00:00:14 Brannock: I feel like mantis gets used? maybe that's just because I use it. 00:00:22 it does 00:00:26 but I mean stuff like this 00:00:41 well 00:00:44 !source process.txt 00:00:46 https://github.com/crawl/crawl/blob/master/crawl-ref/docs/develop/process.txt 00:00:49 Just skim through that 00:00:53 ^status 00:00:53 31 Crawlers. CBRO disk usage=77% (135GB) | RAM usage=20% (4GB)| uptime/CPU= 00:00:52 up 26 days, 4:39, 2 users, load average: 1.24, 1.48, 1.54 (4 Cores) http://status.berotato.org 00:01:00 much better 00:01:10 activity on mantis for like implementables and such? 00:01:13 Most of it pertains to the 2008-2012 days I think 00:01:17 at a guess 00:01:18 certainly for bug reports it gets used 00:01:41 I mean that Mantis doesn't get many patches these days, nor does the dev team frequently post implementables on Mantis for other coders to take up 00:01:43 yeah, that is pretty out of date 00:01:47 It's primarily bug reports these days 00:02:03 oh yeah, patch submissions 00:02:10 The mailing lists also aren't used nearly as much, see: https://sourceforge.net/p/crawl-ref/mailman/crawl-ref-discuss/ 00:02:14 I remember reading this when I was added 00:02:32 I'm trying to clean this up but the whole process has changed so significantlY I don't know whether it's better to just rewrite it 00:02:51 ah 00:02:57 take a gander at https://crawl.develz.org/mantis/view.php?id=7188 00:03:05 the good old days of mantis patch submissions! 00:03:09 nice 00:03:23 hats of jumping?! 00:03:28 there's a ton more but they went directly to a now-deleted branch 00:03:36 hats of jumping were the new hotness for a while 00:03:46 ahh, back in the days when gammafunk was fresh and young ;) 00:03:47 AND Dragoons were cool 00:04:11 was that before you were a dev gammafunk? 00:04:41 seems like it, yeah 00:05:05 hmm, scarf of jumping?!? 00:05:11 no. 00:05:16 scarf of wall jumping 00:05:29 it's amazing how many jumping-related antics we've had 00:06:04 jump attack, Ru jump, shadow step, wall jump 00:06:04 gammafunk: how do you feel about monster/item count in the_grid 00:06:26 gammafunk: don't forget when Crawl jumped the shark by removing mountain dwarves 00:06:32 are you trying to pre-empt my use of non-uniform distributions? 00:07:00 I haven't looked, but I imagine that vault modifies it in some way based on the vault choice? 00:07:03 !vault the_grid 00:07:13 Can't find the_grid. 00:08:27 hrm 00:08:31 gammafunk: not talking about distribution, just the overall amount 00:08:34 !source the_grid.des 00:08:34 was this vault inspired by V:5? 00:08:35 https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/variable/the_grid.des 00:08:47 gammafunk: no more than it was inspired by any other vault that uses subvaults 00:09:00 yeah I just saw the "prize" terminology 00:09:02 gammafunk: as in, do you think it has too many monsters/items 00:09:20 well tbh I'd really have to sit down and look at it 00:09:29 or I could make up some BS based on my one playthrough 00:09:40 it was certainly quite a thing on D:11 00:10:00 gammafunk: I was asking for the anecdotal evidence of one playthrough, yes, since you're the one person who I know has encountered the vault and happens to be here 00:10:33 are you asking purely in terms of numbers here? it doesn't have too many monsters numerically speaking 00:10:57 its going to continue having 8s and so is the sequel 00:14:13 I personally like to just do things like make the vault place and run objstat on that level, look at the results 00:14:22 but the one thing about it is that it's so large, the variant I got 00:14:53 the_grid sequel development statistics so far: 4 accidental swastikas, 1 accidental exact duplicate of existing subvault, 1 accidental pepsi logo 00:15:01 sort of sat awkwardly in the middle of the level, and felt like it should have just had reasonable encompass containing it 00:15:13 but I doubt that's a thing you care about either 00:16:17 "// debugging output -- oops, I probably