00:03:49 Unstable branch on crawl.develz.org updated to: 0.11-a0-2067-gcb14100 (33) 00:11:59 -!- Textmode has joined ##crawl-dev 00:18:28 -!- ussdefiant has quit [Ping timeout: 245 seconds] 00:20:05 Windows builds of master branch on crawl.develz.org updated to: 0.11-a0-2067-gcb14100 00:33:21 03elliptic * rb342e0b8f0b2 10/crawl-ref/source/ouch.cc: Don't give immunity to Sunray to invisible creatures. 00:36:53 -!- Pingas has joined ##crawl-dev 00:38:36 -!- Textmode has quit [Ping timeout: 255 seconds] 00:41:52 -!- Textmode has joined ##crawl-dev 00:47:43 -!- ainsophyao has quit [Remote host closed the connection] 00:48:10 03dolorous * r694e5a1c36d2 10/crawl-ref/source/view.cc: Fix capitalization of some Ashenzari auto-id messages. 00:51:33 -!- MarvinPA_ has joined ##crawl-dev 00:54:22 -!- Textmode has quit [Ping timeout: 255 seconds] 00:55:01 -!- MarvinPA has quit [Ping timeout: 248 seconds] 00:55:02 -!- MarvinPA_ is now known as MarvinPA 00:58:04 -!- G-Flex has quit [Quit: Time for us to bug out.] 01:39:31 -!- Guest43275 is now known as ChrisOelmueller 01:44:33 How resource-intensive is it if a player does nothing but take stairs repeatedly? 01:46:02 -!- Dixie has joined ##crawl-dev 01:51:27 A new serial vault (https://crawl.develz.org/mantis/view.php?id=5709) by infiniplex 02:10:02 dtsund: damn badly intensive, I'm afraid 02:10:03 kilobyte: You have 4 messages. Use !messages to read them. 02:10:16 Pity. 02:10:29 -!- SkaryMonk has quit [Read error: Connection reset by peer] 02:10:44 I'm trying to think of a way to implement Brogue-style inter-level chasing that isn't heinous 02:10:58 -!- SkaryMonk has joined ##crawl-dev 02:11:29 currently: saving the level and everything that's not tied to a level (ie, the player, global data, travel cache, etc) 02:11:54 actually, that would _reduce_ the pressure on everything but memory 02:12:16 Perhaps if I saved the level itself and the monsters into two separate chunks, so I could modify the latter directly 02:12:16 do you think only the next level or everything? 02:12:23 Only the next level. 02:12:45 Current plan is something like: 02:12:57 a simple solution: it could be kept in memory for several turns after you move 02:13:38 -!- ainsophyao has joined ##crawl-dev 02:13:41 When you take stairs, record all monsters within LOS that are chasing you, and how many turns they'd need to reach the stairs. Store copies of them in a list external to all levels. 02:14:12 When they'd reach the stairs, pop them from the list and now store their original indices in a "to be removed" things 02:14:16 *thing 02:14:25 When you return to the level, delete those monsters 02:15:02 But yeah, perhaps we could just actively simulate two floors at once. 02:16:30 (This has become a relatively high-priority thing for Light, since otherwise so-called "mummystabbing" is possible because of food removal.) 02:18:03 dtsund: I'd say three floors would be needed probably. 02:18:32 Also brogue looks SO good. 02:18:51 It's somewhere between "good" and "infuriating" 02:19:17 If you're anywhere near a dragon and you step on a paralysis trap, you die, full stop. 02:19:51 -!- SkaryMonk1 has joined ##crawl-dev 02:20:03 dtsund: What is? 02:20:20 Brogue. 02:20:50 I've only played to "D:2" but the "graphics" are neat :) 02:21:10 Oh, you meant "looks" literally? 02:21:33 brogue is so pretty 02:21:36 it's not even fair 02:22:36 dtsund: Yep. 02:48:21 brogue traps are ridiculous, but you can deal with them fine as long as you search excessively or wear a ring of awareness. problem is that the first option is tedious, and the second is rare 02:49:14 evilmike: That dragon scenario I described is something that actually happened to me 02:49:14 also, i would think a dragon would burn away the paralysis gas... the real BS combo that i've found in that game is goblin conjurer + paralysis gas. their summons are immune to it 02:49:20 I had the amulet and was on my way out 02:49:23 ah 02:52:24 also great is getting trapped in a room with an ally and confusion gas. great way to starve or get beaten to death by your friend (who keeps stepping on the pressure plate). 02:52:30 (yeah brogue traps are ridiculous ugh) 02:52:41 At least with confusion there are things you can do about it 02:52:48 You can still use staffs perfectly while confused 02:53:05 (super secret bonus: you can still use staffs while obstructed) 02:53:29 obstruction is very handy yes 02:53:52 i like brogue, every new version i usually play it and get 1 or 2 wins. still can't hold my interest beyond that, though 02:54:07 i cant really win reliably, all i do is find something overpowered and exploit it (which isn't hard) 02:55:54 brogue was pretty great until I got sick of it yeah. maybe I'll get back into it sometime 02:57:02 the latest version is the best one yet, in my opinion. the early game actually poses a challenge, aside from level 1 (which is an utter waste of time) 02:57:19 What's the most recent version? 