00:13:44 <Cbrotell> Unstable branch on cbro.berotato.org updated to: 0.33-a0-764-g13ce909368 (34)
00:32:30 <NotCheibriados> 03DracoOmega02 {regret-index} 07* 0.33-a0-765-g12f01c111f: Refactor/unify immune to sleep/paralysis and blame handling 10(3 days ago, 6 files, 54+ 42-) 13https://github.com/crawl/crawl/commit/12f01c111f2d
00:32:30 <NotCheibriados> 03DracoOmega02 {regret-index} 07* 0.33-a0-766-gf4f8fbaca6: New status effect: Vexed 10(3 days ago, 37 files, 204+ 17-) 13https://github.com/crawl/crawl/commit/f4f8fbaca64b
00:32:30 <NotCheibriados> 03regret-index02 07* 0.33-a0-767-g6ac722a880: Split sphinxes into earlier marauders and later guardians 10(68 minutes ago, 56 files, 288+ 146-) 13https://github.com/crawl/crawl/commit/6ac722a88032
00:46:03 <cgettys> Uh-oh, yet more content and chanes :P
00:46:07 <cgettys> changes
00:46:11 <cgettys> why can't I type :D
00:55:40 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.33-a0-767-g6ac722a880
04:32:15 <Lantell> Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-5261-gd9800d219b
05:29:06 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.33-a0-767-g6ac722a (34)
07:59:44 <NotCheibriados> 03dolorous02 07* 0.33-a0-768-gbf5a252ee1: Fix comment typo. 10(55 seconds ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/bf5a252ee1ed
09:22:17 <Discord|> <09h​ellmonk> crossposting from roguelike discord
09:22:17 <Discord|> <09h​ellmonk> https://cdn.discordapp.com/attachments/747522859361894521/1330210664886571132/wtf.png?ex=678d26b8&is=678bd538&hm=2c54af179e36b72b64a2b96389a416cc7237302bc3ef13a755a7bc55aaca4fb4&
09:25:05 <Discord|> <09h​ellmonk> seems like there is a KITEM:   . = any weapon randart line that got added (for testing I assume)
09:31:18 <NotCheibriados> 03Hellmonk02 {GitHub} 07* 0.33-a0-769-gdcb3d5a981: don't spawn 500 randarts in hall_of_zot 10(66 seconds ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/dcb3d5a9812c
09:38:48 <geekosaur> lol
09:40:29 <Discord|> <09h​ellmonk> if your new thing doesn't break something in trunk it's not a real feature tbh
09:47:15 <Discord|> <08n​icolae> not many github projects include phrases like "apotropaic variety of sphinx" in their commit messages
10:14:48 <Discord|> <09g​ammafunk> woopsie
10:19:27 <Discord|> <09g​ammafunk> @regret-⸸nde※ Looks like you left a present for players in the hall of zot in 6ac722a (fixed just now)
11:19:43 <Cbrotell> prawnwizard (L23 DsMo) ERROR: range check error (20 / 20) (Depths:1)
11:38:43 <Discord|> <09g​ammafunk> !crashlog
11:39:06 <Discord|> <04C​erebot> 29773. prawnwizard, XL23 DsMo, T:36193 (milestone): https://cbro.berotato.org/morgue/prawnwizard/crash-prawnwizard-20250118-181935.txt
11:41:06 <Discord|> <09g​ammafunk> is this a known crash?
13:06:28 <Discord|> <03i​mplojin> looks like it might be blowing past the cap on tile overlays by ... trying to pack shoal waves while in depths? somehow?
14:23:48 <cgettys> _pack_wall_shadows
14:24:23 <cgettys> "trying to pack shoal waves while in depths? somehow?" - wouldn't that have th shoal waves function? unless the functions are identical and fold together, gotta love optimizing compilers
14:25:00 <Discord|> <04d​racoomega> I believe it may be Disaster Area's fault
14:25:17 <cgettys> Me too
14:25:18 <Discord|> <04d​racoomega> By hitting the same spot over and over and over again
14:25:40 <Discord|> <04d​racoomega> (In which case, I am surprised this hasn't come up more often, but maybe the odds of hitting one spot that many times is still very low)
14:27:50 <cgettys> Is hitting the same spot even expected to do anything?
