00:09:56 -!- Quipyowert2 is now known as Quipyowert 00:46:32 Seems like that approach could result in guess and check submitting pulls to find out what works 01:16:32 -!- amalloy_ is now known as amalloy 01:32:58 well, i achieved something, but using lowercase words ain't great: https://imgur.com/2rWU2Ot.png 01:35:37 ooh. "of the Pure slimy Ghosts" 👍 01:35:41 who ya gonna call? 01:38:41 Unstable branch on crawl.develz.org updated to: 0.25-a0-1119-g28b6fd0c3b (34) 01:56:47 much better, although "Pure" seems to come up a lot and i'm pretty sure that's not my fault 01:56:49 https://imgur.com/MWk6Ixy.png 02:00:24 lmao these are great 02:10:43 Windows builds of master branch on crawl.develz.org updated to: 0.25-a0-1119-g28b6fd0c3b 02:17:56 thx 02:18:55 is it good/cool/fun that 50% of weapons & armour are given letter-salad names? like "Fuiqhhabl" 02:19:38 i'd rather they got more stuff like "of the Fifty Burnt Souls" 02:32:18 https://github.com/RojjaCebolla/crawl/commit/31111dc53828711f524fabc9b74b937534a977cb 03:18:45 Unstable branch on crawl.beRotato.org updated to: 0.25-a0-1119-g28b6fd0 (34) 03:26:55 I feel like bladehands at 4% miscast shouldn't be turning me into a fungus or a tree making me unable to move, it's a bit harsh 03:28:07 what miscast level is that 03:28:17 white 03:28:28 grey is lowest, then white right? 03:30:11 Kramin: ask ebering about that, since he's the one behind that project; badforms at white/4% does seem excessive 03:30:23 that project being miscast reform 03:30:29 yeah ok 03:30:41 also 03:30:45 !streak 03:30:54 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3054-g1a545eca1e 03:30:59 gammafunk has 19 consecutive wins (GhAs, FoEE, KoSu, DDMo, DsIE, DECj, VSDK, TrAK, CeWn, VpEn, BaAE, OgTm, GnWr, DrFE, GrSk, SpVM, OpGl, HuAs, DgAM). 03:31:04 get good 03:31:08 lmao 05:17:02 New branch created: pull/1414 (3 commits) 13https://github.com/crawl/crawl/pull/1414 05:17:02 03RojjaCebolla02 07https://github.com/crawl/crawl/pull/1414 * 0.25-a0-1120-g6471c02: Add verbs to some randart names 10(3 hours ago, 2 files, 154+ 2-) 13https://github.com/crawl/crawl/commit/6471c02eda3b 05:17:02 03RojjaCebolla02 07https://github.com/crawl/crawl/pull/1414 * 0.25-a0-1121-g49b5c22: Add more randart name variety 10(2 hours ago, 2 files, 52+ 4-) 13https://github.com/crawl/crawl/commit/49b5c2234034 05:17:02 03RojjaCebolla02 07https://github.com/crawl/crawl/pull/1414 * 0.25-a0-1122-gc874474: Add adjectives to the randart name generator 10(88 minutes ago, 2 files, 152+ 0-) 13https://github.com/crawl/crawl/commit/c87447412b5d 06:28:19 -!- arnold is now known as Guest70702 06:29:43 Hey, can anyone tell what are the different accessibility changes for visually impaired users? I'm kind of curious. 07:28:56 -!- amalloy is now known as amalloy_ 07:44:43 03RojjaCebolla02 07https://github.com/crawl/crawl/pull/1414 * 0.25-a0-1123-g7d227ba: Further randart name tweaks 10(2 hours ago, 4 files, 25+ 9-) 13https://github.com/crawl/crawl/commit/7d227ba26ba5 08:17:18 Servitor has an outdated list of spells 13https://crawl.develz.org/mantis/view.php?id=12264 by Ge0ff 09:23:49 cebolla: letter salad names are a crawl classic 09:23:55 ??fuk de sak 09:23:56 fuk de sak ~ fuk da sac[1/27]: http://bit.ly/1bHFx6X 09:26:43 Kramin: yeah I need to do something about that. 10:37:01 yeah, not sure cat on keyboard randart names are really a problem per se, and actually that name code has a bunch of stuff that prioritizes plausible syllables structure and so on 10:37:41 though that commit doesn't tweak the probabilities *that* much 10:39:05 oh I see the "Future Goals" note 10:39:12 yeah, I wouldn't want to remove those :-) 12:56:44 since the new_annotate is almost finished, I might have a look at 'Force pickup with ;;', which looks to be a short project 12:56:44 and then either the Wield weapon information screen 12:57:05 or Shortcuts for R 12:57:23 but for those last 2 I'd like to go beyond the request on the wiki 12:58:04 for weapon information I think it'd be worth simplifying