block it above ... dang! {dlb}" 00:16:29 I guess the variant I got was just a grid of four squares, so there's not a lot to do with that 00:16:56 new encompass: slaughtergrids 00:16:59 the _grid_plus and _grid_plus_closed have even higher dimensions 00:17:36 what's that vim command to count chars/lines 00:18:36 who are our administrators? I'm guessing amethyst, MPA, and dpeg? 00:18:45 admins for what? 00:18:54 people who can grant commit access 00:18:59 all of us can 00:19:08 oh, right, the GitHub move probably affected that, didn't it? 00:19:11 yeah 00:19:15 man this thing is ancient 00:19:22 once you're an org member, you can add others 00:19:34 and various other things various other people have various levels of access 00:19:44 I know Lasty has a bunch of tavern access 00:19:47 it's highly centralized! 00:19:54 technically I think I can't but that's more because github was being confusing 00:19:54 well Lasty's a mod, like me 00:19:58 "Administrators will make final design decisions where necessary, or where there is confusion or debate." 00:20:00 I'm only a "Member" 00:20:02 but doesn't really have any other access 00:20:19 archaeo and njvack currently have more tavern control than either of us, since they're tavern admins 00:20:24 but I asked kin about making MPA one 00:22:08 There can be no formal rules for solving such problems, because we want 00:22:08 neither design stagnation (i.e. adding only with mutual agreement) nor 00:22:08 dictatorship (because other people do have good ideas sometimes, even if it's 00:22:08 hard to believe). 00:22:49 if you have any general design guru wisdom, probably just best to put it in the orientation I made 00:23:11 the new dev guideline? 00:23:13 yeah 00:23:22 Yeah I'm getting to that one after I eviscerate process.txt 00:23:29 Will submit my changes as a branch, by the way, so you can review them 00:23:30 I made sections specifically about that, some others suggested modifications 00:23:39 sure, or just commit them even 00:23:49 the new dev one was in much better shape than process.txt 00:23:57 it's not meant to be a doc that one person controls 00:24:20 I had totally forgotten that process.txt existed 00:24:45 one of the SA goons was talking about adding to it (over the WJC mess, "what to expect when your contribution is merged") and then I looked at it and whoa it's nearly five years old 00:24:50 the info in it, I mean 00:24:52 it's probably older! 00:24:58 hrm, it could be nice, though 00:25:09 I guess to have something about "the process" for non-devs 00:25:17 for PRs and such 00:25:19 possibly all the links in the new dev doc should go to github? right now they all go to the develz.org mirror 00:25:28 yes, that was a thing I wanted to fix 00:25:31 someone was talking about updating that a while ago 00:25:34 possibly even you/gammafunk! 00:25:52 I fixed the links, but I made them all go to the mirror 00:26:03 since some of them already did 00:26:51 it might be possible to split out the "design guidelines" portion of the new dev doc into its own thing 00:27:08 and refer to that in the new dev doc, and refer PR authors to it 00:27:34 maybe some bits are specifically for dev team members wrt design, but I sort of doubt there'd be anything like that 00:28:13 I mean the "design philosophy" stuff, but then again some of that talks about "when to merge" 00:29:37 who is edlothiol? 00:31:18 new_dev_checklist still mentions the CSZO server......... 