14:29:00 <cgettys> ah, yeah, it is
14:29:02 <cgettys> "Each explosion is discouraged from overlapping. "
14:29:03 <Discord|> <03i​mplojin> cgettys: we have various optimizations going on here and it's unclear to me where the error lies
14:29:11 <cgettys> discouraged != cannot happen :D
14:30:14 <cgettys> And if you're above 16.7 invocations, you can get 2 tile radius
14:30:33 <cgettys> which would be 3x3?
14:31:03 <cgettys> // TODO: should count get a cap proportional to targets.size()?
14:32:09 <Discord|> <03i​mplojin> %git afd660ec2
14:32:09 <Discord|> <04C​erebot> Aidan Holm * 0.26-a0-128-gafd660ec27: Move pack_cell_overlays() call to vbuf load time (4 years, 7 months ago, 2 files, 12+ 9-) https://github.com/crawl/crawl/commit/afd660ec271a
14:32:19 <Discord|> <03i​mplojin> %git 7d2dc893a16f
14:32:20 <Discord|> <04C​erebot> advil * 0.26-a0-1139-g7d2dc893a1: Do basic deduplication when rendering tile overlays (4 years, 1 month ago, 1 file, 6+ 1-) https://github.com/crawl/crawl/commit/7d2dc893a16f
14:32:47 <cgettys> // TODO: should count get a cap proportional to targets.size()?
14:32:55 <Discord|> <03i​mplojin> %git f17e09bd7
14:32:55 <cgettys> whoops didn't mean to send that twice
14:32:56 <Discord|> <04C​erebot> DracoOmega * 0.32-a0-1893-gf17e09bd7d: Make tides look a little better around some pseudo-walls (regret-index) (6 months ago, 1 file, 12+ 1-) https://github.com/crawl/crawl/commit/f17e09bd7d5f
14:33:00 <cgettys> Invocations     |          X          |   1   |    20    |   24  |  11491 | 476/1085
14:33:09 <cgettys> invocation level 24
14:34:02 <cgettys> you.skill_rdiv(SK_INVOCATIONS, 1, 2),
14:34:05 <cgettys> would be up to 12
14:34:19 <cgettys> random2avg(you.skill(SK_INVOCATIONS, 2), 2)
14:36:06 <Discord|> <04d​racoomega> Oh, I didn't even know about trying to not bother with rending the same overlay multiple times (though this makes sense)
14:36:13 <Discord|> <04d​racoomega> However, it looks like it just checks if they are sequentially the same
14:36:23 <Discord|> <04d​racoomega> Disaster area can be random
14:36:25 <Discord|> <04d​racoomega> Oh!
14:36:37 <Discord|> <04d​racoomega> That means something else is probably relevant
14:37:36 <Discord|> <04d​racoomega> The air and ice rolls from Upheaval (also used by Disaster Area) both used to have the same tile on this:
14:37:40 <Discord|> <04d​racoomega> %git 7e861bf89c551b7f90f3e74c08fb6e9460dcb891
14:37:42 <Discord|> <04C​erebot> regret-index * 0.33-a0-606-g7e861bf89c: Loose, miscellaneous tiles tweaks and additions (3 weeks ago, 29 files, 71+ 29-) https://github.com/crawl/crawl/commit/7e861bf89c55
14:38:04 <Discord|> <04d​racoomega> Which could make this crash more likely to trigger than before
14:39:18 <Discord|> <04d​racoomega> Just probably a very rare situation where enough hits of different elements alternatingly hit the one spot
14:42:30 <Discord|> <04d​racoomega> Really, probably the most sane approach is to start erasing overlays from the bottom, if we're going to hit a limit
14:42:41 <Discord|> <04d​racoomega> It's not like one is likely to be able to make them out usefully anyway
14:42:53 <Discord|> <04d​racoomega> (I cannot immediately tell where this is supposed to happen, though)
14:52:08 <Discord|> <04d​racoomega> Possibly _add_overlay in tilecell.cc should actually check for overflows and do cleanup in that case?