the damage formulas, so that we could display eg 12:59:21 the damage as D13 + 25% fire 13:01:38 For R, I'd like to make it a menu to Remove any piece of equipment, with shortcuts <> for rings, " for amulet, 1 for shield, 2 for helmet, 3 for cloak, 4 for gloves, 5 for boots, 6 for body, or something generally like that 13:02:13 numbers for cloak etc could be reused for rings for octopodes 13:02:25 Does either of these projects sound acceptable, and which would be more worth my time? 13:11:44 petercordia: I wouldn't link a desire to change the damage formula to the W display improvements 13:12:36 while crawl's core formulas are complicated they produce gameplay that we like 13:13:11 changing the fundamental damage formulas is perhaps a small code change, but it's a massive gameplay change 13:16:55 I would really like the wield wepon information screen 13:18:40 Unstable branch on crawl.akrasiac.org updated to: 0.25-a0-1119-g28b6fd0 (34) 13:32:27 yes, the problem with basically all of these terrible formulas in crawl is that they do correspond with good "feel" and attempting to simplify them is notoriously hard to get right 13:59:44 What is this idea based on, that the damage formula's have a good feel? I know any change feels uncomfortable, but.. I haven't seen any attempts to change the damage formulas. 14:00:26 I feel like the damage formula's should be simplified, not only so that players can understand them but also to enable future cool development 14:01:50 Hence making a wield weapon information screen before simplifying the formulas feels like a waste 14:02:23 with simplify formulas I could add information like 'dps vs 10AC + 15EV' 14:02:30 and I could allow the player to pick a monster from which to copy the stats 14:03:47 and resistances 14:06:09 I mean, I see that you've participated in some of the tavern discussions I was about to link to, so 14:06:35 can you remind me anyway? 14:06:39 I don't think anyone really believes that the current math is good, but coming up with a replacement is a nightmare 14:06:48 first one that came up was https://crawl.develz.org/tavern/viewtopic.php?t=26078&p=336163 14:06:56 where no one agrees about anything 14:10:16 re feel I think some of this may come from fairly ancient attempts to change things or changes that happened in forks 14:10:33 but maybe I'm wrong 14:10:59 is the high variance in the late game important? 14:11:44 It wouldn't be too hard to make a distribution where the average damage is controlled by a single number, and the variance is controlled by XL 14:12:22 I'm relatively good with maths & probability distributions 14:19:51 I have a vague memory that 4.1 alpha changed the damage formulas quite a bit 14:20:05 petercordia: the devteam has quite a few phd mathematicians on staff 14:20:09 the mathematics isn't the hard paart 14:20:12 the gamefeel is 14:20:22 yes, where we don't understand why a current distribution works, approximating it can be a viable strategy for simplification 14:20:47 if the end result is simpler 14:21:16 Doesn't new acq incentivise players to write down the options and bank the scroll until it's more clear what they want sometimes? 14:21:20 sometimes the end result is "this is a high precision integer polynomial computed to fit the old bizzare curve found by sampling" (looking at you spell failure) 14:21:33 haha yeah was just about to mention spell failure 14:21:37 !blame me 14:21:37 I pronounce me... Guilty! 