03:07:16 Unstable branch on cbro.berotato.org updated to: 0.29-a0-451-gfbfa42419e (34) 03:43:36 Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-4699-g54b3c284c3 04:28:02 Hey I'm a novice programmer and I'm wondering if I wanted to use some of the map data like in the .des files for a project, would I pretty much have to make my entire project open source, or how does that kind of stuff usually work? 06:22:42 Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-4699-g54b3c284c3 06:54:32 jDawg: Replied on Discord 08:29:51 Unstable branch on crawl.akrasiac.org updated to: 0.29-a0-451-gfbfa424 (34) 09:16:10 jDawg: yes, the des files (and most of dcss that is not art) are covered under the GPL, so a project that made use of them would need to also be licensed under the gpl 09:31:19 <06a​dvil> @gammafunk do you know if there is any sort of N/A-like value that can be used in objstat outputs? 09:31:59 jDawg: just to be clear, reusing the syntax would be OK 09:33:20 <06a​dvil> well, you'd have to rewrite the parsing 09:33:29 <12e​bering> If you re-use our yacc/lex files/level compiler that’s all gpl 09:33:46 <06a​dvil> I mean, it's all ok, just comes with obligations 🙂 09:56:51 <10P​leasingFungus> idea: e is still unused, right? make it a generic equip/unequip button, listing all carried weapons, armour, & jewels (but keep w, W, T, etc around for experienced gamers) 10:07:25 <06a​dvil> I have a long term plan sort of like that 10:07:43 <06a​dvil> it's (as usual for me) more complicated than that though 10:08:33 <06a​dvil> I sort of think we need multiple control schemes that can be easily switched between 10:11:35 <06a​dvil> but I do have a specific plan for a menu kind of like that 10:14:04 <06a​dvil> it's true that e has a nice mnemonic for that regardless of how involved this goes 10:14:06 <10P​leasingFungus> oh, neat! 10:14:11 <10P​leasingFungus> let me know if i can help, then 🙂 10:14:34 <06a​dvil> basic idea: a generalized equip menu that has panes or modes corresponding to all the current equip menus 10:15:24 <06a​dvil> so CMD_WIELD or whatever just opens the specific pane, left or right switches between those (really left/right bumper(!)) 10:16:50 <06a​dvil> (and it has a mode that shows everything) 10:18:06 <10P​leasingFungus> into it 10:19:20 neat 10:20:21 <06a​dvil> lol who is placing a guardian mummy on D:10 10:21:14 a single key to use items would be neat too 10:21:17 I think c and p are unused? 10:21:25 though I have them custom-bound personally 10:21:30 <06a​dvil> p is used 10:21:34 p is uesd 10:21:38 c isn't though 10:21:56 too bad a/A are though, would have made a handy [a]pply button 10:22:17 yeah, my bad 10:28:12 just wanted to say, ty << 31 for showing the damage rating and formula for weapons 10:28:19 idk when it got added but I just noticed it now 10:30:43 <06a​dvil> @PleasingFungus random though re OoD placement, maybe too overcomplicated: I wonder if the monster's placement distribution should have an impact on this somehow? Sort of weird that this can (e.g. FALL vs RISE) have a big impact at the bottom of the range that is then ignored outside that range 10:31:17 <10P​leasingFungus> hm? 10:31:30 <10P​leasingFungus> don't quite understand 10:31:42 <10P​leasingFungus> doesn't distribution already apply? 10:31:48 <06a​dvil> hm, does it? 10:32:09 <06a​dvil> that's possible, I don't really understand this code 10:32:26 <06a​dvil> oh, maybe the ranges are just shifted and then the regular probably for the shifted ranges applies? 