14:53:58 <cgettys> are the overlays rendered top to bottom? or are they just alpha blended or something?
14:54:22 <cgettys> e.g. is the base tile rendered separately then overlays, or it jus tall one big stack
14:54:42 <Discord|> <04d​racoomega> I mean, I assume base tile is rendered first, but I'm pretty sure it's just bottom to top in general
14:54:43 <Discord|> <03i​mplojin> i'm not even sure if this overlay limit is a real limit or was just picked as an arbitrary "well this is a lot more than we have now" number back in ~2011
14:54:50 <Discord|> <04d​racoomega> That too
14:55:03 <Discord|> <04d​racoomega> I mean, it seems very hard to hit outside of something like Disaster Area, and unlikely even then
14:55:05 <cgettys> another interesting thing abou tthis
14:55:07 <Discord|> <04d​racoomega> Or we'd have seen this more often
14:55:12 <cgettys> the weighting
14:55:18 <cgettys> oh jeez
14:55:23 <cgettys> this function is super not efficient in edge cases
14:55:29 <cgettys> idk if it's super not efficient enough to matter
14:55:59 <cgettys> it's O(count*target.size())
14:57:12 <Discord|> <04d​racoomega> I mean, most tiles don't have many overlays in almost all cases
14:57:26 <cgettys> But then again, modern cpus, count can't be more than uh, 27*2 = 54
14:57:59 <cgettys> I don't mean rendering, I mean qazlal_disaster_area
14:58:46 <Discord|> <04d​racoomega> Oh, well, Disaster Area is a bit of a mess, but I can't imagine performance matters on something called exactly once in response to a player action (which then takes a long time to animate afterward)
14:59:25 <Discord|> <04d​racoomega> (There have been complaints about the algorithm being somewhat nonsense in terms of unexpected/unintuitive biases in where it hits things, but that's quite a separate concern)
14:59:42 <cgettys> Probably could fix both at once
15:00:06 <cgettys> Wait
15:00:17 <cgettys> _upheaval_radius = pow / 100 + 1
15:00:20 <Discord|> <04d​racoomega> Well, 'fixing' the algorithm involves figuring out how it mechanically 'ought' to work
15:00:39 <cgettys> pow = 6 * invocations
15:00:51 <Discord|> <04d​racoomega> Which is much less straightforward than actually implementing that
15:01:19 <cgettys> so 27 * 6 / 100 = 1.62 + 1 = 2.62
15:01:36 <cgettys> maybe another place it'd feel beteter to probabilistically round up or not
15:01:47 <cgettys> but yeah, deciding mechanisms probably the hard part
15:02:14 <Discord|> <04d​racoomega> (For instance, there'd been talk about mechanically decoupling its properties from Upheaval's so that it didn't scale in quite such a nonlinear fashion - it's quite underwhelming until you have very high invo, to the point that less informed players may not realize how good it even is at high invo)
15:02:55 <cgettys> stupid q uestion, is radius inclusive or exclusive of center?
15:03:31 <cgettys> e.g. LOS_RADIUS = 8, does that mean if C_SQUARE, it's a 17x17 square?