02:57:29 1.6.2, I think 02:57:32 ooh 02:57:39 I'll just autoplay the first level as usual 02:57:47 unless it has a door puzzle or something 02:57:49 yeah, autoplay has like a 100% success rate on level 1 02:57:54 I'm really bad at all roguelikes still... 02:58:01 past that, you can actually die (OOD monsters can spawn on 2) 02:58:15 I think vaults don't start until 3, so there is literally no reason to play normally on level 1 02:58:25 I can't understand why the game is designed that way :P 03:16:48 -!- Ragdoll has joined ##crawl-dev 03:35:51 -!- ark____ has joined ##crawl-dev 03:44:07 -!- ens has quit [Ping timeout: 252 seconds] 04:03:56 -!- ens has joined ##crawl-dev 04:10:28 -!- edlothiol has joined ##crawl-dev 04:13:22 -!- Mottie has quit [Ping timeout: 245 seconds] 04:15:26 -!- barbs has joined ##crawl-dev 04:18:38 -!- ens has quit [Ping timeout: 244 seconds] 04:24:09 -!- ens has joined ##crawl-dev 04:31:57 -!- HousePet has joined ##crawl-dev 04:35:40 I have a few poison spell suggestions, do I put them in Consolidated Spell Proposals? 04:45:33 if you'd like. could also make a tavern thread if you want a bit more attention 04:46:00 like, put your stuff on the wiki + post the same stuff in a thread 04:46:08 i guarantee you'll get more feedback that way 04:46:24 hmmm 04:49:39 -!- evilmike_ has joined ##crawl-dev 04:50:42 -!- evilmike has quit [Ping timeout: 256 seconds] 04:51:47 * kilobyte mutters something about the tavern's suckage. 04:52:09 typically, if it suggests something, I'd choose the exact opposite way 04:53:16 so if i post in both places at the same time you will be ambivilant? 04:54:41 well, I suggest the wiki because it's less of a hole than the tavern is... and the tavern because you might be interested in more feeback, even if not all of it is good 04:55:08 i also find it preferable to have feedback/discussion concentrated in a thread, rather than a bunch of weird comments mucking up a wiki page 04:55:16 (maybe that's just me, though) 04:57:02 -!- alefury has joined ##crawl-dev 04:58:25 discussion in a forum thread is more useful for discussion, weird comments on a wiki page are better for reviewing important points of a discussion later 04:58:32 imo 05:00:07 OSX builds of master branch on crawl.develz.org updated to: 0.11-a0-2069-g694e5a1 05:05:37 now i'm unsure whether this should go in Game Design Discussion or Contributions :/ 05:05:52 game design discussion, unless you have a patch or something 05:06:16 gotcha 05:14:07 -!- dtsund has quit [Ping timeout: 240 seconds] 05:24:21 Posted in Tavern. Let the rotten tomatoes begin :) 05:35:13 And wikied: https://crawl.develz.org/wiki/doku.php?id=dcss:brainstorm:magic:spells:propose:new_spells#poison_spell_suggestions 05:35:40 HousePet: you know about the "berserker rage" spell, right? 05:35:48 .... 05:35:56 >_> 05:35:58 it was removed 05:36:02 for a reason 05:36:08 lol 05:36:53 salt storm sounds like something that could thrash slime 05:37:25 and imo holy water rain should be like ozocubu's refrigeration 05:37:49 i would say, no holy spells 05:38:09 Oh, I forgot to suggest evil gods get pissy about it 05:38:15 crawl avoids letting you cast holy spells... those belogn to gods 05:38:33 there was a "holy spell school" of sorts 05:38:41 I wrote up a big post in response to that, but discarded it 05:38:53 Salt Storm is weird... a spell nearly completely useless except for a single branch, but awesome there. 05:39:31 also, ponder the desc of Hat of Pondering... 05:40:25 I agree about evilmike wrt holy spells, but there's no need to require holiness for a demon-specific poison 05:40:30 i had considered suggesting holy water be poison/invoc 05:40:31 ' there was a "holy spell school" of sorts' there was? what was in it? 05:40:56 all that comes to mind right now, is a couple of healing spells, and a "holy light" spell 05:41:57 must be ancient 05:42:06 curiousity makes me want to know what crawl's original spells were. lmao 05:42:21 wait, so only the name/flavour of holy water is bad? 05:43:40 I think, as far as crawl theme goes: 05:43:46 -If a spell involves one of the elements, it needs to INVOLVE that element 05:43:58 For example, scorpion teleport which is currently themed as fire/transloc, probably isn't going to end up being fire as well 05:44:12 It's not like, for example, HoMM3 where bloodlust is fire and weakness is water and so on 05:44:13 what do you mean by "INVOLVE" 05:44:29 monqy: physically manifests or physically manipultes that element somehow I guess 05:44:38 like, every poison spell makes, transmutes or removes poison as far as I can remember 05:44:39 you can explain away a lot with flavouring 05:44:41 same for every fire spell 05:44:47 yes, but that requires actually making flavour first 05:45:30 Yeah that one is really tenuous 05:46:01 Restore Strength (2), Restore Intelligence (2), Restore Dexterity (2), Lesser Healing (2), Greater Healing (6), Cure Poison (3), Purification (5), Smiting (4), Repel Undead (3), Holy Word (7), Heal Other (3), Guardian (7), Pestilence (4), Thunderbolt (Holy/Air 6), Flame of Cleansing (8), Shining Light (7), Summon Daeva (8), Abjuration II (4) 05:46:09 scorpion teleport had nothing to do with fire 05:46:16 wow 05:46:21 look at all those overpowered spells 05:46:24 (and a few I have no clue what they do) 05:46:28 "involve" is a vauge, but good term. A fire spell should either use fire, shoot fire at stuff, or be connected with fire in some way 05:47:43 Patashu: most of those are in the game, in some form or other 05:48:01 (just not as spells) 05:48:11 Cure Poison is 05:48:16 oh, yes 05:48:34 they'd be more powerful as spells, yes 05:48:44 like holy word is extremely powerful. so it only exists as a rare scroll 05:49:03 is thunderbolt anything like lightning bolt? 05:49:06 self healing is really powerful, so it only exists in limited quantity (or on elyvilon piety, which if you overuse will require a lot of effort to get back) 05:49:12 it sounds like smite targeted lightning 05:52:10 look like Thunderbolt _is_ Lightning Bolt 05:52:55 you could cast the spell as Conj/Air 6, or Holy/Air 6, only the name and school differs 05:53:03 haha 05:53:56 same as Abjuration being Summ 3 or Holy 4 -- the same spell, even the visible name is the same 05:54:30 it's SPELL_ABJURATION_I / SPELL_ABJURATION_II only to implement different schools 05:55:38 hmm, interesting 05:57:00 is Pestilence anything like Miasma as in the gloorx/d.drake/tar.soh spell? 05:57:11 pretty amazing at level 4, though I guess cblink was level 4 too huh 05:57:30 pestilence sounds like it might be summon swarm. just a wild guess, though 05:57:43 oh could be that too 05:57:51 i doubt a holy spell would involve nuking enemies with miasma 05:58:07 wait 05:58:09 cblink was level 4? 05:58:12 in what version of crawl? haha 05:59:01 SPELL_PESTILENCE and SPELL_SHINING_LIGHT are unimplemented 05:59:23 Pestilence has a working Zin invoc version, which is indeed Summon Swarm 06:00:06 Shining Light has no code, apparently 06:00:59 TSO has an "Annihilate Undead" invocation, which is same as Dispel Undead (which exists too) 06:01:06 Patashu: 4.00b26 06:03:28 and haste used to be level 8 or thereabouts 06:03:47 there's an amusing comment somehwere about it being lowered to 6, because it's almost the same thing as swiftness. this was pre-nerf haste 06:05:11 perhaps it could be brought back to level 8 06:05:18 and I mean, without unnerfing 06:07:00 sigh 06:07:27 I could see it being 7 06:09:54 level 8 haste would be ridiculous, I could see *removing* the spell though 06:10:42 if haste was removed, charms would be worthless to train, wouldn't it? 06:11:12 there do exist other charms spells... 06:11:29 any higher level than haste and worth casting? hmm, I'll check the wiki I guess 06:11:41 death's door 06:11:43 oh, deflect missiles is still a charm 06:11:44 and ddoor 06:11:50 ok, I could see haste being level 7 06:12:34 elliptic: oh, now that you're here, let's discuss mon-pick? 06:14:01 kilobyte: sure... I guess you've seen all the messages I spammed you with earlier? 06:14:43 the -5/+5 limit seems for all practical purposes absolute -- there's no branch that could plausibly fail to pick a monster after 1000 tries (n_relax_margin) 06:14:58 that was my impression, yeah 06:15:37 and it's an "once in the life of the Universe" kind of chance 06:17:11 this narrowing changes the shape of the quasi-parabola, though 06:17:18 not sure how to represent that 06:18:23 probably what we should do is add another parameter to vary how flat the distribution is 06:18:55 OODs at least have a simple effect: I don't see anything that would make them work differently from just acting like you were at that depth 06:19:16 I think evilmike_ was suggesting that this would be a good idea anyway for some monsters? giant eyeballs? 06:20:05 yeah, they are the only ones that immediately come to mind, but there are probably others 06:20:10 If not flatness, what about having single giant eyeballs at depth X, packs of two giant eyeballs at depth Y, etc 06:20:14 (that's how brogue does it) 06:20:28 boggarts would perhaps be another monster, although you wouldn't want those early 06:20:36 but otherwise they'd be reasonable from mid-D all the way to 27 06:21:15 I insist on replacing the current crazy distribution (which is some random quartic polynomial) with something more sensible, anyway 06:21:15 -!- Nomi has quit [Ping timeout: 260 seconds] 06:21:29 -!- ainsophyao has quit [Remote host closed the connection] 06:21:38 and the weight number shouldn't affect the shape of the distribution the way it currently does, it should just scale the whole thing 06:21:49 that odd <= instead of < for chance should be probably removed, I doubt it's intentional 06:21:59 -!- ainsophyao has joined ##crawl-dev 06:22:13 kilobyte: the random2avg(100,2) is garbage anyway 06:22:20 -!- ark____ has quit [Quit: Leaving] 06:22:35 -!- ark____ has joined ##crawl-dev 06:22:37 the only effect is 1/3400 chance to ignore "chance" and instead picking any eligible monster from -5..