10:32:57 <06a​dvil> in which case, nm 10:33:28 <10P​leasingFungus> yes, i believe what happens is: 10:33:49 <10P​leasingFungus> 1. we pick a level (current or shifted by ood fuzz, vault glyph, etc) 10:34:17 <10P​leasingFungus> 2. we pick a monster for that level (which is based on frequencies by range) 10:37:46 <06a​dvil> ok interesting, so then the MONS_CYCLOPS example from yesterday is maybe a bit rarer than we were calculating, both before and after the fix, because its dist is SEMI 10:43:09 <10P​leasingFungus> well, no 10:44:12 <10P​leasingFungus> or, sort of 10:44:26 <10P​leasingFungus> we were calculating the odds of any 5 level ood being generated 10:44:37 <10P​leasingFungus> not cyclops specifically 10:46:15 <10P​leasingFungus> true that it was extra bad luck to get a monster at the bottom of its range 10:46:48 <10P​leasingFungus> but, once we fuzzed the level, that was equally likely with or without the bug 10:48:59 <06a​dvil> would something with RISE or FLAT be more likely than something with the same bottom range and FALL or SEMI? 10:50:03 <10P​leasingFungus> i think you have that half backwards? 10:50:53 <10P​leasingFungus> FALL and FLAT are more likely at the bottom of their respective ranges than RISE and SEMI are, right? 10:51:25 <06a​dvil> oh yeah, I was swapping rise and fall 10:51:28 <10P​leasingFungus> oh, depending how you define bottom 10:51:38 <10P​leasingFungus> i meant in the sense of earlier levels 10:51:49 <10P​leasingFungus> not physically deeper ones 10:53:03 <10P​leasingFungus> of course, a very high value SEMI entry might be more likely than a low value FLAT 11:14:28 03advil02 07* 0.29-a0-452-g763a1779be: feat: add some basic monster depth info to objstat 10(2 minutes ago, 1 file, 23+ 4-) 13https://github.com/crawl/crawl/commit/763a1779becc 11:17:05 <10P​leasingFungus> neat! 11:17:37 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-452-g763a1779be (34) 11:19:36 <06a​dvil> just trying to figure out how we might notice accidental changes to OoD chance 😄 11:19:44 <06a​dvil> gotta close that barn door 11:25:02 <10P​leasingFungus> “after the yaks have fled?” 11:34:16 <06a​dvil> lol 11:34:21 <06a​dvil> more like the death yaks 11:46:33 <08n​icolae> more like the shouldna been standin there yaks 12:48:52 <09g​ammafunk> I'd have to look into it to see how it currently handles things. I guess the answer to your question involves both what's a good representation on the C++ side and what's a good one in the spreadsheet output side, in terms of what sort of value sorts best in a numeric column 12:49:08 <09g​ammafunk> one thing I can say is that the counts are all stored as ints now, I believe 12:50:26 <09g​ammafunk> or rather all data is counted as integer values before any averaging and conversion to floating point values for conversion to string 14:11:32 <06a​dvil> Yeah, it doesn’t look like there was one, I just used 0 in the relevant cases in ^ commit 14:32:15 <09g​ammafunk> oh dangit 14:32:35 <09g​ammafunk> I'm almost done running a full objstat job for 0.29-a 14:32:38 <09g​ammafunk> guess I can just run it again 15:21:16 <06a​dvil> oops 15:21:34 <06a​dvil> I think I am going to add something explicit about ood as well 15:24:07 <06a​dvil> how many iters do you do for those runs? 15:26:16 <09g​ammafunk> 1000 iterations. It takes something like 12 hours on CDI to run all four jobs over 2 vcpus. The 3 rune jobs each seem to take about 5 hours, but the job I canceled a little while ago (after seeing your commit) seemed weirdly slow at the end 15:27:03 <09g​ammafunk> so I'm hedging my bets and saying it takes 12 hours total 15:27:36 <09g​ammafunk> @advil are you going to make another commit for objstat soonish? if so, let me know and I'll just wait to run the big job until that's pushed 15:27:45 <06a​dvil> yes, I think so 15:27:58 <09g​ammafunk> just thought it would be nice to see current data with the ood fix, seems like we both had the same idea! 15:28:14 <06a​dvil> haha yeah 15:28:22 <09g​ammafunk> ok, I'll wait and start after that's pushed then 15:28:59 <06a​dvil> the thing I added this morning is useful but on looking at some sample data I think it would be also good to just explicitly see ood info (which you can't get very directly from that) 15:29:31 <09g​ammafunk> one question I did have based on your commit message 15:29:33 <09g​ammafunk> ??ood 15:29:34 <04C​erebot> orb of destruction[1/9]: Iskenderun's Orb of Destruction - level 7 pure conjuration homing missile spell. Moves at speed 30 with inertia - if something casts this at you, sidestep, but watch out behind you! It does less damage at short range; at melee range it only does about half damage. Only in the book of Power and randart books. 