14:21:57 although I think I got it from minmay 14:23:11 %git 99c12e3834af 14:23:11 07advil02 * 0.20-a0-1022-g99c12e3: Replace spell success breakpoints with a polynomial 10(3 years, 1 month ago, 2 files, 61+ 30-) 13https://github.com/crawl/crawl/commit/99c12e3834af 14:24:16 I still think this was a relative improvement, but hard to say whether it crosses the threshold of "good" 14:27:24 Pinkbeast it's certainly a thing that is possible now, I'm not sure if that is good 14:27:36 *if doing that is good 14:27:41 I don't think they have to write anything down, at least 14:28:02 If (say) I got a scroll on O:2 I'd definitely bank it and look at the shops 14:28:24 under old acquirement in that scenario it might also be better to wait and look at shops 14:28:35 in fact I bet it would still be best to wait before reading it 14:28:43 at least if you are aiming for diverse items 14:29:44 the real question is what if you get it on O:1 and it's a tough orc level 14:30:21 oh huh this gives me another acquirement idea that would solve some of the problems allowing canceling solves: have the scroll gate you to a mini-trove 14:30:45 I guess it would be usable as an escape item, maybe not ideal 14:31:04 I was thinking just of making you use it up before you do anything that takes time 14:31:13 Or maybe just before you move 14:31:20 I thought about that but I'm not sure how implementable that is 14:31:35 I mean, I'm sure it's possible in principle, but messing with time is usually tricky 14:32:04 could you use the moment where you leave the level as a break point? 14:32:38 I know it would feel a bit weird, but it would stop most exploits 14:33:20 yeah maybe 14:33:26 I guess you could also do something like a turn limit 14:35:12 hm in fact 0 turns would be the limiting case of implementing it as a turn limit 14:36:19 And surely as short as possible is right - a longer limit where you had to shave turns to make it to a shop as fast as possible would not spark joy 14:37:11 I'm still not really sure there is an exploit here 14:37:55 Pinkbeast: re: saving acquirement for later: I would like to offer a counterargument 14:38:00 !lg * !win cv>=0.24 s=place% 14:38:03 722205 games for * (!win cv>=0.24): 138189x D:1 (19.13%), 121047x D:2 (16.76%), 89753x D:3 (12.43%), 70783x D:4 (9.80%), 38845x D:5 (5.38%), 29371x D:6 (4.07%), 23944x D:7 (3.32%), 23591x D:$ (3.27%), 22870x D:8 (3.17%), 16212x Abyss:1 (2.24%), 15212x D:9 (2.11%), 12507x D:10 (1.73%), 10546x D:11 (1.46%), 6523x D:12 (0.90%), 6078x Abyss:2 (0.84%), 6036x Lair:1 (0.84%), 5838x Orc:2 (0.81%), 4834x L... 14:38:14 heh 14:38:29 yeah, if anything I suspect waiting is almost always a trap still 14:38:47 (though then it would make sense potentially to remove the option based on that) 14:39:18 I think an implementation for 0 turns might be easier than I thought, barbs gives a good model 14:39:45 If choosing between gold or a staff or energy, pre-orc, I'd wait 14:40:41 If it's a seductive fallacy, all the more reason to stop it. 14:41:25 Particularly barbs hangs off movement too, so you can stop in one place and swap jewellery or whatever to help you decide what item you want 14:49:38 !messages 14:49:38 (1/1) cebolla said (1d 16h 4m 49s ago): hey, what kinda rcfile shenanigans are you talking about? is there like, some magic lua that people use to make your spells & arrows prefer to target smiters and malmutaters when possible? 14:55:22 !tell cebolla most simply there's automagic_enable and autofight_throw, but using clua you can wrap a custom targeting/firing function macro as whatever keybind you'd like.. the complexity of such is more or less 'whatever you care to write'; lots of crawl gamestate is exposed through clua: see https://doc.dcss.io/ 14:55:23 Implojin: OK, I'll let cebolla know. 14:57:10 !tell cebolla i don't have a good example of "autoranged" to hand, someone might have written something like that if you search the tavern though; you can look at some examples of how to write targeting in qw.rc and autofight.lua / automagic.lua 14:57:11 Implojin: OK, I'll let cebolla know. 15:03:35 ebering was just attempting to regenerate the lua docs finally, and got the following error, any tips? `/Users/advil/repos/crawl/crawl-ref/source/dat/dlua/userbase.lua:368: ?: 'class' cannot have multiple values; {function,function}` 15:08:20 !tell cebolla target prioritization in clua usually goes something like: "iterate you.los() for monsters, compare exposed monster state to figure out what the most dangerous thing present is; a lot of user scripts rely on m:threat() for this, but you can check any of the exposed status in the monster module if you want: 15:08:21 https://doc.dcss.io/modules/monster.html 15:08:21 Implojin: OK, I'll let cebolla know. 15:11:45 <|amethyst> advil: probably you need a blank line (or something other than a -- comment anyway) before "--- Acquirement scroll hook." 15:12:48 thanks, that at least got me moved on from that error 15:12:55 now I'm getting internal ldoc errors 15:13:33 for Force pickup with ;; , I could try to make autopickup() remember it's been called before with a static variable. Is that a good approach? 