15:03:43 <cgettys> eh let me just answer my own question :D
15:04:30 <cgettys> Or just assume it's 17x17 for purpose of back of envelope
15:05:17 <cgettys> if it's 17x17, there are at most 289 positions
15:05:36 <cgettys> that it can be centered on
15:06:17 <cgettys> Honestly, I'd be tempted to build a hashmap (actual or just a 289 element array)
15:06:25 <cgettys> Decouple the rendering from the impact and call it a day
15:06:36 <cgettys> only render clouds once per each tile hit that way
15:10:23 <Discord|> <04d​racoomega> They animate in sequence, though, so you see it incrementally filling up
15:10:48 <Discord|> <04d​racoomega> (And seeing spaces repeatedly hit is relevant)
15:12:29 <cgettys> Ah, good point
15:15:48 <cgettys> The worst case I think would be if you rolled what, 2 54s for an average of 54
15:15:50 <cgettys> for count I mean
15:16:22 <cgettys> Each target can only be hit once directly, if I read this right
15:18:13 <cgettys> so the worst case would be what, a 3x3 square next to the player, where all 9 squares get hit, hitting the middle tile 9 tiles? or is 2 tile radius bigger than I think
15:19:04 <cgettys> oh wait, but multiple effects can stack
15:19:59 <cgettys> So 9 * 4 = 36 would be the max number of hits on one tile, I think
15:20:43 <cgettys> for the beams anyway
15:51:08 <cgettys> I'm getting a bit sidetracekd here, but some things don't make sense to me
15:51:30 <cgettys> why does tile_flavour have both floor_idx _and_ floor, and didto for wall and feat
15:51:35 <cgettys> for example
15:53:59 <Discord|> <04d​racoomega> The floor/floor_idx is weird and complex to me, and even though I've had to deal with it, I only partially understand also
15:54:30 <Discord|> <04d​racoomega> Has to do with tile variants? But also has to do with how it's marshalled/unmarshalled??
15:54:50 <Discord|> <04d​racoomega> Since some of them can change if you save/reload a level (I fixed a couple bugs related to that at some point in the past)
15:55:05 <Discord|> <04d​racoomega> (Really, I basically hate huge chunks of the rendering system)
15:55:22 <Discord|> <04d​racoomega> And almost every time I deal with it for anything non-trivial, it's just miserable to work with
16:03:43 <cgettys> the thing is there's places where a tileidx_t (uint64) is shoved in there
16:03:53 <cgettys> into an unsigned short
16:03:56 <cgettys> which is uint16
16:04:01 <cgettys> so one of those types is probably wrong?
16:40:52 <Lantell> Unstable branch on underhound.eu updated to: 0.33-a0-769-gdcb3d5a981 (34)
17:26:48 <NotCheibriados> 03dolorous02 07* 0.33-a0-770-ga9c169229f: Fix inconsistent ellipses. 10(5 minutes ago, 4 files, 4+ 4-) 13https://github.com/crawl/crawl/commit/a9c169229f9d
17:26:48 <NotCheibriados> 03dolorous02 07* 0.33-a0-771-g289bc3cdf1: Add Zonguldrok messaging for unrand hats. 10(7 minutes ago, 2 files, 18+ 0-) 13https://github.com/crawl/crawl/commit/289bc3cdf16f
17:29:26 <Discord|> <04d​racoomega> @dolorous_84348 Most of those aren't actually 'inconsistent'. It's standard for ellipsis that form the end of a sentence to be 4 dots. (Technically this is 3 dots for the ellipsis and then 1 dot for the actual period, but effectively it's 4 in a row)
17:30:13 <Discord|> <04d​racoomega> So I... didn't know what to say. but I didn't know what to say....
17:32:11 <Discord|> <09g​ammafunk> I love the idea that we need flavor text for zonguldrok with each unrand hat
17:33:22 <Discord|> <09g​ammafunk> hrm, I feel we should possibly also have zonguldruk be ok with a mundane hat if it's max enchanted
17:33:42 <Discord|> <04d​racoomega> imo, he can be picky
17:33:59 <Discord|> <04d​racoomega> Must be fancy!
17:34:01 <Discord|> <09g​ammafunk> well, is a +2 hat actually worse than an ego +0 hat?