00:48:33 edlothiol is a dev who did a ton of work on tiles/webtiles 00:48:39 maybe mostly webtiles, I forget 00:48:46 hasn't been active for some time though 01:10:12 this file is a pain in the ass, no wonder it's been neglected 01:14:07 process.txt? 01:14:10 yes 01:14:24 !source process.txt 01:14:24 https://github.com/crawl/crawl/blob/master/crawl-ref/docs/develop/process.txt 01:14:33 I'm basically rewriting it entirely. I'm not sure what to do with the talk about the CDO tracker and the devteam structure sections 01:14:51 This seems like useful information but the rewrite I'm doing doesn't really have room for that 01:15:23 well, again some of that is covered in the orientation 01:15:25 https://pastebin.com/Sn01VW2h This is what I have so far 01:17:17 wip, obviously 01:17:30 I don't think anything after the section on CDO/wiki needs preserving 01:17:36 that's covered in the other doc 01:17:46 and the perspective for contributors you're already covering above 01:18:00 yeah 01:18:05 mentioning bug reports and mantis is good probably 01:18:09 yes 01:18:37 We could mention the doku wiki as a tool for brainstorming, I guess 01:18:50 I don't think granting commit rights is covered anywhere else 01:18:56 though I'm not sure that needs to be explicitly laid out in a file 01:19:05 no there's a new dev checklist 01:19:17 which certainly will mention granting commit access 01:19:21 oh, I mean, for the whole "should we add this guy?" email stuff 01:19:41 lines 193+ in the pastebin 01:20:02 though I guess that culture is pretty well established by now 01:20:05 uh, I'm not sure that it needs to go in a doc per se, but this doc wouldn't be the one imo 01:20:07 probably wasn't back when this file was originally written 01:20:08 yeah 01:20:09 Unstable branch on crawl.develz.org updated to: 0.20-a0-927-g1ae2c61 (34) 01:20:10 this is for outside contributors 01:46:39 -!- Marvin is now known as Guest36380 01:50:38 gammafunk: i just got an idea 01:51:09 gammafunk: custom crawl version that auto-generates vaults from twitch chat 01:57:59 Windows builds of master branch on crawl.develz.org updated to: 0.20-a0-927-g1ae2c61 02:02:58 03Brannock02 07* 0.20-a0-928-gaefa9c1: Totally rewrite and rename process.txt 10(9 minutes ago, 2 files, 123+ 152-) 13https://github.com/crawl/crawl/commit/aefa9c1719e4 02:02:58 03Brannock02 07* 0.20-a0-929-g5127991: Minor cleanups to new_dev_checklist.md 10(9 minutes ago, 1 file, 13+ 12-) 13https://github.com/crawl/crawl/commit/5127991fafd7 02:02:58 03Brannock02 07* 0.20-a0-930-gbddf555: Relocate migrants from process.txt to new_dev_guidelines.md 10(3 minutes ago, 1 file, 17+ 3-) 13https://github.com/crawl/crawl/commit/bddf5556d2b6 02:07:31 minmay: so what would trumptrain input result in, vault-wise 02:07:45 or mps input, for that matter 02:18:00 gammafunk: whenever a level is generated, run a level generator on a small area stored in memory, say 30x30, and make a tunnel leading out of it. 02:18:35 gammafunk: twitch chat lines are then parsed for monster and item names which are added in random places (with limits so we don't exceed crawl's monster limit etc) 02:19:21 gammafunk: the most frequent character out of 'b', 'c', 'v', 'x', 'g', and '8' determines what this vault's walls are made of 02:19:34 this sounds like an excellent april fool's joke 02:19:35 gammafunk: when you enter the next level, that vault is placed on it as a floating vault, and the process repeats 02:19:52 since it contains all the stairs, you're forced to play at least some of it 02:21:16 I guess you could also just use this method to generate an entire level, and then also parse twitch chat lines for vault names and try to place the first X mentioned vaults 02:21:46 actually nah it shouldn't be first X it should be a random selection of all the vaults mentioned (with duplicate checking so someone can't just spam the same vault over and over) 02:34:32 people are really bad at knowing vault names though 02:34:47 unless cheibrodos happens to be in my chat or a few other players who even know what they're called 02:36:35 I was thinking like, use each word as a regex and pick a random vault matching it if any 02:37:27 so for example you'll get evilmike_hourglass a lot because 'ass' is a substring of it 02:37:54 and you'll get a the_grid vault in every game because 'the' is a substring of it 02:38:22 also they can just spam "minmay" if they want a lot of minmay vaults for example 02:39:41 just choosing existing vaults would be a lot