, otherwise autoexplore could walk you into back to monsters or into already explored territory) 05:19:44 <12g​e0ff> Another solution could be activating a full-los gold pull when your gold aura ends. So you killed a bunch of enemies, the screen is full of coins everywhere, the flavour is over the top. And when the battle ends, you auto-collect all the 🪙s 05:21:42 <12g​e0ff> (so even if you have no time to rest, the gold will be pulled to you) 05:22:00 New branch created: pull/4427 (1 commit) 13https://github.com/crawl/crawl/pull/4427 05:22:00 03WizardIke02 07https://github.com/crawl/crawl/pull/4427 * 0.33-a0-1181-g2e8e60c804: Don't waste a turn when summoning with no space (NormalPerson7) 10(3 minutes ago, 9 files, 541+ 84-) 13https://github.com/crawl/crawl/commit/2e8e60c80434 05:23:31 Unstable branch on crawl.akrasiac.org updated to: 0.33-a0-1189-g0c5ade9 (34) 05:24:07 03WizardIke02 07https://github.com/crawl/crawl/pull/4427 * 0.33-a0-1181-gabb0a8dc44: Don't waste a turn when summoning with no space (NormalPerson7) 10(5 minutes ago, 9 files, 541+ 84-) 13https://github.com/crawl/crawl/commit/abb0a8dc44f1 05:37:19 <06m​umra> I think .. if turncount is a big motivator of wanting this feature, making Gozag the defacto god for turncount games purely down to interface convenience seems .. not ideal? (Gozag doesn't even seem to care about gems, only actual gold ...) Obviously there is apportation to help with traps/slime but yeah it is mainly just an inconvenience once all the monsters are dead. Possible solutions that could help all players, not just Gozag 05:37:19 followers; 1) allow apportation to just grab all gold in sight, in one cast, if you target a single tile of gold (gold is smaller and easier to manipulate than other bulky items) 2) don't ever drop gold (or any items?) onto traps/slime/other dangerous features, shift them to nearby floor instead (seems awkward) 3) let gold just be picked up automatically and instantly from adjacent tiles as you walk past, spending a turn to pick up gold doesn't seem to 05:37:20 serve much purpose, unlike consumables or a weapon/ammo that you could want to use immediately in a combat situation (but just seems weird to have gold work so differently to any other item) 05:40:22 <12g​e0ff> having special rules for picking up one and only one type of loot would be unusual and surprising for the players 05:41:11 <12g​e0ff> (although, your solution #2 feels less special-casey than the rest) 05:42:28 <06m​umra> yeah it still has unfortunate conditions like "what happens in a tomb room where the whole floor is traps" etc 05:43:30 <12g​e0ff> Also, it not only about turncount players, it's about convenience of playing Gozag. Compare it with the latest version of "auto pray on corpses": it was mostly automated in the end, but the whole concept of requiring the player to do the praying on items was removed anyway, cause it doesn't add anything fun to the game! 05:45:01 <06m​umra> it might seem reasonable for gozag to treat dangerous features the same as deep water in general and just give you the gold 05:45:18 <12g​e0ff> fr: Gozag's 💳 05:45:39 <12g​e0ff> (you automatically get your gold into your Gozag account) 05:46:23 <06m​umra> has nobody made a gozag meme coin yet 😂 05:46:46 <12g​e0ff> i agree with PF that having gold everywhere is flavourful, but it turns the player into a Roomba 05:47:12 <12g​e0ff> 💡: what if the gold was purely decorative, like blood-splatters? 05:47:58 <12g​e0ff> (only the gold from Gozag's gold touch, i mean) 05:48:57 <06m​umra> (one-off gozag ability to buy a single shop item "on credit" and you have to pay it back with extortionate interest?) 07:43:55 <09h​ellmonk> we have the infrastructure for this (descent mode shop voucher) but its kind of antisynergistic w/ the rest of the god 07:44:16 <09h​ellmonk> like this idea 07:50:13 03dolorous02 07* 0.33-a0-1190-gc9706edb85: Various food cache message updates. 10(6 minutes ago, 1 file, 35+ 11-) 13https://github.com/crawl/crawl/commit/c9706edb85c3 08:07:10 <06m​umra> not so sure it's completely anti-synergistic ... like if you can afford to summon a shop but then don't have the money to buy a very good item that you see, the shop was kind of a waste of money in the short term. but if gozag can front you the money ... but then will take 50% of all gold you find until it's paid off plus interest ... so you can still build gold back up for other abilities but the item will make things easier right now 08:16:22 <06m​umra> (i think it'd have to only work at gozag-created shops ... otherwise you could just convert to gozag with little gold, buy an item then abandon them ... i suppose the item could be destroyed on abandonment if it's not yet paid off in full) 08:50:28 What is this new player ghost level is it a replacement for ossuary 08:52:18 I really like the look of the place nice blue and purple hues 08:53:16 and star light esque water/trees/funerary flowers 08:53:25 Death doesn't have to be monochrome and boring! 