+5 depth 06:23:53 (ie, the only way other than OODs to pick something from farther than sqrt(rarity), but not farther than 5 away) 06:23:55 -!- ainsophyao has quit [Remote host closed the connection] 06:24:06 the parameters that we should have are: min depth, max depth, something controlling the shape of the distribution, and a weight (that just scales everything) 06:25:30 for your remark that the monsters should be sorted by mindepth: the list I posted is sorted by avgdepth, not sure if that's better 06:25:37 obviously, trivial to change 06:26:13 not a big deal, yeah... just felt to me like putting mindepth first and sorting by it might make it easier to read 06:27:09 should floor mindepth to avgdepth-5 and maxdepth to avgdepth+5 06:27:24 and perhaps somehow add that "flatness" 06:27:55 on the other hand, its tempting to remove parameters to make things simpler 06:28:01 no idea, though 06:30:57 -!- ainsophyao has joined ##crawl-dev 06:30:59 I don't have a great idea about what distribution to use except that it shouldn't be a quartic polynomial :P 06:33:41 I have bad new for you: we have to pick something 06:33:51 Does crawl have a 'place OOD monsters on generation' mechanic? 06:33:51 er, news 06:33:58 Or does it rely on the current weird formula to do it for it? 06:34:43 Patashu: both, vaults have "8" and "9" symbols, for depth+5 and (depth+2)*2, respectively 06:35:07 -!- Ragdoll has quit [Ping timeout: 240 seconds] 06:35:10 and MONS: place:Branch:13 for full control 06:35:29 03evilmike * r7ff2355b00cf 10/crawl-ref/source/dat/ (3 files in 2 dirs): A huge columns-themed serial vault. (infiniplex) 06:35:30 But not in general dungeon generation, ok 06:35:34 03evilmike * r05b8b70f3074 10/crawl-ref/source/rltiles/ (8 files in 2 dirs): New crumbled column tiles. (ontoclasm) 06:35:35 (needs lua if you want it relative to current depth) 06:35:58 I was thinking you could make it all linear distributions from mindepth to maxdepth, and instead of having parahyperbolas or whatever just make it so x% of monsters are one depth OOD, y% are two, etc down to some maximum deepness 06:35:59 Patashu: regular generation has a chance for any spawn to be OOD 06:36:08 03evilmike * reb5f7f930c72 10/crawl-ref/source/dat/des/serial/ (column_ruins.des column_ruins2.des): Rename column_ruins.des. 06:36:49 Patashu: besides the OOD roll, the -5..+5 rule is a hard cut-off 06:37:05 -!- ainsophyao has quit [Remote host closed the connection] 06:37:13 -!- ainsophyao has joined ##crawl-dev 06:39:59 -!- mikee_ has quit [Quit: HydraIRC -> http://www.hydrairc.com <- Po-ta-to, boil em, mash em, stick em in a stew.] 06:40:15 unless I made some mistake, the chance to avoid -5..+5 tends to be around 10^-22ish for most branches 06:49:41 kilobyte: I will propose replacing the random2avg(100, 2) with random2(100) and setting chance = rarity*flatness_param/100 at depth = min_depth or depth = max_depth, chance = rarity at depth = (min_depth+max_depth)/2, and linearly interpolate in between 06:49:57 you might be right that having the extra flatness parameter isn't worth it though 06:51:14 you could just use a gaussian distribution that is cut off at +-5. std controls sharpness of the peak. Of course with varying total weight its not a real probability, but assuming constant total weight its quite easy to interpret. About 2/3 of events lie within +-std, something like 95% within +-2std. 06:51:53 we could also use a real distribution and truncate, yes 06:51:55 a curve might be better, especially since random2avg(100, 2) effectively squares low effective chances 06:52:22 are there decent distributions that naturally reach 0 at ends of range? 06:53:14 kilobyte: the linear thing I was suggesting is actually a sort of reasonable approximation to the current distribution 06:53:40 about reaching 0 at ends of range, that would be more natural but it also isn't how the current distribution works at all for high rarity values 06:53:41 also, if the idea of specifying range rather than avg and diff holds, there's a minor issue of where the peak is if max-min is odd 06:55:03 elliptic: hmm right, my mistake, I reversed where the squaring is in the parabola, it may indeed be countered by squaring of low parts by random2avg 06:55:14 at rarity 70, say, the actual probabilities you get after taken into account are proportional to 4186, 5995, 7636, 8911, 9730, 10011, 9730, 8911, 7636, 5995, 4186 06:55:46 -!- HousePet has quit [Quit: Divide by cucumber error] 06:56:17 I'm not sure how much we should care about trying to approximate the old behavior, though... I think it will be pretty difficult to do so for all rarity values at once 06:57:22 elliptic: the formula you proposed sounds pretty good. 06:57:46 no idea about approximation, but it seems readable and reasonable 06:58:34 I'm sort of liking the truncated normal distribution idea though 06:58:39 kilobyte: about the peak for intervals of uneven length, there could just be two 06:58:43 s/uneven/odd 07:00:13 i think the distribution is pretty irrelevant as long as it has some kind of peak, and the sharpness of the peak can be scaled 07:00:30 your suggestion seems easier to interpret than a cut off normal distribution 07:01:21 perhaps the stddev could eliminate the ranges? 