17:34:02 <Discord|> <04d​racoomega> Not merely well-made
17:34:25 <Discord|> <09g​ammafunk> I guess if the lore is about its specialness, yeah
17:34:43 <Discord|> <09g​ammafunk> ...but what if we do add a brand armour scroll...
17:35:33 <Discord|> <04d​racoomega> No need to future-proof one dead necromancer, imo 😛
17:38:22 <Discord|> <08n​icolae> obviously we do, it just makes sense. perhaps i will try to think of a few lines for the other hats.
17:38:58 <Discord|> <09g​ammafunk> ??best hat
17:38:59 <Discord|> <04C​erebot> crown of dyrovepreva[1/2]: +3 unrandart hat {rElec Int+2 SInv}. A large crown of dull bronze, set with a dazzling array of gemstones. This is the best hat.
17:39:02 <Discord|> <09g​ammafunk> !learn q best_hat
17:39:03 <Discord|> <04C​erebot> best hat[1/1]: do {.echo $(!randhat) This is the $(if (> (rand 2) 0) "best" "worst") hat.}
17:39:10 <Discord|> <09g​ammafunk> !cmd randhat
17:39:10 <Discord|> <04C​erebot> Command: !randhat => .echo $(replace _ ' ' (ldb (randnth (split & (=randhat.list)))))
17:39:16 <Discord|> <09g​ammafunk> !cmd =randhat.list
17:39:16 <Discord|> <04C​erebot> Command: =randhat.list => .echo hat of the alchemist&hat of the bear spirit&hat of the high council&hat of pondering&crown of dyrovepreva&mask of the dragon&maxwell's etheric cage
17:39:36 <Discord|> <09g​ammafunk> needs updating
17:40:40 <Discord|> <08n​icolae> Now this is a fine visage to cover a skull of my caliber! (mask of the dragon)
17:40:40 <Discord|> <09g​ammafunk> also really wish more people would use sequellese lists and hashes instead of rolling their own string functions
17:41:07 <Discord|> <09g​ammafunk> fine, but it feels pretty generic relative to the theme of the mask
17:41:43 <Discord|> <08n​icolae> yeah...
17:44:37 <Discord|> <08n​icolae> Dragon scale armour for the head-sized customer. Wonderful!
17:46:38 <Discord|> <08n​icolae> Whoever or whatever a Dyrovepreva is, they do make a fine crown.
17:47:16 <Discord|> <09g​ammafunk> ??mask of the dragon
17:47:17 <Discord|> <04C​erebot> mask of the dragon[1/1]: +0 hat with SInv Will+ Slay+3.
17:47:39 <Discord|> <09g​ammafunk> sadly I think it's extremely not dragon scale armour
17:47:51 <Discord|> <09g​ammafunk> something about "dragon slaying"?
17:47:56 <Discord|> <09g​ammafunk> like a play on that
17:48:00 <Discord|> <08n​icolae> ...if you've got your regular head slot melded, and then you get the skull of zonguldruk, and then you get the fungal fisticloak, can you put the fungal hat part on zong
17:48:26 <Discord|> <09g​ammafunk> do orbs meld?
17:48:29 <Discord|> <09g​ammafunk> they do don't they
17:48:37 <Discord|> <08n​icolae> er, i meant cursed
17:49:06 <Discord|> <08n​icolae> but mostly i was wondering, how does zong interact with multi-slot unrands that need a head
17:49:13 <Discord|> <03i​mplojin> Today, on good idea bad idea: Getting rid of staircases.
17:51:06 <Discord|> <08n​icolae> This hat smells like a thousand failed laboratory experiments.
17:51:46 <Discord|> <09g​ammafunk> !lg * status=no_stairs
17:51:49 <Discord|> <08n​icolae> Now THIS is a crown fit for a head such as mine! (vainglory)
17:51:53 <Discord|> <04C​erebot> 1. cadi the Fencer (L18 BaIE of Sif Muna), slain by a radroach on Spider:4 on 2025-01-14 10:02:17, with 206592 points after 40268 turns and 1:10:58.