simpler to implement 02:40:03 compared to generating an entirely new vault, I mean 02:40:33 but I'm not sure choosing existing vaults would be super interesting 02:43:26 ok, better idea: twitch chat gets to mess with pop_entries 02:44:13 like you can type "black mamba+" to increase the weight of black mambas in all branches (but not change their depth) 02:45:06 so a coordinated effort could give you a lair that's almost exclusively oklob plants, for example 02:45:37 you'd want a limit on how often one user can do this of course 02:47:02 since they still couldn't place monsters out of depth it would be hard to make the game totally unplayable, although you'll probably want to do the first few dungeon levels as fast as you can before the orc priest weight gets too high 02:52:14 can't wait for exclusively OoF zot 02:52:28 finally, a good reason for FDA 02:53:58 Monster database of master branch on crawl.develz.org updated to: 0.20-a0-930-gbddf555 03:00:04 minmay: yeah I think you'd have the essential fun if you could just generate monsters from chat, probably giving them vault names based on twitch chat names 03:00:25 and maybe likewise do for items (and use twitch chat names for artefacts) 03:01:01 the vault names bit would make things kind of ugly, I guess, but you could limit the length 03:10:15 Unstable branch on crawl.jorgrun.rocks updated to: 0.20-a0-930-gbddf555 (34) 03:12:41 Unstable branch on crawl.beRotato.org updated to: 0.20-a0-930-gbddf555 (34) 03:52:40 -!- amalloy_ is now known as amalloy 05:49:21 -!- amalloy is now known as amalloy_ 05:53:52 -!- amalloy_ is now known as amalloy 07:04:02 -!- amalloy is now known as amalloy_ 07:27:48 -!- zkyp_ is now known as zkyp 08:20:10 New branch created: pull/503 (2 commits) 13https://github.com/crawl/crawl/pull/503 08:20:10 03sooheon02 07https://github.com/crawl/crawl/pull/503 * 0.20-a0-931-g1c201bd: Remove Monks' starting weapon choice 10(38 minutes ago, 1 file, 2+ 3-) 13https://github.com/crawl/crawl/commit/1c201bd77922 08:20:10 03sooheon02 07https://github.com/crawl/crawl/pull/503 * 0.20-a0-932-g400500b: Move Monks to Adventurer classification 10(7 minutes ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/400500b6a60f 08:45:36 Any devs around? 08:52:50 SteelNeuron: I am for 8 min, then I have to run out for a bit 08:52:54 sure 08:53:02 I think I have a possible implementation for wall jump 08:53:06 that could be good 08:53:42 let's leave accidental wall jumps aside for a second, since the solution for that is sort of unrelated. But here's what I'm proposing: 08:54:04 wall jump is free and usable everywhere again, but it takes 20 AUTs to do (for a 10 aut character) so it's essentially speed neutral with walking 08:54:22 However, enemies around the jumping and landing spot always get distracted for 10 AUTs (or half of whatever you took) 08:55:12 we may then roll the distract chance separately from that. End result however is that monsters in your vicinity react to your action in a similar way to old wall jump (if you wall jump away, even if you took 20 AUTs, they only get 10 AUTs, so they would only make it halfway to you) 08:55:29 but monsters that are a bit further catch up since they travel at your same speed 08:55:46 it also does not help you outrun ranged monsters 08:56:24 I think this is ideal, since it still allows you to pull off local positioning tricks, without harming other aspects of gameplay or requiring a limitation in how and when you use it, or a price 08:56:48 Ah, interesting. I think that resolves the objections I feel strongly (no use as travel, strong consistency) and also adds the interesting nuance of the difference of the effect on hit versus non-hit monsters. 