15:29:38 <09g​ammafunk> ??ood[2 15:29:38 <04C​erebot> orb of destruction[2/9]: Monsters with shields might block the orbs. If two orbs collide they will explode with radius 1 and Gong-level noise (40). Orbs that haven't travelled more than one square dissipate instead of exploding. 15:29:39 PTOANNNG! 15:29:52 <09g​ammafunk> well, it's what does the fuzzing max it, isn't it 5 levels? 15:29:59 <06a​dvil> yes 15:30:24 <09g​ammafunk> you mentioned that a d:9 fire dragon is beyond ood since it starts generating normally at d:14 15:30:29 <09g​ammafunk> but wouldn't that make it not beyond ood? 15:30:36 <09g​ammafunk> maybe I misunderstood your meaning 15:30:42 <06a​dvil> I was looking at D:6 data iirc 15:30:57 <06a​dvil> oh, no that was from a range 15:31:12 <06a​dvil> yeah maybe I just did the math there wrong? 15:31:18 <09g​ammafunk> right, ok 15:31:21 <06a​dvil> or I was looking at D:8? 15:31:22 <06a​dvil> not sure 15:34:56 <09g​ammafunk> think having out of depth summaries will be useful, since we haven't had a very convenient way to look at that before 15:36:23 <06a​dvil> yeah, should be interesting to see, I'm also planning on backporting to 0.28 15:36:34 <09g​ammafunk> I guess it might be nice to have an out of branch count like you mention in the commit, but that can be derived already from the NumInBranch field, and those counts alone don't give you much data anyhow, since there's no good way to score such monsters 15:36:46 <09g​ammafunk> I suppose you could look at monster hd, but that's not going to be very reliable probably 15:37:26 <06a​dvil> one possibility might be to look at absdepth across all mon-pick tables 15:37:34 <06a​dvil> I could easily just flip the logic for NumInBranch 15:37:49 <06a​dvil> if it's more useful to see NumOutOfBranch 15:38:22 <09g​ammafunk> it does seem more useful as a standalone stat, probably 15:39:18 <09g​ammafunk> vaults are kind of a shadow factor for this kind of balancing, they are pretty difficult to control in a systematic way 15:40:09 <09g​ammafunk> but maybe we can look at some initial comparisons and get a sense of how big of a problem they are (as opposed to level spawns) 15:40:16 <06a​dvil> the out of branch counts are slightly inflated by things like uniques though 15:41:29 <09g​ammafunk> oh, if you wanted to look across monsters, you mean 15:41:44 <09g​ammafunk> it's probably fair to consider uniques in-branch? 15:41:59 <09g​ammafunk> since they do have something very much like mon pick data 15:42:35 <06a​dvil> yeah, possibly, though they need to be not considered in that count for that Num field's role as a denominator 15:58:10 advil and cot, thank you for clarifying. i might create similar vaults of my own with the same idea, but if i use any of the code i will go full GPL since I might want my source released anyway. Thank you. 15:59:33 And I wil definitely credit DCSS as inspiration! love this game :) 16:00:28 good luck with your game :) 16:00:54 is it a roguelike, or just a 2d game with some procgen? 16:00:59 I love the randomness, the gameplay philosophy, the penalties for dying, the shops... not too keen on the turn-based nature of time, it doesn't feel quite right 16:01:12 cot im a total novice, its nothing yet 16:01:20 ah 16:01:38 but i found godot engine 16:01:44 its got great tutorials 16:02:32 and open game art has some great tilesets and such 16:10:11 but yeah, 2D with a LOT of procgen, I played cubeworld and it works but the dev didn't understand principles of artistic design (like so many games unfortunately) 16:11:10 i might introduce platforming later for combat, but for now bird's eye view auto battles are fine 16:11:37 procgen is a way for a small team to make a lot of content 16:13:14 and vaults are genius, i swear 16:37:48 <06a​dvil> lol at the D:6 fire crab 16:38:00 <06a​dvil> also who is placing deep elf pyromancers and whatnot there? 