07:01:29 I guess the problem with a normal distribution is approximating it discretely is going to be pretty silly in small range cases 07:01:48 -!- ainsophyao has quit [Remote host closed the connection] 07:01:55 sure, you just need mean and std for a normal distribution 07:02:33 the range would just always be +-5 then 07:08:27 any opinions on removing the +-5 restrction, btw? as evilmike_ said, some monsters work well throughout most of D. 07:09:39 well I assume we'd be removing that regardless... the only reason I brought it up is that kilobyte's extraction of the current ranges wasn't respecting that 07:09:49 after we switch to a new system, we can do whatever we want 07:09:56 especially zombies/skeletons: for melee monsters, they could even be tied to the depth for their base 07:10:14 Normalization would be tricky for a cut off normal distribution btw. Normalize without cutoff? With cutoff? Just set the max = rarity? 07:10:35 elliptic: yeah, I extracted that assuming quite a part of the old formula would be kept, which can indeed be scrapped 07:11:00 any kind of nonstandard normalization kills easy interpretation of std 07:11:42 also, rarities > 100. Currently, they affect only the flatness. 07:12:00 alefury: I think max=rarity is most natural probably from our perspective, but you are right it is a bit problematic 07:12:18 just one more reason why i prefer your suggestion 07:12:39 kilobyte: yeah... not many of those though, and if we have a flatness parameter we can just set that to something high for those 07:13:03 the extractor script can use some approximation 07:14:09 so what then? Four parameters? 07:14:40 could add a fifth controlling skewedness, but four are probably good enough 07:14:45 -!- ussdefiant has joined ##crawl-dev 07:15:42 kilobyte: if we want to approximate the old behavior, I think 4 parameters, and set the smoothness parameter based on the rarity 07:15:42 I was looking more into a reduction 07:15:44 (settable depth of maximum for elliptics suggestion instead of always using avgdepth) 07:16:10 -!- evilmike_ has quit [Read error: Connection reset by peer] 07:16:40 kilobyte: I could also see not worrying about smoothness though... I'm not sure how great it is in gameplay terms 07:16:55 also s/smoothness/flatness/ 07:16:58 kilobyte: for 3 parameters you would have to use avgdepth with a fixed range, or set flatness based on rarity. 07:17:49 or just fix it to 20 or something 07:17:53 or 0 07:18:10 not 0. 07:18:25 setting flatness based on rarity is really weird IMO, one of the features of the current system that makes it hardest to understand 07:19:36 i dont see the problem with having 4 parameters. flatness could just be 10 for almost all monsters. 07:20:00 the added flexibility would be nice in some cases, so having it is better than not having it imo. 07:20:14 alefury: range is currently fixed for all but rarest monsters 07:20:33 if you want to get fancy you can even allow omitting it. 07:20:57 kilobyte: i think flexible range would be great, for stuff like eyeballs and boggarts, as evilmike said 07:20:59 ie, by placing that on the right side of the table? 07:21:08 it could be approximated by simply having duplicate entries, though 07:21:23 I formatted it with the monster last since it's the only field with non-constantish width 07:21:44 heck, yeah, the new format allows for duplicates 07:21:49 if flatness was "default" for most monsters that wouldnt matter 07:22:07 not sure what duplicates would be good for, though 07:24:42 assymetric distributions by having two symmetric ones with different mean. also, with fixed range (for example +-5) you could have one spawn from 7-17 and one from 18-28 or something. 07:26:35 I think having flatness on the right side if it was "default" for almost all entries would be fine. If you want to vary flatness to better approximate the old behaviour it would be bad. 07:30:52 -!- clouded_ has joined ##crawl-dev 07:33:03 actually allowing for omitting entries would probably be pretty bad for clarity 07:33:39 -!- ussdefiant has quit [Ping timeout: 252 seconds] 07:35:48 another idea: instead of 1. going through MONS_0 .. NUM_MONSTERS, picking eligibles, copying them to a list, 2. picking from the list, rolling chance and rejecting -- perhaps we could calculate effective chance in the first loop, and roll against the sum? 07:51:21 actually all this stuff is constant, so you could put eligible monsters and their chances into some kind of db if you wanted. 