08:57:06 It also lets wall jump be a decent engage power against a ranged monster, say 08:57:30 This seems like a good option to me. 08:57:41 I'd like to hear whether this resolves any of the issues MPA has with it 08:57:44 also gammafunk 08:57:49 sure 08:58:17 the (unrelated) solution for accidental wall jumps would be to require two bumps against the wall (the first is free, and just MPR's "You put your feet on the wall", and the second executes) 08:58:56 I think this has value over using the "a" menu, especially if we can get rid of the piety cost idea. One less keypress, no menu navigation. 08:59:30 I'd personally argue you get used to not miskeying near walls fast enough that this wouldn't be necessary, but I'm open 09:03:50 give d:1 slime walls, people will quickly learn not to miskey near walls :p 09:20:58 I wish I understood a bit better what all these description levels were used for 09:21:39 like is basename supposed to include missile brands? 09:21:56 like qualname it only includes them if they're prefixes (so poison, but not confusion) 09:22:34 but not in terse mode 09:23:01 I guess the better question is, what if any save compat issues could I run into by messing with this 09:23:21 hopefully none? 10:08:43 aha 10:08:43 %git dc39ea48bec184b30a6d3 10:08:43 07|amethyst02 * 0.11-a0-1818-gdc39ea4: Omit enchantment from missile base/qual/db name. 10(4 years, 10 months ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/dc39ea48bec1 10:08:56 but why isn't this just about DESC_DBNAME? 10:11:11 I think that fix shouldn't have changed for QUALNAME, at least relative to the current description code 10:24:11 03SteelNeuron02 07https://github.com/crawl/crawl/pull/502 * 0.20-a0-929-g04395ec: Rework Wall Jump 10(in the future, 3 files, 50+ 33-) 13https://github.com/crawl/crawl/commit/04395ecd0225 10:27:07 03elliptic02 07* 0.20-a0-931-gb4023b8: GSC base delay: 1.9 -> 1.8. 10(10 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/b4023b899515 10:32:21 03SteelNeuron02 07https://github.com/crawl/crawl/pull/502 * 0.20-a0-929-ga730685: Rework Wall Jump 10(in the future, 3 files, 50+ 33-) 13https://github.com/crawl/crawl/commit/a730685b0bc7 10:34:21 03advil02 07* 0.20-a0-932-g5a89274: Regularize qualname and basename for missile brands 10(12 minutes ago, 1 file, 4+ 4-) 13https://github.com/crawl/crawl/commit/5a8927434310 10:34:21 03advil02 07* 0.20-a0-933-ga5223d7: Improve display for collapsed missile searches 10(3 minutes ago, 2 files, 13+ 5-) 13https://github.com/crawl/crawl/commit/a5223d7ca325 10:52:06 03advil02 07* 0.20-a0-934-geb9d9bd: Remove use of spam threshold to silently exclude other levels 10(3 minutes ago, 1 file, 0+ 2-) 13https://github.com/crawl/crawl/commit/eb9d9bd04685 11:09:27 Unstable branch on crawl.jorgrun.rocks updated to: 0.20-a0-934-geb9d9bd (34) 11:13:54 SteelNeuron, you've fixed a lot of what I found problematic with wall jump 11:14:54 Brannock: Good to hear! I'm finding this one more promising than piety costs tbh 11:15:06 Yes, same 11:15:09 I did not like piety costs on jumping 11:15:46 SteelNeuron, to be clear, wall jumping now means that distant enemies get the chance to double-attack the jumper? 11:15:48 since they are not distracted 11:15:57 ranged ones do 11:16:18 melee ones wouldn't, because they get that energy decrease whether they are adjacent to your origin or landing tile 11:16:31 I think that's acceptable. After all, a complex maneuver like that is going to be, notionally, easy to line up a shot against 11:16:57 SteelNeuron: wait, a melee monster 3 tiles away from the landing point of your wall jump would get its energy decreased? 11:17:03 It also encourages players to use the cloud ability 11:17:10 no Lasty_ , only adjacent ones 11:17:13 oh whew 11:17:18 good 11:17:28 That's what I'd expect :D 11:18:30 it doesn't cause any issues with melee damage I think 11:18:38 the number of attacks is calculated over 10 auts normally 11:18:45 so you don't get two turns of hitting enemies 11:19:23 and the retaliation you get when landing among enemies is also one turn worth of damage. 