00:28:26 <Eksell> Unstable branch on CRAWL.XTAHUA.COM updated to: 0.21-a0-538-gd16b88a (34)
00:30:39 -!- mikee__ is now known as mikee_
00:57:06 -!- mikee__ is now known as mikee_
01:12:27 -!- mikee__ is now known as mikee_
01:20:58 <Gretell> Unstable branch on crawl.develz.org updated to: 0.21-a0-538-gd16b88a (34)
01:22:51 -!- mikee__ is now known as mikee_
02:00:20 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.21-a0-538-gd16b88a
02:16:20 -!- mikee__ is now known as mikee_
02:32:28 <Lantell> NullDivider (L11 SpAr)  (Lair:1)
02:32:29 <Lantell> Webtiles server stopped.
02:32:29 <Lantell> Webtiles server started.
02:37:07 -!- mikee__ is now known as mikee_
02:53:29 -!- mikee__ is now known as mikee_
02:55:17 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.21-a0-538-gd16b88a
03:04:08 <SteelNeuron> !messages
03:04:09 <Sequell> (1/3) advil said (15h 29m 51s ago): new proposal for walljump UI: (i) have two modes, one walljumps via movement with no delay (i.e. your initial version), and one walljumps using an ability + targeter, and (ii) have a pseudo-ability that toggles between the two modes in game (could still keep the rc option to set the default)
03:04:19 <SteelNeuron> !messages
03:04:20 <Sequell> (1/2) advil said (15h 28m 18s ago): I think the current version could work if crawl had more things you could do in a similar way, and I have found that it reduced my accidental walljumps, but it's too far off typical crawl abilities and maybe a bit too much in-between the other two versions & isn't quite careful enough for careful players
03:04:23 <SteelNeuron> !messages
03:04:23 <Sequell> (1/1) advil said (15h 27m 57s ago): Another specific concern is that it's not cancellable, which would be straightforwardly fixed in an ability version
03:05:08 <SteelNeuron> !tell advil I like your proposal. For me personally, the ideal would be defaulting to the initial version, and having a RC file option for the ability version; the toggle seems a bit too fiddly
03:05:09 <Sequell> SteelNeuron: OK, I'll let advil know.
03:06:39 <SteelNeuron> !tell advil I generally push to keep it movement-triggered because I know it's important to a few of the players that enjoy the god, it's that difference that makes it unique I think; but it's very valuable to have another option.
03:06:39 <Sequell> SteelNeuron: OK, I'll let advil know.
03:07:35 <SteelNeuron> !tell advil Not sure if I got your response on this... I may have but I forgot. What do you think of a single [Y/n] when walking into a wall? Too wordy/annoying? It would 100% stop accidental wall jumps even when double tapped at least
03:07:35 <Sequell> SteelNeuron: OK, I'll let advil know.
03:21:27 <Rotatell> Unstable branch on crawl.beRotato.org updated to: 0.21-a0-538-gd16b88a (34)
04:21:15 -!- Evablue_ is now known as Evablue
04:58:03 -!- amalloy is now known as amalloy_
04:59:11 -!- amalloy_ is now known as amalloy
05:47:39 -!- amalloy is now known as amalloy_
05:57:52 -!- Amnesiac_ is now known as Amnesiac
09:41:28 <|amethyst> hm, I might try cross-grading my home i386 Debian install to amd64 this weekend
09:41:55 <|amethyst> that ought to be fun
09:44:44 <|amethyst> (especially since this is a Debian unstable install that has been periodically upgraded over the course of 15 years)
09:50:30 <gammafunk> is your cpu actually 64-bit?
09:50:35 <|amethyst> yeah
09:50:42 <|amethyst> I was running an amd64 kernel for a long time
09:50:55 <|amethyst> but Debian no longer provides amd64 kernels in the i386 repo
09:51:00 <gammafunk> hah
09:51:39 <gammafunk> It never occured to me to try to have a 64-bit kernel with 32-bit userspace
09:51:57 <|amethyst> and now I'm running into the problem that I/O in an 32-bit-with-PAE kernel gets very very slow I/O when you have more than about 8GB of RAM
09:52:33 <|amethyst> (because the kernel spends most of its time shuffling around memory to put the things it needs in the low 4 GB)
09:53:28 <|amethyst> (actually, just its 1GB slice of the low 4GB)
09:53:38 <gammafunk> I guess that's a delay for merging food-unification?