07:54:52 constant for a given depth, not for a branch 07:58:10 yes 07:58:34 no idea if speeding up monster generation is at all relevant 08:02:43 it'd actually slow it down -- flatness calculations would be done for the whole array instead of just picked values, but it'd allow using rarities > 100 without surprises 08:03:42 on the other hand, rarities > 100 can be removed, and checked for in -test 08:03:56 the table so far: http://sprunge.us/bCLN 08:04:04 a hard to read wall of numbers :( 08:04:40 there should at least be a comment on top saying what number is what 08:04:52 other than that i dont see much of a problem with a wall of numbers 08:05:12 flatness is a function of rarity, too 08:05:17 kilobyte: maybe round the flatness parameter to just a few values 08:05:38 like, could have it be 20/40/60/80/100 only, or something 08:05:56 increments of 10 would probably be fine too 08:05:57 or even fewer values 08:07:23 i think crawl is fairly resilient to balance changes other than during the first few levels of D 08:08:10 so for the most part i dont think approximating the old behaviour as well as possible is not very important 08:08:56 s/dont / 08:10:39 another really unobvious thing is the OOD cap 08:10:50 !learn add alefury toenail golems make mephclouds when hit by fire 08:10:51 alefury[5/5]: toenail golems make mephclouds when hit by fire 08:10:54 yes, i was wondering about that 08:10:58 especially if it is lower than the max depth of the branch 08:11:48 ie, Hells and Zot 08:12:04 ummm, okay... what does that do? 08:13:59 !learn del alefury[3] 08:14:00 Deleted alefury[3/5]: look up that good hexes breaking through MR idea on the tavern 08:15:01 alefury: depths deeper than the OOD cap take from the cap instead 08:15:34 just for OOD monsters or for all? 08:15:55 this shows up most prominently on Vaults:8 as the vault consists mostly of "8" and "9" symbols, which take from Vaults:13 and Vaults:15, respectively 08:16:20 alefury: normally just OOD, except for Hells and Zot 08:19:11 03dolorous * r9a4ba214cf27 10/crawl-ref/source/dat/descript/features.txt: Reword broken pillar description. 08:21:32 in D/Vaults, the only eligible monsters are: yaktaur captain, deep troll, fire giant, frost giant, lich, phantasmal warrior, glowing shapeshifter, tentacles monstrosity, {storm,shadow,bone,quicksilver,iron,golden} dragon, deep elf {hpriest,demonol,annih,sorc,death mage}, titan, ancient lich 08:27:18 huh? eligible at what depth? 08:28:29 D:31/Vaults:15 (the OOD cap) 08:29:11 that seems like a fine set of high threats 08:29:42 V:8 works pretty great in general 08:30:40 maybe we could buff phantasmal warrior, it's sort of pointless as is 08:31:32 by the way, reading the backlog: I support removing the haste spell 08:33:02 interesting, not a single V:8 vault uses a monster other than "8" or "9", with a heavy stress on the latter 08:34:22 on phantasmal warrior, how about making it fast and giving it rock worm properties? 08:34:47 rock worm seems to be odd 08:34:51 fuck rock worms 08:34:57 and that, yeah 08:35:15 on the other hand, passing walls but not being able to attack from there could be good 08:35:59 rock worms are a pain in the backside because you need a scummy tactic to fight them 08:37:23 hrm no, passing walls but not attacking would need extra AI, for hardly any gain 08:37:59 it would also raise the question why not all ghosts can go through walls 08:38:53 causing balance problems with early pghosts 08:44:41 rock work should only block attacks 50% or 75% of the time 08:44:44 100% is just pointless 08:52:09 -!- Grunt has joined ##crawl-dev 08:53:21 yeah... the rock worm has to at least partially leave rock to attack you (theme-wise), in a system without discrete turns you'd be able to strike when it moves out 08:53:56 -!- ToBeFree has joined ##crawl-dev 09:08:53 -!- ark____ has quit [Ping timeout: 246 seconds] 09:10:16 03dolorous * r08dabaeaf122 10/crawl-ref/source/view.cc: Avoid unnecessary duplication of Ashenzari warning strings. 09:16:03 -!- ainsophyao has joined ##crawl-dev 09:20:22 Unstable branch on crawl.akrasiac.org updated to: 0.11-a0-2073-g9a4ba21 (33) 09:49:07 -!- ussdefiant has joined ##crawl-dev 10:07:01 -!- Patashu has quit [Ping timeout: 248 seconds] 10:12:19 -!- MarvinPA_ has joined ##crawl-dev 10:12:45 jesus christ, when is spider going to be not guaranteed i am sick of this branch 10:12:48 -!- Mottie has joined ##crawl-dev 10:15:37 -!- MarvinPA has quit [Ping timeout: 240 seconds] 10:15:38 -!- MarvinPA_ is now known as MarvinPA 10:32:11 -!- ens has quit [Ping timeout: 244 seconds] 10:53:08 -!- SkaryMonk1 has quit [Quit: Leaving.] 10:55:19 -!- _wh1te has quit [Ping timeout: 252 seconds] 11:02:01 -!- ark____ has joined ##crawl-dev 11:09:17 -!- barbs has quit [Ping timeout: 246 seconds] 11:14:07 -!- ussdefiant has quit [Ping timeout: 240 seconds] 11:17:07 -!- ussdefiant has joined ##crawl-dev 11:17:12 -!- G-Flex has joined ##crawl-dev 11:18:54 -!- Ragdoll has joined ##crawl-dev 11:31:32 -!- ark_ has joined ##crawl-dev 11:32:09 -!- ark__ has joined ##crawl-dev 11:32:28 -!- ark____ has quit [Ping timeout: 252 seconds] 11:35:55 -!