17:56:39 <Discord|> <08n​icolae> Our dark harvest shall come from the shadows, scion! (hood of assassin)
17:57:30 <Discord|> <08n​icolae> I can't bear puns, scion. Keep that in mind. (bear spirit)
17:58:59 <Discord|> <08n​icolae> Or: I can't stand puns, scion. Bear th-- keep that in mind.
17:59:43 <Discord|> <08n​icolae> Even in death, I can feel the rage of this powerful spirit! (bear again)
18:01:30 <Discord|> <08n​icolae> It feels a bit... flimsy, but protective nonetheless. (alchemist)
18:02:07 <Discord|> <08n​icolae> bear spirit: WHOO! I am DEFINITELY not afraid of the fucking POLICE right now!
18:02:24 <Discord|> <08n​icolae> okay i'm done for now.
18:20:34 <Discord|> <06d​olorous_84348> I thought ending a sentence with an ellipsis meant the sentence wasn't actually ended. At least the change is easily undone.
18:22:17 <NotCheibriados> 03dolorous02 07* 0.33-a0-772-g4d2cd2bd01: Revert "Fix inconsistent ellipses." 10(80 seconds ago, 4 files, 4+ 4-) 13https://github.com/crawl/crawl/commit/4d2cd2bd0113
18:41:57 <Discord|> <04d​racoomega> (I double-checked with a bunch of sites to see that I wasn't misremembering or misunderstanding this, before I said it, but what I searched up did seem to agree with what I said there)
18:47:08 <geekosaur> what you said is correct
19:03:58 <Discord|> <08n​icolae> i agree, ellipses at the end of a sentence get four dots, one for the period.
19:12:18 <Discord|> <06d​olorous_84348> Okay.
19:15:09 <Discord|> <06d​olorous_84348> Is there a distinction to be made between a complete sentence and what is technically a complete sentence but could be longer, in the case of a speaker's trailing off?
19:15:39 <Discord|> <04d​racoomega> 'Trailing off into silence' counts as the end of a sentence
19:16:00 <Discord|> <04d​racoomega> Basically 'Is the letter afterward capitalized or not?'
19:16:13 <Discord|> <04d​racoomega> (More or less)
19:17:21 <Discord|> <06d​olorous_84348> Ah. I'm just trying to figure out whether the line I added, "I must contemplate what to do next..." needs a fourth period or not, because it's trailing off at the end.
19:17:39 <Discord|> <04d​racoomega> I think 4 would be correct
19:18:06 <Discord|> <06d​olorous_84348> Okay, I can do that.
19:21:01 <NotCheibriados> 03dolorous02 07* 0.33-a0-773-g379047849a: Fix an ellipsis at the end of a sentence. 10(3 minutes ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/379047849a42
19:23:30 <cgettys> Ok, I have a branch with changes to the overlay stuff now for that crash from earlier
19:23:33 <cgettys> no clue if it works yet :D
19:25:52 <Discord|> <04d​racoomega> viewmap_flash_colour() has already been changed in trunk again, incidentally
19:26:05 <Discord|> <04d​racoomega> Oh
19:26:09 <Discord|> <04d​racoomega> Wrong overlay stuff
19:26:12 <Discord|> <04d​racoomega> My mistake ^^;
19:26:43 <Discord|> <04d​racoomega> (I was going to say that it's a hard crash to test, but I guess it's not hard if one changes the code to force piling a ton of stuff into one place)
19:26:58 <cgettys> That's not the approach I was taking at all
19:27:05 <cgettys> though I should
19:27:11 <cgettys> Should spam the chat in a minute :D
19:27:26 <NotCheibriados> New branch created: pull/4253 (96 commits) 13https://github.com/crawl/crawl/pull/4253
19:28:38 <NotCheibriados> 03gammafunk02 {dolorous} 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-1-g67e854b3c2: fix: Add some missing KMASK to a vault (dilly) 10(5 months ago, 1 file, 5+ 3-) 13https://github.com/crawl/crawl/commit/67e854b3c248