16:41:52 <10P​leasingFungus> guppyfry_early_elven_hall 16:42:54 <06a​dvil> quick response! 16:43:23 <06a​dvil> zyphyrmancers are 11 levels OOD according to the population tables at that point 😄 16:44:14 <10P​leasingFungus> i had to scour through a bunch of des files, oughta script it somehow 16:44:19 <10P​leasingFungus> des parser. des explorer 16:52:24 https://github.com/geekosaur/vaultgrep :) 16:52:39 <09h​ellmonk> Could be a vault that originally placed like deep elf mage from 0.10 or something 16:53:51 <06a​dvil> yeah, seems plausible 16:54:04 <06a​dvil> in some of its range when it places pyromancers those are valid ood 16:54:17 <06a​dvil> anyways, this seems like it could be useful for finding stuff like this to tweak 16:55:52 oh I've gotten that vault D:6 before 16:56:26 surprisingly looks like I've never died to it 16:59:23 <06a​dvil> it's kind of a run away check 17:01:21 PleasingFungus: nitpicking, but mft's invisible/unearned medals tooltip has ".." instead of "..." for its ellipse. 17:02:17 03advil02 07* 0.29-a0-453-gf113fcf677: feat: add some monster OOD info to objstat 10(7 minutes ago, 3 files, 21+ 4-) 13https://github.com/crawl/crawl/commit/f113fcf677d9 17:19:01 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-453-gf113fcf677 (34) 17:34:30 <06a​dvil> silly question, why are objstat outputs named .txt and not .tsv 17:39:44 <09g​ammafunk> I guess I've been making tab delimited text for years and using txt 17:40:00 <09g​ammafunk> probably reasonable to change that, although personally I've never seen the tsv extension 17:40:21 <06a​dvil> ah, it's pretty common in "data science" stuff I think 17:40:39 <06a​dvil> well, more people use csv I expect 17:40:41 <09g​ammafunk> change sounds good to me, then! 17:40:43 <06a​dvil> the advantage is that a spreadsheet handles it automatically 17:40:51 <09g​ammafunk> yeah, and I never liked use of .csv when text in fact used tab 17:41:43 <09g​ammafunk> aside from autoimport, also makes it easy to grab all the relevant output without copying random source-related text files 17:41:54 <09g​ammafunk> if you use something like .tsv, I mean 17:42:12 <06a​dvil> ah, yeah 17:42:23 <06a​dvil> or more specifically, the extension causes an appropriate app to open it, and then it imports correctly 17:42:38 <09g​ammafunk> seems like a good change 17:43:44 FWIW you should always use tsv over csv given the option. The extension "tsv" is fairly common, about equal to .csv's usage. 17:43:50 If maybe a bit less 17:44:48 <06a​dvil> lol why are iron dragons and shadow dragons even in the D population 17:45:06 <06a​dvil> (placed by D:1 trapped monster stuff) 17:45:38 isn't there at least one arrival vault with an oof behind glass? 17:45:54 <06a​dvil> yeah, but it's not in the D population so doesn't show up in these stats 17:54:14 <06a​dvil> well, most ood for this run prize goes to a d:2 great orb of eyes 17:54:41 <06a​dvil> (haven't checked if it's trapped or in a runed vault or something,) 17:55:59 <06a​dvil> ah, probably trapped in hellmonk_ecumenical_altar_watching_god 18:00:25 <06a​dvil> too bad for these stats there's no decorative monster flag 18:02:03 <09g​ammafunk> iron and shadow are at the end of the OOD list 18:02:09 <09g​ammafunk> depth 27 18:02:29 <06a​dvil> yeah, I'm just wondering why they're in the list at all 18:02:41 <06a​dvil> I suspect they can't place in D except for those entry vaults 18:02:46 <06a​dvil> maybe for some derived undead reason? 18:03:04 <09g​ammafunk> They could simply place as 8 spawns by some vault, right? 18:03:10 <09g​ammafunk> maybe I misunderstand your question 18:03:51 <06a​dvil> I don't quite understand how 8s work but I think an 8 on D:15 cannot be depth 27? 18:04:10 <09g​ammafunk> yeah, it can, depth from 8 is 2xdepth + 2 I think? 