- alefury has quit [Quit: ChatZilla 0.9.85 [Iceape 2.0.11/20120421074340]] 11:36:05 -!- ark_ has quit [Ping timeout: 248 seconds] 11:40:47 -!- MarvinPA_ has joined ##crawl-dev 11:42:51 -!- MarvinPA has quit [Ping timeout: 244 seconds] 11:42:51 -!- MarvinPA_ is now known as MarvinPA 11:54:17 -!- ens has joined ##crawl-dev 11:59:50 -!- _wh1te has joined ##crawl-dev 12:03:29 jesus christ, when is spider going to be not guaranteed i am sick of this branch 12:06:40 * ens has never got spider :/ 12:13:00 -!- ussdefiant has quit [Ping timeout: 260 seconds] 12:14:15 -!- ussdefiant has joined ##crawl-dev 12:18:06 -!- ark__ has quit [Ping timeout: 244 seconds] 12:18:32 -!- ark____ has joined ##crawl-dev 12:18:56 -!- alefury has joined ##crawl-dev 12:29:13 -!- G-Flex has quit [Quit: Time for us to bug out.] 12:40:29 -!- MarvinPA has quit [Read error: Connection reset by peer] 12:47:17 -!- ussdefiant has quit [Ping timeout: 246 seconds] 12:48:37 -!- Dixie has quit [Read error: Connection reset by peer] 12:48:51 -!- Dixie has joined ##crawl-dev 12:53:33 -!- ussdefiant has joined ##crawl-dev 13:09:39 -!- MarvinPA has joined ##crawl-dev 13:15:46 -!- blabber has joined ##crawl-dev 13:46:10 -!- dtsund has joined ##crawl-dev 13:58:26 -!- ainsophyao has quit [Remote host closed the connection] 14:12:20 -!- ussdefiant has quit [Ping timeout: 246 seconds] 14:14:47 -!- MarvinPA_ has joined ##crawl-dev 14:17:09 -!- MarvinPA has quit [Ping timeout: 248 seconds] 14:17:09 -!- MarvinPA_ is now known as MarvinPA 14:22:03 -!- heteroy has quit [Quit: ChatZilla 0.9.88.2 [Firefox 12.0/20120420145725]] 14:22:25 -!- HangedMan has joined ##crawl-dev 14:22:39 -!- ussdefiant has joined ##crawl-dev 14:33:10 -!- ToBeFree has quit [Read error: Connection reset by peer] 14:35:17 some notes on brogue traps talked about earlier: the paralysis gas trap is going to have the plate distant from the gas to allow emergency action in 1.6.3, it tries to balance the horribly breakable ally system, and you can make sure that no traps repeatedly activate (ally and confusion problem) by throwing a dart onto the plate 15:24:13 -!- MarvinPA has quit [Read error: Connection reset by peer] 15:36:12 -!- Nomi has joined ##crawl-dev 15:41:56 -!- __duncan has quit [Remote host closed the connection] 15:51:29 -!- Pingas has quit [Read error: Connection reset by peer] 15:58:53 -!- blabber has quit [Quit: leaving] 16:01:28 -!- blabber has joined ##crawl-dev 16:04:20 -!- ussdefiant has quit [Ping timeout: 246 seconds] 16:07:04 -!- ussdefiant has joined ##crawl-dev 16:08:23 -!- ussdefiant_ has joined ##crawl-dev 16:11:20 -!- ussdefiant has quit [Ping timeout: 246 seconds] 16:11:34 -!- ussdefiant_ is now known as ussdefiant 16:11:49 Nemelex's Choice 16:11:51 Nemelex' Choice 16:12:08 so uh... 16:12:24 n's choice 16:13:36 can we start using "Nemelex's Choice" for the tourneys? 16:13:44 x' seems awful and wrong 16:17:20 -!- MarvinPA has joined ##crawl-dev 16:17:38 -!- ussdefiant has quit [Ping timeout: 245 seconds] 16:18:07 -!- Mottie has quit [Ping timeout: 252 seconds] 16:19:22 to me x' seems right 16:19:43 why? 16:19:45 but im not a native speaker, so that doesnt really matter much 16:19:54 well, x kind of ends in s when spoken 16:20:13 so i wouldnt put an additional one 16:20:15 that doesn't really lend itself to the s-possessive rules 16:21:55 now, i admit i'm basing this on american english standards, because that's my native grammar, as opposed to australian english 16:23:43 if there's some australian standard that says x', then cool. 16:23:56 alefury: right, but do you pronounce it "nemelex choice" or "nemelexes choice" 16:24:18 nemelex choice 16:24:47 i tend to add an "-ihs" sound in cases like this 16:26:23 eg. Rex's, Alice's, Lexus's 16:27:13 well, as i said, im no native speaker 16:33:50 Zannick: I tend to agree about 's being more correct, I just haven't changed it for consistency with past tournaments 16:51:59 How about "Glory of Nemelex" or something? Sidestep the whole issue. 16:52:09 "Pride of Nemelex" 16:53:50 -!- MarvinPA has quit [Read error: Connection reset by peer] 17:00:19 -!- blabber has quit [Quit: gn8] 17:02:29 pit boss 17:10:10 -!- Wensley has joined ##crawl-dev 17:19:46 -!- Ragdoll has quit [Quit: Ragdoll] 17:21:55 -!- _wh1te has quit [Quit: leaving] 17:22:51 -!- heteroy has joined ##crawl-dev 17:32:41 -!- alefury has quit [] 17:59:19 -!- Mottie has joined ##crawl-dev 18:03:18 -!- clouded_ has quit [] 18:05:39 -!- bhaak has quit [Ping timeout: 276 seconds] 18:12:53 -!- bhaak has joined ##crawl-dev 18:13:42 -!- mikee_ has joined ##crawl-dev 18:15:06 elliptic: about that "flatness" parameter: what if we limit it to, say, just three profiles? 18:15:16 fine by me 18:16:48 also, without enabling C++11 mode, there is no way to specify an array or vector literal with sub-tables of different length 18:17:11 I wonder if there is a way to hack it around using global constructors or some such 18:17:26 c++ is like a game 18:17:34 like global thermonuclear war 18:17:35 otherwise, I'd have to split it into per-branch arrays 18:18:03 the only way to play is not to win 18:18:06 wensley, not to playt 18:18:14 * kilobyte tosses some Cobol, or even worse, Python, towards Wensley. 