08:55:08 "Wyrmholes have climbed to the top of killers in Depths since their 08:55:08 addition three months ago, which is admittedly further than I 08:55:09 particularly planned- while it's good for Depths to have more 08:55:09 extremely distinct monsters from V:$ and most of Zot, they're not meant 08:55:10 to quite quash down any other comparative threats in the branch. (It's 08:55:10 hard to tell how many people also died to not bothering to cancel the 08:55:11 teleport, though.) They've been pushed deeper into the Depths 08:55:11 depth-scaling and are somewhat less common (with alderking weight 08:55:12 pushed up slightly to compensate), as a consequence. They also have had 08:55:12 some of their numbers slashed slightly- less hp, lower cast rate for 08:55:13 their breath, and speed 12 over speed 13. I'll still be keeping an eye 08:55:13 on them, but I'd rather reduce their numbers a little further if needed 08:55:14 rather than otherwise nerf their suite of defenses or their central 08:55:14 displacement gimmicks. 08:55:15 " 08:55:17 I can answer that question kind off 08:56:53 less people died because the "evil" teleport is a blessing in disguise becaus eit means you no longer get hammered by 3d999 irresistible damage blasts from a super tank 08:59:43 btw found easy counterplay to nobody 08:59:59 8-9 crystal spears did the trick 09:00:11 anyone can do it really 10:16:32 Hi, is there a Discord specificlly for crawl development? 10:18:25 there is, but it's restricted to committers 10:18:44 mostly because of moderation concerns 10:19:05 it's bridged here though 10:19:34 Ahhh fair, thanks for the info! Just wanted to peer into the crawl design choice mentality haha 10:23:49 you'll see most of it here, some of it on the regular Crawl discord (#dcss on the Reddit roguelikes Discord) 10:24:50 some of it in other places but there will be pointers here when that happens and I think they're usually out-of-the-way but open 10:25:17 crawl devs aren't generally into the "sekrit cabal" business 10:43:01 Djinn can't put on Lear's because they have no feet : ( 10:43:07 Happy Xom's Day btw 10:43:16 <3 that 11:10:23 03mumra02 07* 0.33-a0-1191-g126c83ba82: fix(webtiles): Make rcfile editing change backwards compatible 10(3 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/126c83ba8240 11:20:59 Can you get a "useless"** spell on Djinn like tempering without having any other forgecraft spell 11:26:33 04Build failed for 08master @ 126c83ba 06https://github.com/crawl/crawl/actions/runs/14202900733 11:44:03 <04d​racoomega> You shouldn't 11:44:18 <04d​racoomega> There are specific checks for things like that (same with battlesphere and having gotten at least something that can trigger it first) 11:44:37 <04d​racoomega> Now, it mightn't be an exciting thing to use it with, but you will always have something 11:46:40 dracoomega thank you 12:29:51 New branch created: pull/4428 (1 commit) 13https://github.com/crawl/crawl/pull/4428 12:29:52 03NormalPerson702 07https://github.com/crawl/crawl/pull/4428 * 0.33-a0-1150-gc65fe420b9: Split up dart/javelin brands for autopickup 10(2 days ago, 41 files, 351+ 177-) 13https://github.com/crawl/crawl/commit/c65fe420b989 12:37:18 03NormalPerson702 07https://github.com/crawl/crawl/pull/4428 * 0.33-a0-1192-gf5476fb577: Split up dart/javelin brands for autopickup 10(2 days ago, 41 files, 352+ 178-) 13https://github.com/crawl/crawl/commit/f5476fb57739 12:44:44 03nlavsky02 07* 0.33-a0-1192-gbf953fe936: Remove quokka 10(3 minutes ago, 1 file, 0+ 18-) 13https://github.com/crawl/crawl/commit/bf953fe936e9 12:44:44 03NormalPerson702 07https://github.com/crawl/crawl/pull/4428 * 0.33-a0-1193-g4f45f511bd: fix: Don't convert old curare needles to poisoned darts 10(83 seconds ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/4f45f511bd00 12:48:59 03NormalPerson702 07https://github.com/crawl/crawl/pull/4428 * 0.33-a0-1194-g25a6a645c1: checkwhite 10(2 minutes ago, 2 files, 2+ 2-) 13https://github.com/crawl/crawl/commit/25a6a645c1be 15:39:01 Unstable branch on underhound.eu updated to: 0.33-a0-1192-gbf953fe936 (34) 16:44:09 New branch created: pull/4429 (1 commit) 13https://github.com/crawl/crawl/pull/4429 16:44:09 03mumra02 07https://github.com/crawl/crawl/pull/4429 * 0.33-a0-1190-gbb89c02515: Add a cachebusting key to lobby static assets 10(8 minutes ago, 3 files, 6+ 3-) 13https://github.com/crawl/crawl/commit/bb89c025153a 18:45:20 Hey - wondering if anyone has experience building for Android on Windows. 