11:19:50 (whenever I say 10 auts, I mean player_move_speed() obviously, whatever that is) 11:20:58 SteelNeuron, have you tested what happens with Slow/Chei/Naga? 11:21:01 or all of the above combined 11:21:12 well, I guess chei doesn't apply here 11:21:17 I need my morning breakfast 11:21:36 haha 11:21:41 Breakfast of Cheimpions 11:24:34 -!- amalloy_ is now known as amalloy 11:25:24 Brannock: about Naga, yep, tested it, works sort of as you'd expect 11:25:38 wall jump away from an adjacent monster, it loses your move speed worth of energy 11:25:46 so it gets one... normal naga turn of time to get back to you 11:26:03 same as old walljump 11:26:20 distant enemies, they would catch up exactly as much as if you had just walked away the same number of tiles 11:26:29 good 11:26:31 consistency is important 11:26:31 Does it matter if you execute this move from or into shallow water? 11:28:09 hm 11:28:23 I wonder if that would need some special handling 11:28:29 it doesn't feel like it would 11:28:35 but I'd need to double check 11:35:42 anyway SteelNeuron you've been doing a really good job on this, you should feel encouraged 11:40:28 thanks :) I've come to realize it's going to be impossible to please everyone, so I'm learning to make more compromises 11:41:26 I still feel for the people who say that martial attacks should be stronger than tab, because that's what they found fun, because there's a part of me that agrees 11:41:32 but nah, this is healthier 11:47:29 -!- amalloy is now known as amalloy_ 12:01:03 -!- Marvin is now known as Guest84546 12:01:08 elliptic, I probably would have favored increasing GSC damage over reducing its base delay. The commit you referenced was trying to differentiate the two weapons. I suppose now with Trolls/Ogres both having bad/'bad' M&F getting to 22 skill is still a significantly higher investment than 18 for GC. 12:01:26 but I guess in general Crawl weapons are already poorly differentiated anyway 12:02:28 Brannock: I considered that but GSC damage is already so high 12:02:33 it is, yeah 12:03:56 I could also see an argument for merging the two weapons 12:04:57 this change seemed the smallest change to keep the two weapons somewhat competitive with each other (though I suspect most chars should still not get GSC) 12:26:19 oh cool, lliptic caught dcss up to yiufcrawl (: 12:39:58 -!- Mert is now known as Roach_Mortician 12:53:49 -!- amalloy_ is now known as amalloy 12:55:31 -!- MarvinPA_ is now known as MarvinPA 13:10:40 Unstable branch on crawl.akrasiac.org updated to: 0.20-a0-934-geb9d9bd (34) 15:01:09 heh, trying to figure out if find_home_near_place(c) will ever choose c 15:01:24 surprisingly not straightforward! 15:03:31 -!- Marvin is now known as Guest61745 15:08:08 gammafunk: it looks to me like it chooses c preferentially to any other place 15:08:40 oh, no 15:10:33 okay, i concur that it is surprisingly difficult 15:13:19 gammafunk: when you figure it out, extract the for loop and the if section to function with names that explain what the heck they are doing, for next time :P 15:13:46 welcome to every other block of code I've been reading lately 15:14:01 thank got the pathfinding code is well-documented 15:14:06 s/got/god/ 15:17:39 !source find_home_near_place 15:17:40 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/monster.cc#L4828 15:18:07 what is this used for? 15:18:41 wow, so I figured out how SquareArray<> works 15:18:42 oh, abyss generation? 15:18:53 it's used in a number of areas 15:18:57 just does what it says on the can 15:19:14 well, a bit more than that, since it actually moves the monster to the chosen location 15:19:58 Brannock: also, for example, beogh recall or maybe stair transit as i recall? 