18:18:15 erm 18:18:20 got that backwards, didn't i 18:18:23 yes 18:18:45 eh, still means something close 18:18:50 kilobyte: I see your cobol and raise you rpg-le 18:19:33 from war import globalthermonuclear 18:19:43 Zannick: pep 8 violation 18:20:07 -!- HangedMan has quit [Quit: showstopper] 18:20:08 oh no i violated a pep 18:20:20 nine demerits 18:20:33 oooh, what does that allow me to buy? 18:20:39 * Zannick goes shopping at the python store 18:21:48 Zannick: nine demerits will buy you three curly brace plushies 18:21:56 they are on fire sale at the python store 18:22:03 maybe i'll just start saving or investing 18:22:04 I had to fix something in a Fortran 77 program once, that gave me a good insight how bad meaningful whitespace is 18:22:10 we can't even give these braces away 18:22:28 -!- HangedMan has joined ##crawl-dev 18:22:55 kilobyte: meaningful whitespace is soooo bad. fortunately whitespace in python is no more meaningful than forcing you to put spaces between keywords in c++ 18:23:32 publicstaticvoidmain(Stringargs[]) <-- valid java 18:23:39 wensley vaults should totally have meaningful whitespace, abyss shenanigans are unable to randomly morph bits of vault without it 18:23:48 seriously? that's valid java? 18:23:54 (no) 18:24:03 i didn't think so :P 18:24:17 also not having to have repetitive barely-mutated vaults for certain things (MAP G ENDMAP) 18:45:59 -!- |amethyst has joined ##crawl-dev 18:58:23 -!- heteroy has quit [Quit: ChatZilla 0.9.88.2 [Firefox 12.0/20120420145725]] 18:59:50 -!- MarvinPA has joined ##crawl-dev 19:03:08 -!- Patashu has joined ##crawl-dev 19:10:19 -!- edlothiol has quit [Ping timeout: 272 seconds] 19:21:47 -!- evilmike has joined ##crawl-dev 19:29:40 -!- ainsophyao has joined ##crawl-dev 19:39:00 -!- ainsophyao has quit [Remote host closed the connection] 20:04:46 -!- ussdefiant has joined ##crawl-dev 20:46:11 -!- ark____ has quit [Quit: Leaving] 20:51:13 -!- ainsophyao has joined ##crawl-dev 20:52:39 -!- MarvinPA_ has joined ##crawl-dev 20:56:17 -!- MarvinPA has quit [Ping timeout: 245 seconds] 20:56:17 -!- MarvinPA_ is now known as MarvinPA 21:07:14 -!- Pingas has joined ##crawl-dev 21:07:37 -!- HangedMan has quit [Quit: showstopper] 21:37:12 -!- MarvinPA has quit [] 21:40:21 -!- capablanca has joined ##crawl-dev 21:40:31 -!- mikee_ has quit [Disconnected by services] 21:40:31 -!- capablanca is now known as mikee_ 21:55:29 enigmoo (L10 VpNe) (D:9) 22:06:24 <|amethyst> I'm adding "any rod" to the itemspec syntax at hangedman's request... would it make sense to have a way to ask for any non-rod magical staff? 22:07:42 <|amethyst> and, for that matter, is there any reason other than inertia (and the size of the change) not to split OBJ_STAVES ? 22:08:06 <|amethyst> I guess there are things, like acquirement, that lump them together 22:09:42 <|amethyst> but vaults could use e.g. "any magical staff / any rod w:1" for the current effect 22:11:25 -!- G-Flex has joined ##crawl-dev 22:21:29 -!- Patashu has quit [Read error: Connection reset by peer] 22:21:39 -!- Patashu has joined ##crawl-dev 22:36:06 -!- ussdefiant_ has joined ##crawl-dev 22:46:40 -!- evilmike_ has joined ##crawl-dev 22:50:04 -!- evilmike has quit [Ping timeout: 252 seconds] 23:00:00 -!- monqy has quit [Quit: hello] 23:00:21 -!- Sequell has quit [Read error: Connection reset by peer] 23:09:21 Debian builds of master branch on crawl.develz.org updated to: 0.11-a0-2074-g08dabae 23:30:48 -!- ussdefiant_ has quit [Ping timeout: 244 seconds] 23:31:57 <|amethyst> Also, is there anything wrong with rob's suggestion in https://crawl.develz.org/mantis/view.php?id=2554 (from two years ago) 23:32:30 <|amethyst> more specifically, to veto positions that do not have LOS_SOLID to the leader 23:34:44 <|amethyst> even better would be something that let monsters be placed around nearby corners but not across walls; limited-depth pathfinding would be one option, checking LOS_SOLID to any previous member (not necessarily the leader) another 23:52:02 <|amethyst> do any existing vaults depend on placing band members across walls/grates/etc ? 23:52:35 <|amethyst> or in an area with limited visibility 23:52:44 not that I know of, and I disapprove of vaults that do things like that 23:53:11 a couple take advantage of summoning through walls 23:53:22 <|amethyst> I'm not changing summoning 23:53:30 <|amethyst> so those won't be an issue 23:53:39 -!- Grunt has quit [Quit: leaving] 23:54:53 looks like there are only a few vaults that place bands at all. The majority are uniques 23:57:41 <|amethyst> what about Grum? 23:57:52 Grum is a unique. 23:58:00 <|amethyst> his room isn't big enough for all his dogs 23:58:32 -!- SkaryMonk has left ##crawl-dev 23:59:48 <|amethyst> oh, never mind