09:53:56 <gammafunk> You mentioned you were going to do that soon(ish)
09:54:06 <|amethyst> eh, I can click the purple button from a laptop :)
09:54:12 <|amethyst> actually, you want to do that?
09:54:24 <gammafunk> hrm, it would be a good idea for me to fully over that commit list anyhow
09:54:49 <|amethyst> yeah, several of those bug fixes could be squashed down
09:54:49 <gammafunk> I'd like to finish this spell I'm working on, but if I don't finish that today, I'll set about looking that over/merging
09:55:02 <gammafunk> if it's going to get merged it deffo needs to be pretty soon
09:55:30 <gammafunk> a new spell going in a couple weeks before release is one thing, a new food system is a bit more impactful
09:55:52 <gammafunk> longer we can have that tested before release the better
09:56:09 <|amethyst> I never did get around to implementing auto_eat_rations
09:56:20 <|amethyst> oh, something strage I noted
09:56:33 <|amethyst> !lg * / starvation s=cv o=.
09:56:37 <Sequell> 9674/8174504 games for *: 159/503108x 0.21-a [0.03%], 131/357160x 0.20 [0.04%], 156/558103x 0.20-a [0.03%], 184/372120x 0.19 [0.05%], 229/493909x 0.19-a [0.05%], 201/312683x 0.18 [0.06%], 553/418096x 0.18-a [0.13%], 890/323782x 0.17 [0.27%], 817/499777x 0.17-a [0.16%], 494/275038x 0.16 [0.18%], 329/323663x 0.16-a [0.10%], 351/223863x 0.15 [0.16%], 245/264211x 0.15-a [0.09%], 231/143404x 0.14 [0.16...
09:56:41 <|amethyst> !lg * / starvation s=cv o=. -graph
09:56:44 <Sequell> 9674/8174504 games for *: https://shalott.org/graphs/fb2bfac51683a828d48fc1cb5aae2d0adf7a2191.html
09:57:25 <|amethyst> I guess 0.17-a was when hungry ghosts were first changed
09:57:46 <gammafunk> !kw hungrycobs
09:57:46 <Sequell> Keyword: hungrycobs => vlong>=0.17-a0-1659-gf62aa6c vlong<0.18-a0-1264-g90af9db
09:57:49 <gammafunk> yeah
09:57:50 <|amethyst> but why has starvation been so low since 0.18?
09:58:04 <|amethyst> it's much lower now than it was before hungry cobs
09:58:12 <|amethyst> s/has starvation/have starvation deaths/
09:58:33 <gammafunk> what was the big food change in 0.18-a?
09:58:37 <gammafunk> ??changelog
09:58:37 <Sequell> changelog[1/3]: https://github.com/crawl/crawl/blob/master/crawl-ref/docs/changelog.txt - often incomplete and/or out of date, see {changelog[3]} for an exhaustive list of changes.
10:00:03 <gammafunk> I could guess something like af_hunger being changed
10:00:15 <gammafunk> and it spiked with death cobs but then we removed af_hunger on those
10:00:20 <gammafunk> but hungry ghosts still had new af_hunger
10:00:34 <gammafunk> so essentially we nerfed hungry ghosts
10:00:52 <|amethyst> hm, I would have expected more starvation deaths to bad play than to hungry ghosts
10:01:15 <gammafunk> sadly it's difficult to track
10:01:21 <|amethyst> especially with the way it's often higher for releases than trunk (see 0.13-0.16)
10:01:23 <gammafunk> due to hungry ghosts not being the killer
10:01:37 <advil> was't that death cobs?
10:01:37 <Sequell> advil: You have 6 messages. Use !messages to read them.