15:14:02 Also, what would be a good statistic which is likely to change when the player does something? 15:17:53 by the way, i do appreciate your efforts in writing/maintaining those clua docs 15:18:07 probably not a lot of players use them, but it's a heck of a lot nicer than grepping the source 15:24:53 -!- amalloy_ is now known as amalloy 15:35:24 ebering (or anyone) ok new question, I'm hitting https://github.com/stevedonovan/LDoc/issues/300 when trying to run ldoc on some of our code, do you know a workaround? 15:35:24 I don't remember this before 15:35:24 maybe I was using a different version though 15:35:54 seems like ldoc is dormant 16:05:36 by the way, i think there's a potential ui improvement with regard to shop windows: it would be nice if, by default, any items that would be eligible for autopickup were automatically added to the shopping list 16:06:18 i implemented something like this in clua, but it requires overriding the default autopick and so my implementation is probably not great for merging 16:20:21 advil: I found ldoc googling around the lua community, it worked when I wrote those docs 16:23:13 advil: yeah looking at my system I have 1.4.6-2 installed 16:47:03 I don't think it's broken, I just seem to be hitting this one bug 16:47:21 I also have 1.4.6-2 16:47:35 any chance you could run make api and see if you get the same issue? 16:47:44 sure 16:47:50 I know I have run this before, though I did have to reinstall ldoc to try this time 16:50:08 1. I also had to reinstall. 2. I get the same error 17:32:10 are the new miscast mechanics documented anywhere? 17:36:20 what kind of documentation are you after? 17:39:38 what happens for each school and at what level the special effect begins to take place 17:40:14 there's some in spell descriptions 17:40:24 e.g. "Miscasting this spell will cause magic contamination and also deal up to 109 electric damage." 17:40:26 additional effects begin to have a nonzero chance at spell level 2 17:41:16 database question for new scoring 17:41:26 https://github.com/crawl/dcss_tourney/tree/new-scoring 17:41:31 there's the tree if you want to browse 17:41:51 the new functions I've added at the end of query.py feel to me like they should be moved to the schema as predefined views 17:41:52 oh wow lots of check marks 17:42:16 while the schema is a bit different writing csdc was good practice 17:42:31 but really I don't know what to do with databases at all 17:42:39 so I dunno what is and isn't a good practice &c. 17:42:55 if it's anything like the state of scoring when I got to it, there is very little good practice to be had 17:43:06 hahaha 17:43:57 one thing is that if you're concerned about efficiency, I'm not sure that super clean db design is needed for 2 weeks of tournament data 17:44:08 most of the terrible stuff in scoring took years to be an issue 17:44:13 indeed 17:44:20 I would say that at least some of those *could* be views 17:44:29 but it might be easier as we iterate to keep them in python 17:44:29 views aren't any sort of performance advantage though 17:45:18 they're transparent to the query optimizer and executer 17:45:22 I don't think views make sense here, because it's harder to modify a view than a dynamic query 17:45:35 yeah, that's what I had in mind 17:45:57 if you find the code is using the same query in multiple parts, maybe extract the query template to a string you can export 17:46:17 queries.TOP_GOD.format(username=....) 