18:04:23 <09g​ammafunk> obvious not below d:8 where there's a special rule 18:04:23 <06a​dvil> oh, huh 18:05:59 <09g​ammafunk> yeah, it's Depths * 2 + max(0, depth - 8, 4) 18:06:38 I don't like tabs as a delimiter because their size is variable visually 18:06:39 <09g​ammafunk> er, not quite, there's an appropriate min in there as well, but that's the idea 18:06:58 <09g​ammafunk> tdt is really meant to be opened in a spreadsheet jDawg 18:07:02 <09g​ammafunk> not in a normal text editor 18:07:12 then the delimiter doesn't matter 18:07:13 <09g​ammafunk> that's the use case of objstat, at least 18:08:01 <06a​dvil> in practice tsv has a mildly better chance of being readable in a text editor than csv ime 18:09:09 <09g​ammafunk> yeah, I guess it's fair to say that the reason why tdt is preferred is probably because you can at least make out that data in a text editor, whereas it's harder to read with commas 18:09:29 <09g​ammafunk> I'm sure various data science people prefer some other format altogether 18:09:36 <06a​dvil> it's mostly csv there 18:09:43 <09g​ammafunk> welp 18:10:20 <09g​ammafunk> would have no problem with an objstat option for using a different delimiter, but it probably should continue to default to tab 18:10:54 <06a​dvil> hm I guess objstat_Info is not a tsv 18:11:02 <06a​dvil> perhaps its a tsv with one column? 18:11:07 <06a​dvil> *it's 18:11:08 <09g​ammafunk> yeah, that sounds right 18:11:23 <09g​ammafunk> I don't think it would break any of my stuff to change format of that file 18:11:25 <09g​ammafunk> if it matters 18:11:41 <09g​ammafunk> the dream is to move objstat out of crawl and have crawl just make a bunch of levels that a libcrawl-linked python program could use to do the data analysis 18:12:19 <09g​ammafunk> so the dream is to have a libcrawl for outside programs and just have an easily parallizeable batch of crawl processes run to make level data, I guess 18:12:57 03advil02 07* 0.29-a0-454-g53bdb1016d: fix: tweak precision for NumOOD in objstat 10(7 minutes ago, 1 file, 3+ 1-) 13https://github.com/crawl/crawl/commit/53bdb1016df8 18:12:57 03advil02 07* 0.29-a0-455-g96ad54b138: feat: use .tsv for objstat outputs instead of .txt 10(85 seconds ago, 2 files, 2+ 1-) 13https://github.com/crawl/crawl/commit/96ad54b138d6 18:14:01 <06a​dvil> definitely would speed up objstat to parallelize it! 18:14:09 <06a​dvil> I assume there's no way of doing that right now? 18:14:17 <09g​ammafunk> well, I do use gnu parallell 18:14:52 <09g​ammafunk> but that's to run multiple jobs, it can only scale up to 4 cpus, since only 4 jobs 18:15:01 <09g​ammafunk> and in case of cdi I just have 2 18:15:34 <09g​ammafunk> no trivial way to make it parallel more so than that, that I'm aware of 18:16:08 <09g​ammafunk> I guess I do need to update my google script to look for tsv instead of txt, but that's an easy fix I think 18:17:38 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-455-g96ad54b138 (34) 18:19:19 <10P​leasingFungus> perryprog: oops, thanks! 18:22:41 jDawg delimiter does matter, because what if you have text with commas in it? Then you have to escape it. And as soon as you introduce escaping, it's a whole other thing. (There's also your "record separator", but that's a little lazy crazy since most people agree that you separate records with a newline that's escape as \n in text, but even then it can be inconsistent.) 18:23:32 The advantage of tsv over csv is that tsv is less likely to have inconsistent parsing, basically. See also http://thomasburette.com/blog/2014/05/25/so-you-want-to-write-your-own-CSV-code/ 18:23:51 Unstable branch on underhound.eu updated to: 0.29-a0-453-gf113fcf677 (34) 18:48:04 I mean, you still have to deal with the delimiter occuring in entries, it might be less common, but what good is making a bug less common rather than fixing it 18:49:29 there's about 5 different ways to "escape" delimiters in csvs… one of which is "ignore it" 18:57:25 jDawg, you don't, because it's two different standards. Tsv isn't "csv but tabs instead of commas". See https://en.wikipedia.org/wiki/Tab-separated_values versus https://en.wikipedia.org/wiki/Comma-separated_values for a better breakdown. 