10:01:51 <advil> oh nm
10:01:58 <advil> it's in scrollback
10:02:04 <|amethyst> I wouldn't expect death cobs to have much of an effect at all
10:02:18 <|amethyst> because of how few games get that far
10:02:27 <|amethyst> otoh, these are all tiny numbers
10:02:45 <gammafunk> yes, and death cobs spiked zot kills real good as I recall
10:02:50 <gammafunk> .cobbed
10:02:54 <Sequell> 135. jazmu the Slayer (L26 GrWn of The Shining One), starved to death on Zot:2 on 2016-09-30 04:12:15, with 557974 points after 70343 turns and 4:41:46.
10:03:00 <gammafunk> .cobbed s=cv
10:03:01 <Sequell> 135 games for * (hungrycobs place=zot|crypt ((ktype=starvation || ikiller=death_cob))): 62x 0.17, 45x 0.17-a, 28x 0.18-a
10:03:20 <gammafunk> yeah for numbers like ~200, 45 more is a lot
10:03:33 <gammafunk> well
10:03:36 <gammafunk> that's not 45 more
10:03:46 <gammafunk> .cobbed starve
10:04:00 <advil> man did I hate those things :)
10:04:01 <Sequell> 91. jazmu the Slayer (L26 GrWn of The Shining One), starved to death on Zot:2 on 2016-09-30 04:12:15, with 557974 points after 70343 turns and 4:41:46.
10:04:05 <gammafunk> .cobbed starve s=cv
10:04:18 <Sequell> 91 games for * (hungrycobs place=zot|crypt ((ktype=starvation || ikiller=death_cob)) starve): 50x 0.17, 29x 0.17-a, 12x 0.18-a
10:04:30 <wheals_> auto eat chunks got turned on by default in 0.18, maybe the people who were dying to starvation weren't eating their chunks before they rotted
10:04:31 <gammafunk> yeah, just 29 more for that numerator
10:04:43 <gammafunk> Yeah I was wondering about that too
10:04:47 <gammafunk> I didn't see it mentioned in the changelog
10:05:08 <|amethyst> wheals_:  hm
10:05:11 <|amethyst> wheals_:  I bet that is it
10:05:18 <wheals_> oh huh
10:05:25 <wheals_> i misread
10:05:27 <gammafunk> oh you said before they rotted
10:05:30 <gammafunk> ghouls can't starve
10:05:35 <wheals_> that was 0.17
10:05:47 <gammafunk> oh you mean before the chunks rotted
10:06:07 <wheals_> yeah
10:06:11 <gammafunk> but yeah I was wondering if auto eat might have lead to less starvation deaths somehow
10:06:19 <gammafunk> but yeah didn't see that changed in 0.18 changelog
10:06:42 <wheals_> yeah i was looking at the 0.17 section by accident
10:06:42 <|amethyst> hm, 0.17 would have been mixed in with the cob/ghost thing
10:06:55 <|amethyst> so once those were resolved in 0.18
10:08:35 <advil> there were a bunch of small things that could have had some impact, i.e. the more detailed starvation levels, autoeat when starving
10:10:20 <advil> tweaks to when you could use high hunger abilities
10:10:50 <|amethyst> berserk hunger change
10:12:34 <|amethyst> gammafunk:  one more thing that I'd kind of like to do for food unification, though it's not necessary before a release
10:12:52 <|amethyst> gammafunk:  I'd like to eliminate about half the chunks in the game, and approximately double their food value
10:13:21 <|amethyst> gammafunk:  because, to balance total amount of food, chunks in food-unification are worth significantly less nutrition than before
10:13:42 <|amethyst> err
10:13:43 <advil> !lg *  starvation / trog s=cv o=.
10:13:43 <Sequell> 2370/9674 games for * (starvation): 32/159x 0.21-a [20.13%], 36/131x 0.20 [27.48%], 35/156x 0.20-a [22.44%], 50/184x 0.19 [27.17%], 97/229x 0.19-a [42.36%], 75/201x 0.18 [37.31%], 151/553x 0.18-a [27.31%], 274/890x 0.17 [30.79%], 185/817x 0.17-a [22.64%], 99/494x 0.16 [20.04%], 65/329x 0.16-a [19.76%], 129/351x 0.15 [36.75%], 50/245x 0.15-a [20.41%], 72/231x 0.14 [31.17%], 100/379x 0.14-a [26.39%]...