15:21:19 very handy method, just need to make sure I'm using it correctly 15:23:46 look like 15:35:44 looks like it will never choose the center 15:35:51 but it also looks like it's coded incorrectly 15:36:01 in the case for when the destination if e.g. full of monsters 15:36:06 I'm probably reading that wrong somehow 15:36:23 incorrectly how? 15:37:28 if c is surrounded by monsters (a monster on every tile that's adjacent to c and which is habitable grid-wise for the monster) 15:37:47 it *looks* like it's assuming place will be set to a valid location, which it looks like it wouldn't be 15:40:05 gammafunk: i think it just returns false in that case. move_to_pos is only called if nvalid is non-zero, which only happens if check_set_valid_home decides the target was legal 15:40:45 ah yeah, actually nvalid won't be increased in that case 15:40:53 and then the caller is supposed to place the monster in a place that is less desirable. eg, for beogh orcs following you through a staircase, if they can't find enough room near the destination stair, they get dropped randomly anywhere on the level 15:40:57 but it won't return false specifically in the case I described 15:41:01 it will actually continue on 15:41:15 checking the next set of tiles at dist > 1 15:41:21 or specifically at dist 2 15:41:43 so yeah, looks like it works properly 15:42:06 murtidash (L18 DsCK) ASSERT(rc == 0) in 'mutation.cc' at line 2107 failed. (Abyss:3) 15:42:21 and will never choose the center, as nvalid is 0, and when nvalid becomes non-zero, the min. acceptable distance will already have inreased to 1 15:42:26 phew! 15:44:05 I think the only thing I'm going to do is document the method and mention that at doesn't ever choose c 15:44:41 it's doing reservoir sampling on top of all this 15:44:50 through check_set_valid_home 15:46:44 sure. the trick is figuring out what it's sampling from 15:49:14 it looks to me like it's just choosing randomly an acceptable spot at radius 1, and then if it can't find one choosing an acceptable spot at radius 2...up to MAX 15:50:01 trying to do a breadth-first search basically 15:56:21 03sooheon02 {Brannock} 07* 0.20-a0-935-gb48d95f: Remove Monks' starting weapon choice 10(8 hours ago, 1 file, 2+ 3-) 13https://github.com/crawl/crawl/commit/b48d95fd49c1 15:56:21 03sooheon02 {Brannock} 07* 0.20-a0-936-gaf809b6: Move Monks to Adventurer classification 10(8 hours ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/af809b6da4c1 15:59:07 -!- Marvin is now known as Guest75592 16:02:04 4 stealth skill isn't "essentially saved up xp", it's wasted xp 16:02:12 thank you, minmay 16:02:28 (as is the dodging skill mostly) 16:02:48 I like the change but not seeing how it's a buff at all 16:02:57 I don't think it was a buff either 16:03:08 I like very much that now we have a background that starts with 'nothing' 16:03:12 very ascetic! 16:03:37 would it be more interesting if monks just started with like 6-7 fighting and nothing else 16:03:52 !apt fighting 16:03:53 Fighting: Og: 3!, Ba: 2, HO: 2, Mi: 2, Fo: 1, Dr: 1, Mf: 1, Gh: 1, Ko: 1, Gr: 1, Na: 0, Fe: 0, Hu: 0, Op: 0, Mu: 0, Ce: 0, Ds: 0, Te: 0, HE: N/A, Ha: -1, DD: -1, Dg: -1, VS: -1, Vp: -1, DE: -2*, Sp: -2*, Tr: -2* 16:04:20 my impression was that the aim of Mo back when it started with UC was to have the best skills but the worst equipment 16:04:47 and I think that is an interesting gimmick, so I'm glad to see its equipment going back to nothing 16:05:17 but...those skills are just plain worse than Fi or Gl's skills 16:05:36 so there's room for improvement there 16:05:41 technically fighting and dodging 3 -> 4 is 80 more xp, at the cost of 30 lost from weapon skill, you can spin that as a buff of sorts 16:05:55 do monks need to be as good as Fighter? 16:06:00 re starting loadout 16:06:01 I don't think they do 16:06:28 I don't think +1 dodging gives any additional EV for most races 16:06:35 rumflump: skill point cost increases depending on how many skill points you already have, so starting with more levels of a weak skill is actually bad because it makes it harder to get levels in a good skill 16:07:02 I would have to playtest to be sure, but I sincerely predict the background would be stronger with 0 stealth skill and no other changes 16:07:43 would that take in account less likelihood of attracting monster attention with 4 stealth? 