17:46:59 I did that a bit, splitting out the query objects from obtaining the rows 17:47:21 ebering do you have stuff in that project list that you would like someone else to work on? 17:47:57 happy to put my hand up too 17:48:00 anything on the mako templating side 17:48:19 the templates I'm making are bare bones so I can see that my ranking queries are working correctly 17:48:36 redesigning the player page and coming up with a good overview page are things that well 17:48:39 idk 17:49:01 they don't excite me and also I'm bad at web design 17:49:28 there currently aren't any todo entries for clan scoring 17:49:37 so every individual tag needs a corresponding clan tag 17:50:31 ok, these are some things to look into, though it'll take me a bit to get up to speed in actually running tournament scripts too 17:50:48 i'll take a look at some frontend design. I wrote a giant PR that was never merged a long time ago, to update the visuals 17:52:14 that sounds good to me alexjurkiewicz 17:52:14 do you know where that PR was? 17:52:14 I guess this repo may have been somewhere else in the past 17:52:34 it might be lost, i cant see a fork in my profile 17:53:18 by neccessity it's going to be a huge PR though, every frontend file will essentially be rewritten 17:53:51 I'll prototype one file and open a draft pr, then we can tweak the design until you're happy and then i'll convert remaining pages to match 17:54:12 well it was already gonna be a big rewrite because the relevant info has changed! 17:57:42 how would you feel about a design that does the templating on the client side (eg vue/react) rather than server side? 18:00:01 I'd prefer (with this timeframe) to use the (old but at least I understand it) tech that we've been using 18:00:06 I guess the trick is that we do basically want this to be doable in about 2 weeks 18:00:48 right 18:00:49 I'm not opposed to adding some more modern js on top of the server-side templating though 18:02:29 it's something I keep meaning to play with on the cao scoring page too, even floraline's extremely minimal extra js on https://crawl.kelbi.org/scoring/highscores.html makes it a lot more usable 18:03:52 yeah when I say "redesign" what I really mean is "write new mako templates and decide which tables/values belong where on the overview and player pages" 18:04:05 for the player pages I guess the points section can be simplified 18:04:14 and that's the only part that really needs work? 18:04:38 navigation could be better 18:05:59 I could imagine design improvements, too 18:06:38 sure 18:07:27 I'm aiming much closer to the bare minimum tweak-wise to make it happen 18:07:39 yeah, seems like the best thing to shoot for 18:07:43 to start with 18:08:04 the number of checkmarks on your project page does make me feel like we are in good shape though ;-) 18:08:51 getting the bare minimum to work would also let us be firmer about setting a tournament date 18:09:15 I hear we're competing with a nethack tournament 18:12:10 -!- Tiobot is now known as Guest18269 18:12:25 i'm sure both nethack players will be very put out ;) 18:13:07 lol 18:14:06 now now 18:24:08 Unstable branch on underhound.eu updated to: 0.25-a0-1119-g28b6fd0c3b (34) 18:52:25 ebering: what branch are you working on for dcss_tourney? 18:53:37 oh, new_scoring 19:28:05 ugh, the python mysql lib is so old. I can't even compile it on modern debian 19:32:09 ah, I can use a newer version as a drop-in replacement 19:50:24 -!- amalloy is now known as amalloy_ 21:51:51 ebering: is the tourney script running on a mysql 5.7 db? 22:06:49 alexjurkiewicz: it's running on CDO, version 5.5.62 22:10:02 alexjurkiewicz I think I've worked on them locally with mysql 5.7 22:16:18 CDO is the ancient one without docker, right? 22:16:25 no chance of 57 i guess 22:17:29 docker? are any of our servers running on docker? 