10:13:44 <|amethyst> not chunks
10:13:49 <|amethyst> s/chunks/rations/g
10:14:25 <gammafunk> that substitution in both those sentences?
10:14:36 <|amethyst> yeah
10:14:39 <gammafunk> ok
10:14:59 <|amethyst> chunks are unchanged
10:15:31 <gammafunk> ration are placed through food vaults and through normal item gen, the latter is easy to change in the sense of reducing rations
10:15:35 <|amethyst> but rations went from 4400-6500 down to 1900-3400
10:15:45 <gammafunk> however lowering this does increase the numbers of other types of items
10:15:55 <|amethyst> gammafunk:  yeah, I already did a few passes through food vaults
10:16:10 <|amethyst> gammafunk:  to merge everything down to one permafood
10:16:13 <gammafunk> so I guess we'd have to either be careful about what their weight goes to or reduce the number of generated items or both
10:16:47 <gammafunk> yeah, I supose food vaults are actually the easier change since you can just e.g. half the number of rations in the vault or half the rate of placement of said vaults
10:17:09 <gammafunk> for fewer ration items you want to be a bit careful though
10:18:00 <gammafunk> objstat can tell us if we've changed the dist of other types of items, although I don't think I have a summary overall by item class...oh wait maybe we do
10:18:01 <|amethyst> also, doubling ration nutrition and halving ration numbers means Fedhas can require one ration rather than two
10:19:11 <|amethyst> OBJ_FOOD currently has weight 140
10:19:19 <|amethyst> (food_unification doesn't change that, so far)
10:19:52 <|amethyst> 140/1960
10:20:20 <gammafunk> sounds like spreading 70 weight out wouldn't be that bad
10:20:35 <|amethyst> or you'd have to reduce item generation by about 3.5% to compensate
10:21:34 <|amethyst> could assign the extra weight to gold, then change the amount of gold per pile to compensate
10:21:55 <|amethyst> since gold is even more fungible than unified food
10:22:15 <advil> a lower total number might be more susceptible to variance from item gen on the low end
10:22:23 <advil> maybe not a real concern
10:22:36 <gammafunk> you mean like on a per-level basis?
10:23:38 <gammafunk> but yeah with a 3.5% decrease I don't think it could ever be a problem
10:24:06 <advil> I really mean overall, item gen has fairly high variance
10:24:25 <gammafunk> well it does, but the low-end value overall is a very large number
10:24:57 <gammafunk> it'd be a 3.5% decrease of the average, and although I don't know what the decrease on the min would be
10:27:20 <|amethyst> FR: players eat bugs off the ground, so they gain nutrition by stepping onto previously unstepped squares
10:27:26 <|amethyst> I have a branch that may be useful
10:27:27 <gammafunk> I'd imagine that from a player standpoint there wouldn't be a way to percieve this
10:27:30 <|amethyst> %git city-connection
10:27:30 <Cheibriados> 07|amethyst02 * 0.16-a0-3624-gd71b18a: Paint stepped-on squares. 10(2 years, 11 months ago, 5 files, 13+ 1-) 13https://github.com/crawl/crawl/commit/d71b18abbf94
10:27:42 <advil> hmm, can nutrition be deterministically converted into a number of turns (ignoring abilities/spells)?