16:07:47 yes 16:07:51 hmm 16:09:24 Unstable branch on crawl.jorgrun.rocks updated to: 0.20-a0-936-gaf809b6 (34) 16:10:15 Brannock: even if you're okay with backgrounds being unbalanced, surely each background should have *some* sort of advantage over others - like the commit message says, "worse Gl" is not a worthwhile background 16:11:27 part of the problem is that weapon and spell skills are overwhelmingly more important than other skills 16:11:36 extra health and piety bonus counts in my book, and I am aware that you consider the piety bonus to be meaningless 16:11:57 Wanderer doesn't have much of an advantage either: the random loadout is the attraction of the background 16:12:00 so you would rather have 4 weapon skill than like 10 dodging skill 16:12:44 wanderer's random loadout is usually pretty strong and basically always better than Mo's 16:14:15 if you think the piety bonus is enough of an attraction, fair enough; I won't argue against that 16:14:16 wanderers can start with items no other background has 16:14:53 they could start with some potions, maybe 16:14:56 agility? 16:15:12 misc evokes, potions, scrolls 16:15:30 I want to avoid miscs because of how strong they can be, unless it's something funny like fan of air (and probably not even then) 16:15:30 the first one is the notable one since misc evokes are pretty op 16:15:32 again, I think items is the wrong way to do it because the Mo gimmick is having almost no items 16:15:34 yes 16:15:58 but starting with scrolls of tele ID'd is also unique to it 16:15:59 5 scrolls of random uselessness, imo 16:16:16 Brannock: CanOfWorms is talking about wanderers, which currently can start with a box of beasts, fan of gales, lamp of fire, or phial of floods 16:16:20 starting with +2 armor on wn is also quite potent 16:16:22 oh 16:16:25 I misread, yeah 16:16:27 my bad 16:16:53 it uh...also tries to give the box of beasts 10 to 15 charges 16:17:07 i get the feeling nobody else has looked at this code in a while 16:18:13 does looking at it and throwing up your hands in defeat count 16:18:39 if you want a flavourful item for monks you can give them !degen, it makes sense that they would know what it looks liike (in order to avoid it) 16:19:28 a cask of ascetic ambrosia that they brewed on their ascetic mountain with their ascetic comrades 16:19:49 ascetic degeneration, surely 16:20:07 crawl has already established that it's either alcohol, or a byproduct of alcohol :v 16:20:11 monk backstory: you entered the dungeon because you're trying to smuggle this booze into vaults:5 16:20:33 dungeon crawl prohibition soup 16:20:45 forgot about the old booze and pizza vault 16:22:23 the other thing about monk is that unarmed attacks without claws or transforms are really, really bad 16:22:58 you need a couple skill levels just to beat a club 16:23:16 so if you were really playing old Mo to win you would very frequently use a weapon instead of unarmed 16:24:02 the unarmed skill was less a predefined weapon choice and more a crutch to let you kill hobgoblins 16:25:18 it is possible that the base unarmed damage for untransformed form should be increased; that would help new-new-new monk since finding the first weapon would be less urgent 16:25:30 what's the current value? 16:25:37 also it'd nerf spider form and i love nerfing spider form 16:25:38 3 16:25:58 but consult with Other Devs before doing something like that 16:26:26 yes 16:26:36 since that woudl affect mage starts 16:26:39 not heavily, of course 16:27:11 it makes sense theme-wise that any 0 skill weapon is better than 0 skill unarmed but if theres a background intended to start with 0 skill unarmed as its best attack... 16:27:51 (as cow says you could also fix the problem by giving them lots of fighting skill but idk if people would go for that) 16:28:26 I think fighting is too good to give a lot to a background 16:28:28 esp for high apt races 16:30:52 how much hp would an xl1 ogre with 6 fighting (fixed for apts) have 16:31:30