22:17:41 oh, you mean is docker available 22:17:47 I'm pretty sure cdo doesn't have docker but I doubt it is that ancient 22:18:48 n.apkin is not super active in crawl but he does all sorts of other stuff with it 22:19:39 CDO is Debian 8 22:20:20 yeah, napkin is willig to sit down and do upgrade stuff as he has time; we got CDO builds and monster working a week ago, but his time is limited 22:20:20 not even out of LTS! 22:20:27 haha, really? 22:20:31 that's quite an LTS 22:20:32 Jun 30, 2020 22:20:36 wow 22:21:03 advil: did you add the tournament script function to support mysql5.7 btw? 22:21:12 guess I can check the git log! 22:21:13 I don't think I did anything? 22:21:31 no recently, but I just see that function in the scripts 22:21:43 support_mysql57() 22:21:52 maybe CSZO had mysql5.7 22:22:02 CDO is only 5.5 22:22:10 yeah, I don't think that was me, doesn't look like I have committed to master in that repo 22:22:39 gotcha 22:24:11 https://github.com/crawl/dcss_tourney/commit/fcd34bf53d1723f5333232f6f2180024704814c0 22:24:17 haha it was alexjurkiewicz 22:25:07 nice 22:26:01 wow 22:26:23 i was looking at that function just before thinking "what idiot wrote this completely mysterious thing with no documentation" 22:26:24 past alexjurkiewicz, meet present alexjurkiewicz 22:26:38 you must together save alexjurkiewicz of the future 22:26:52 lol 22:27:25 i wonder how much effort would be saved unifying all the services that read and store game data. tourney scoring, normal scoring, sequell, other 22:29:21 well, those things you mention are maintained across three different servers with different owners/admins 22:30:38 for tourney I think it's kind of useful that it's a small one-off database, regular scoring couldn't really handle all the milestone stuff that tourney does without a big rewrite 22:31:26 even sequell is losing it's capacity to really do milestones it seems (even setting aside disk issues, milestone queries have been pretty unreliable for some time) 22:31:57 I disabled all milestone reading on CAO scoring because it didn't scale well at all relative to the like 3 unchanging stats that were being collected 22:32:30 yeah, when operating well, you need to heavily limit versions/games queried to get a milestone result before timeout 22:32:37 of course even then it's usually pretty slow 23:08:22 I think unifying these services would be good, but not necessarily most beneficial 23:09:34 the bottleneck here is the different owners/admins and the fact that they don't have much free time 23:10:07 and that the servers all have other stuff running on them, which hugely magnifies the risk of changes 23:15:54 if we could move smaller services with smaller storage requirements one at a time to linode or something, and give all devs SSH access, I think it would help 23:16:28 but I don't have a clear picture of what services are running where (which would be a good thing to add to the server overview wiki page) 23:21:14 yeah 23:22:58 aidanh: there's also issues of maintainership/authorship. I'm not sure if you're aware who greensnark, the main author and maintainer of Sequell is (see http://crawl.develz.org/wordpress/the-dawn-of-stone-soup) 23:23:44 so he basically built most of sequell and footv and has run the server; it's not the case that we'd want to sort of "take" ownership of sequell server without getting a clear OK from him 23:24:34 it's possible that he'd be ok with that, but for devs/admins who have built up services themselves, it's not necessarilly fair that we'd dictate handing over operation to a central server 23:24:59 not saying a general project to do something like that isn't worthwhile, but it's something to generally keep in mind 23:31:41 oh yeah, for sure 23:33:29 I'm not suggesting that we wrest control away from anyone; if we did migrate services, snark should still have access of course 23:57:08 imo it would be good is to do it in a modular way, alex and I have had the idea of a central pubsub server for milestones/logs/morgues for a while 23:58:50 nothing fancy, just store the raw data and serve it to other services like a tourney script/online scoreboard/more queryable sequell-like database