19:00:02 Oh, thank you, yeah, it is good to have a standardized approach to encoding. TSV is better for that alone 19:01:50 <09g​ammafunk> to be clear, the only reason why objstat is tab-separated is, back when I wrote it, I was used to working with that format for all data analysis (coming more from academic lab setting, where I don't think standards were pervasive at the time) 19:01:59 03advil02 07[stone_soup-0.28] * 0.28.0-50-gd1d40adc96: feat: add some basic monster depth info to objstat 10(8 hours ago, 1 file, 23+ 4-) 13https://github.com/crawl/crawl/commit/d1d40adc965b 19:01:59 03advil02 07[stone_soup-0.28] * 0.28.0-51-g5517f08384: feat: add some monster OOD info to objstat 10(2 hours ago, 3 files, 21+ 4-) 13https://github.com/crawl/crawl/commit/5517f0838499 19:01:59 03advil02 07[stone_soup-0.28] * 0.28.0-52-g7607e2ba6e: fix: tweak precision for NumOOD in objstat 10(56 minutes ago, 1 file, 3+ 1-) 13https://github.com/crawl/crawl/commit/7607e2ba6ef5 19:01:59 03advil02 07[stone_soup-0.28] * 0.28.0-53-g5a82cb258f: feat: use .tsv for objstat outputs instead of .txt 10(50 minutes ago, 2 files, 2+ 1-) 13https://github.com/crawl/crawl/commit/5a82cb258f5d 19:02:39 <09g​ammafunk> I was also the only person using objstat for a while and we weren't really relying on it in any way, but I'm not averse to improving the format now that it's used a bit more and by more people 19:03:39 <09g​ammafunk> big motivation of the time was I was working to remove item destruction, and we knew we had to rebalance consumable generation, since players would have more of those without item destruction, but we didn't have any very good ways to look at generation numbers aside from simply looking at the item weights 19:04:08 <09g​ammafunk> which is actually an alright way of getting an initial feel for how to change things, but vaults really complicate having a complete picture 19:04:13 <09g​ammafunk> so objstat is good for that 19:05:30 I think from the generation side tsv is definitely the easiest 19:05:57 well, comparing a 1k dungeon run from 0.28 to trunk, average ood level has noticeably decreased, but, it's still small values changing to smaller values, I don't know that my new fields have yet hit on a way that would make the recent bug really pop out (maybe I should just produce an average of ood level only for things that are ood) 19:06:54 Have yall considered using a tiling system that makes walls and such look different depending on the surrounding tiles like other walls as opposed to floors etc 19:08:55 there's a complete wang tiling implementation, it's only used in like crypt though 19:09:02 do you have a specific use case in mind? 19:09:23 <08w​ormsofcan> we technically use that for slime and ramparts 19:09:44 I was just wondering if it might be a graphics upgrade in general 19:09:44 <09g​ammafunk> yeah, that implementation is a sort of custom one though 19:09:53 <09g​ammafunk> as advil says we do have wang tiling for certain floors 19:09:59 <09g​ammafunk> but there's no support to extend it to walls 19:10:05 <09g​ammafunk> main issue is we don't have wall assets! 19:10:12 <09g​ammafunk> there are a lot of wall tiles you'd have to rework 19:10:25 <09g​ammafunk> I think it would require remaking rather a lot of wall tiles 19:10:41 <09g​ammafunk> for it to work, you need a perspective shift for the walls, right? 19:10:52 <09g​ammafunk> and I'm not sure how well that would work visually with our monster tiles, either 19:11:26 Hmmm wow that's complicated 19:14:44 yeah, I do think the biggest problem with stuff like this is that someone has to make the art assets 19:16:51 And I guess a lot of the public domaine tilesets are incompatible with stuff ingame already huh 19:17:01 <08w​ormsofcan> you wouldn't need to do a perspective shift but it would still require a bunch of newassets 19:46:27 nitpick: for the damage rating, do we need to use 'x' for multiplication, or can we use unicode's instead? 