18:45:21 I followed the INSTALL.md guide and was able to compile on Windows. For the next step on android, it's unclear what libraries I should install for Msys2 (the recommended windows build env). 18:45:21 When I run `make ANDROID=1 android` I got a segmentation fault after I individually installed libpng through pacman: https://pastebin.com/Xf8Njqzd. 18:48:08 that seems weird. tilegen should build for (and your log shows it being built for) and run on (same) the host, so there should be no difference from building natively for Windows 18:48:17 but, at least in CI, that succeeds 18:51:29 yeah narrowing it down it's just `$ ./tilegen.elf -i ../dc-main.txt 18:51:30 Segmentation fault 18:51:30 ` 18:51:49 i can try building for windows again if that clears things up 18:51:56 sadly I have no idea how you debug on Windows 18:52:15 (I do Linux and FreeBSD only) 18:52:15 no worries - thanks! 18:52:37 maybe i should've used WSL - msys2 was the recommendation. 18:53:25 ah - building for windows also failed 18:53:30 i must have done something wrong 18:53:50 it might be worth trying a regular Windows build to see if tilegen.elf fails the same way (I would expect it to because it's built and run natively in any case) 18:54:12 only the result of running it is needed on a cross platform 18:57:19 did you install msys2 specifically for this? it might be worth removing it and starting over from scratch 18:58:57 yeah i rebuilt again with just make and it's failing. i'll try starting msys2 again. 18:59:01 or trying https://www.mingw-w64.org/getting-started/msys2-llvm/ perhaps 19:02:13 actually I'm not sure that link is the one I'm thinking of, the clang/llvm-toolchain one isn't fully binary compatible with standard msys2 because C++ stuff uses libc++ instead of libstdc++ but that one just installs the toolchain over standard msys2, hm 19:43:31 now my terminal hangs at `CC pcre_fullinfo.o` after make. not sure if this step is just extra slow :/ 19:51:21 dunno, I've never used the contribs, even when I had a Mac I installed them from MacPorts instead 21:23:37 <09h​ellmonk> Does anyone else want to test #4427 before I merge it? Seemed to work quite well in my testing and is a really good change imo. 21:41:26 Bowmetheus (L27 MiFi) ASSERT(it != env.trap.end()) in 'traps.cc' at line 898 failed. (Pan) 21:44:19 <04d​racoomega> It looks far less boilery platey than the previous PR on this issue (which I certainly agreed with the intent of, if not necessarily the implementation). I had actually sort of wanted to return to it myself and try implementing it in a slightly different way. (I'd actually wanted to put all the 'sufficient available space?' checks into a unified function, instead of inside every summon's own code, sort of like how we currently 21:44:20 handle "Is any target in range?" checks. Though doing the latter is less objectionable to me now that it's more concise. ...though I question having extended logic for space checks for spells that aren't even player spells.) I suppose another advantage of having 'is there any room?' checks outside the spell functions is that it might make it easy in future to have monsters consult it themselves in some cases, instead of this all being embedded just in 21:44:20 a player prompt. ...but I suppose even if I wanted to pull the checks out to somewhere else, there's probably no reason this would have to wait on me doing that (especially with all the rest on my plate) (I can do some testing tomorrow, if you'd like, but I'm past spent for the day) 21:46:36 <09h​ellmonk> alright, I'll leave it up for you 22:02:25 <09h​ellmonk> also, would it be worthwhile to post a development plan for 0.34 at some point on the github wiki? I can gather some stuff for that if it'd be useful. We haven't had one since 0.28 22:35:20 Unstable branch on crawl.develz.org updated to: 0.33-a0-1192-gbf953fe936 (34) 22:57:03 Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-1192-gbf953fe936 23:13:39 Unstable branch on cbro.berotato.org updated to: 0.33-a0-1192-gbf953fe936 (34) 23:55:04 Monster database of master branch on crawl.develz.org updated to: 0.33-a0-1192-gbf953fe936