10:28:36 <gammafunk> I want to say that movement/resting are the same
10:28:45 <gammafunk> but I forget if e.g. melee has an additional cost
10:28:52 <advil> heh yeah i don't know either
10:29:02 <advil> maybe this says something about the nutrition system
10:30:21 <advil> right now the objstat tables for one of the 3 rune games has 92 as the min # of comestibles, I'm just curious what 46 newrations translates to in turns (probably a lot)
10:31:18 <advil> could get this in practice by parsing recent morgue files that count eating
10:31:34 <gammafunk> oh you were talking about rations specifically
10:31:41 <gammafunk> I thought you meant overal item gen
10:32:20 <advil> yeah
10:33:55 <|amethyst> melee does have an extra hunger cost
10:39:23 <|amethyst> fortunately, the total amount of food in the game is wack
10:40:57 <|amethyst> if there's too little food on the first few levels (which is where I'd expect problems from increased variance), first_bread could be adjusted
10:41:10 <|amethyst> (first_ration
10:41:13 <|amethyst> in the branch)
10:41:26 <gammafunk> wow
10:41:30 <gammafunk> "first ration"
10:41:37 <gammafunk> way to ruin flavor again, devs
10:41:47 <|amethyst> anyway, I must be going
10:42:04 <|amethyst> things that need playtesting are spriggans and kobolds/felids
10:42:16 <|amethyst> since they get much less food per ration now
10:42:28 <|amethyst> since they have twice as many as before
12:24:35 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.21-a0-538-gd16b88a (34)
12:45:47 <advil> hm, having format strings for door closing in .des is a bit annoying
13:02:23 <gammafunk> advil: items being under doors are not apportable, I assume?
13:02:32 <advil> no
13:02:38 <advil> I'm actually working on the push version right now
13:03:36 -!- amalloy_ is now known as amalloy
13:03:54 <advil> (no = they are not apportable)
13:04:23 <advil> man this warden code is messy
13:11:04 <advil> my current version will still leave an item under a door if there's nowhere for it to go, because that seems to simplify many things (though it has the obvious downside of leading to rare special cases)
13:12:37 <gammafunk> oh, haha, that's what you meant by "push version"
13:12:45 <gammafunk> I thought that was push like "about to git push"
13:13:18 <advil> heh
13:13:23 <gammafunk> hrm, yeah, ideally items would not ever exist under doors if we're pushing them
13:13:47 <gammafunk> it's too complicated to just destroy the item (after trying to push to all possible adjacent locations)
13:13:48 <advil> well, there are really three possibilities: destroy items, refuse to close doors sometimes, or leave them under doors
13:13:56 <advil> I'm anti-destruction
13:14:22 <gammafunk> well, to be clear here, this is only the case where existing desctruction would apply
13:14:25 <gammafunk> right?
13:14:36 <gammafunk> since we're talking about tiles surrounded by deep water or lava
13:14:55 <advil> well, nothing is destroyed now
13:15:04 <advil> it can come up in a few other cases, where there are only walls
13:15:07 <gammafunk> yes, but you're introducing items under doors just for that corner case
13:15:19 <advil> yeah
13:15:23 <|amethyst> except in the branch you'd be able to do it yourself even without a warden
13:15:58 <advil> for wardens, the door wiil currently refuse to close in those cases
13:15:58 <|amethyst> aha
13:16:08 <|amethyst> could just prevent manually closing the door in that case then
13:16:27 <|amethyst> "There's nowhere for the <item> to go!"