19:53:57 <10P​leasingFungus> great question 19:54:08 <10P​leasingFungus> i have no idea if all terminals that we care about support unicode 19:54:11 <10P​leasingFungus> i know ebering's does 🙂 19:55:09 rune already depends on it, no? 19:55:27 there's some unicode in use but there's an rc option that disables it 19:56:07 <12e​bering> we try to stick to CP437 for compatibility reasons for the default charset 19:56:17 <12e​bering> I think succeed even 19:56:20 !source viewchar.cc:10 19:56:21 https://github.com/crawl/crawl/blob/master/crawl-ref/source/viewchar.cc#L10 19:56:30 see comments there 19:57:29 though you'd have to do this in some other way for a unicode char used not in the dungeon display, I guess we must do something for the infinity symbol somewhere? 19:58:06 <10P​leasingFungus> oh yeah, i forgot we use that 19:58:09 <10P​leasingFungus> that's probably my fault, even 19:58:25 <12e​bering> I think we might just break a pure ascii environment in that case 19:59:56 !source output.cc:1806 19:59:56 https://github.com/crawl/crawl/blob/master/crawl-ref/source/output.cc#L1806 20:00:07 smh 20:16:52 <10P​leasingFungus> i suppose we’ve been breaking those hypothetical environments for some years without complaints 20:17:20 <10P​leasingFungus> so perhaps not a massive concern 20:37:11 <06a​dvil> well, I'm going to fix it 20:37:29 <06a​dvil> is there a reason to use codepoints as opposed to literal unicode chars? 20:37:48 <06a​dvil> I admit it looks fancier and all 20:39:10 probably the same reason the table existed: so it could be edited without unicode support 20:43:06 03advil02 07* 0.29-a0-456-g667b474bfb: fix: respect char_set=ascii in hud 10(2 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/667b474bfb69 20:43:16 <06a​dvil> er, not in the hud 20:43:21 <06a​dvil> in the char screen 20:43:22 <06a​dvil> oh well 20:43:41 <06a​dvil> that symbol is not in CP437 but I have to admit I'm not very bothered by that 20:44:00 <10P​leasingFungus> i would probably use ‘inf’ 20:46:43 <06a​dvil> I suppose there's space for it 20:46:44 <06a​dvil> feel free 20:46:58 oh wait, it is in cp437 20:47:27 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-456-g667b474bfb (34) 20:48:20 03advil02 07* 0.29-a0-457-g93c1fc3693: fix: remove wrong comment, use "inf" (PF) 10(30 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/93c1fc3693b3 20:48:37 was surprised to find that the code for that got the spacing right 21:02:32 Unstable branch on crawl.kelbi.org updated to: 0.29-a0-457-g93c1fc3693 (34) 21:09:34 I think we should rewrite in Swift so we can have emojis in our variable names 21:19:24 <10P​leasingFungus> agreed 22:22:39 New branch created: orcify_me_captain (4 commits) 13https://github.com/crawl/crawl/tree/orcify_me_captain 22:22:39 03hellmonk02 07[orcify_me_captain] * 0.29-a0-419-g53cbe7f85a: feat: send Beogh to a HR workshop 10(3 days ago, 3 files, 2+ 5-) 13https://github.com/crawl/crawl/commit/53cbe7f85ada 22:22:39 03hellmonk02 07[orcify_me_captain] * 0.29-a0-420-g9e68f830d5: feat: Beogh species renames 10(3 days ago, 1 file, 76+ 1-) 13https://github.com/crawl/crawl/commit/9e68f830d5c9 22:22:39 03hellmonk02 07[orcify_me_captain] * 0.29-a0-421-gf3c06e2309: feat: orcify Beoghite aptitudes 10(19 minutes ago, 3 files, 25+ 1-) 13https://github.com/crawl/crawl/commit/f3c06e23094f 22:22:39 03hellmonk02 07[orcify_me_captain] * 0.29-a0-422-gcbc1426263: checkwhite 10(35 seconds ago, 3 files, 4+ 4-) 13https://github.com/crawl/crawl/commit/cbc142626314 22:24:38 !!! 22:24:43 exciting 22:25:44 Orcmadillo! 22:27:07 <10P​leasingFungus> this branch name is a ref i do not get 22:30:50 <09h​ellmonk> https://www.youtube.com/watch?v=AvA0drvopjU 22:38:49 <09g​ammafunk> cpp case SP_VAMPIRE: return "Orcula"; 22:45:37 03PleasingFungus02 07[orcify_me_captain] * 0.29-a0-423-g0b761de0a5: Improve 10(10 seconds ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/0b761de0a531 22:49:41 <09h​ellmonk> lol 22:54:38 <08w​ormsofcan> orcmadillo instead of the superior armadillorc? 23:08:06 03PleasingFungus02 07[orcify_me_captain] * 0.29-a0-424-g4a65f1e6ca: Make level 6 orconians less buggy 10(6 seconds ago, 1 file, 2+ 0-) 13https://github.com/crawl/crawl/commit/4a65f1e6ca9f 23:11:53 <09g​ammafunk> we are getting the important stuff implemented first 23:13:22 <10P​leasingFungus> definitely