13:16:39 <advil> it's true, that didn't seem much less fiddly to me
13:17:11 <advil> from the player's perspective
13:17:20 <advil> from ours, yeah, it might be less fiddly
13:17:54 <gammafunk> I'm vaguely worried about what problems might arise from no longer being able to assume items aren't in solid features
13:18:23 <advil> well, there may still be problems but as far as I'm aware most are fixed on that branch already
13:18:25 <gammafunk> But I guess I don't have any specific examples of problems
13:20:44 <advil> I think I'm going to push my current version (in a bit, still working) and you all can see what you think then, it would be easy to undo the changes that let items exist under doors in the branch
13:21:18 <advil> or really to cherry-pick one version or the other from this branch
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-538-g9da711b: Render items under closed doors better 10(11 days ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/9da711be2172
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-539-gc5192a1: Allow closing doors on items (rumflump) 10(11 days ago, 1 file, 0+ 6-) 13https://github.com/crawl/crawl/commit/c5192a118583
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-540-g04b8d97: Don't count closed door tiles as visible for items 10(11 days ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/04b8d97e730e
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-541-gac043e4: Fix a fallthrough 10(11 days ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/ac043e4c880a
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-542-gc5b1387: Rework visibility under doors and handle apport 10(11 days ago, 4 files, 22+ 14-) 13https://github.com/crawl/crawl/commit/c5b1387f6302
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-543-g05b5e5c: Have corpse rot only use accessible corpses 10(11 days ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/05b5e5c9e7e0
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-544-gbfdc9c9: Have fungal bloom use only accessible corpses 10(11 days ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/bfdc9c974ec1
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-545-g5c92a09: Push items out of the way when a door is closed 10(36 minutes ago, 3 files, 133+ 12-) 13https://github.com/crawl/crawl/commit/5c92a0908d8f
13:34:38 <Cheibriados> 03advil02 07[door_items] * 0.21-a0-546-g71a73fb: Have vault wardens use simpler item pushing logic 10(4 minutes ago, 1 file, 9+ 18-) 13https://github.com/crawl/crawl/commit/71a73fb0ac2f
13:36:03 <advil> wouldn't be too hard to add a fail case if there are unmoveable items (and not merge 539,546)
13:36:41 <advil> (the rest of the logic still better handles a few special cases that don't involve doors iirc)
13:48:02 <advil> hm maybe they don't
15:56:08 <advil> weird that deep water death code is still in the game somewhere
16:02:40 <Cheibriados> New branch created: door_items_solidity (1 commit) 13https://github.com/crawl/crawl/tree/door_items_solidity
16:02:40 <Cheibriados> 03advil02 07[door_items_solidity] * 0.21-a0-538-g822b319: Push items out of the way when a door is closed, if possible 10(11 days ago, 3 files, 140+ 14-) 13https://github.com/crawl/crawl/commit/822b3192b421
16:03:03 <advil> I gave in
16:14:26 <advil> ??tomb of doro
16:14:26 <Sequell> I don't have a page labeled tomb_of_doro in my learndb.
16:14:29 <advil> ??tomb of dorokhle
16:14:30 <Sequell> tomb of dorokhle ~ tomb of doroklohe[1/2]: Creates rock walls in one or more adjacent squares, destroying traps and displacing items (but not monsters) in the process; now only seen as the effect of the Tomb card and may be cast by {Khufu}.
16:14:39 <advil> I guess that does still exist
16:56:19 <travis-ci> The build passed. (door_items_solidity - 822b319 #9028 : advil): https://travis-ci.org/crawl/crawl/builds/317116008
17:01:25 <Cheibriados> 03advil02 07[door_items_solidity] * 0.21-a0-539-g191ebc4: Refactor imprison to use new push code 10(18 minutes ago, 3 files, 38+ 26-) 13https://github.com/crawl/crawl/commit/191ebc496c33
17:01:25 <Cheibriados> 03advil02 07[door_items_solidity] * 0.21-a0-540-g608e3f8: Use new push code for abyss generation 10(56 seconds ago, 1 file, 3+ 9-) 13https://github.com/crawl/crawl/commit/608e3f8a10ed
17:03:27 <advil> the door sealing code is such a delicately balanced piece of work I'm afraid of trying to clean it up
17:51:56 <travis-ci> The build has errored. (door_items_solidity - 608e3f8 #9029 : advil): https://travis-ci.org/crawl/crawl/builds/317138173
18:17:48 <Lantell> Unstable branch on underhound.eu updated to: 0.21-a0-538-gd16b88afdd (34)
18:17:57 <Cheibriados> 03advil02 07[door_items_solidity] * 0.21-a0-541-ga2a8c5a: Use refactored push code for door sealing 10(46 seconds ago, 2 files, 71+ 165-) 13https://github.com/crawl/crawl/commit/a2a8c5ad5113
19:17:52 <travis-ci> The build has errored. (door_items_solidity - a2a8c5a #9030 : advil): https://travis-ci.org/crawl/crawl/builds/317165432
21:30:29 -!- mikee__ is now known as mikee_
22:11:49 -!- Evablue_ is now known as Evablue
22:27:51 -!- Evablue__ is now known as Evablue