19:28:38 <NotCheibriados> 03dolorous02 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-2-g34e853cafd: Adjust kraken tentacle msgs to not mention water. 10(5 months ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/34e853cafd8e
19:28:38 <NotCheibriados> 03Medrano8302 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-3-ga61a675df6: Android: Add a full screen option to the launcher 10(5 months ago, 5 files, 76+ 2-) 13https://github.com/crawl/crawl/commit/a61a675df6d1
19:28:38 <NotCheibriados> 03dolorous02 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-4-gb28a3c84aa: Don't give gladiators default kite shield tiles. 10(5 months ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/b28a3c84aa6f
19:28:38 <NotCheibriados> 03Medrano8302 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-5-ge0a8f3941f: Fix Android mousewheel scrolling 10(5 months ago, 2 files, 8+ 2-) 13https://github.com/crawl/crawl/commit/e0a8f3941f23
19:28:38 <NotCheibriados> 03nlavsky02 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-6-gfe437301d8: text: mention that Soul Splinter inflicts weakness (MIC132) 10(5 months ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/fe437301d802
19:28:38 <NotCheibriados> 03PleasingFungus02 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-7-g3eb36626ba: Fix: don't stairs into the abyss (CarefulOdds) 10(5 months ago, 1 file, 1+ 1-) 13https://github.com/crawl/crawl/commit/3eb36626ba1e
19:28:38 <NotCheibriados> 03gammafunk02 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-8-g0b4fe110a3: fix: Don't mark certain spells as self enchantment 10(5 months ago, 3 files, 25+ 31-) 13https://github.com/crawl/crawl/commit/0b4fe110a30c
19:28:38 <NotCheibriados> 03Medrano8302 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-9-g3b3b7eff08: Fix Android touchpad scrolling 10(5 months ago, 1 file, 10+ 1-) 13https://github.com/crawl/crawl/commit/3b3b7eff0895
19:28:38 <NotCheibriados> 03Medrano8302 07https://github.com/crawl/crawl/pull/4253 * 0.32-b1-10-g6bbcbaf917: Fix Android mouse detection 10(5 months ago, 2 files, 2+ 6-) 13https://github.com/crawl/crawl/commit/6bbcbaf91781
19:28:38 <NotCheibriados> ... and 86 more commits
19:28:39 <cgettys> I increased max overlays to 32 (not sure how I feel about that, tempted to see if with the new logic we can cut back to 16 instead of going up to 32)
19:28:39 <cgettys> made it check all overlay sfor duplicates, not just the most recent
19:28:39 <cgettys> And if 32 overflows too, it now will spam console with warnings instead of crashing.
19:28:53 <cgettys> Arg
19:28:54 <cgettys> sorry
19:28:59 <cgettys> I forked off of 0.31
19:29:02 <cgettys> I forked off of 0.32 rather
19:29:04 <cgettys> did not expet
19:29:05 <cgettys> sigh
19:29:39 <Discord|> <04d​racoomega> I can't imagine we even come close to 20 if they're actually pruning duplicates, tbh
19:29:46 <Discord|> <04d​racoomega> (I could be wrong, but would be surprised)
19:30:00 <Discord|> <04d​racoomega> There are very few effects that use multiple different overlays repeatedly
19:30:48 <Discord|> <04d​racoomega> (Fulsome Fusillade actually precalculates affected spots such that there isn't almost any overdraw, even)
19:30:59 <Discord|> <04d​racoomega> Although it could only have done 3 at once anyway, I guess
19:31:23 <cgettys> RE: viewmap_flash_colour - nice, I'll fix up https://github.com/crawl/crawl/pull/4224 to avoid conflicts later
19:31:42 <cgettys> waves do a bunch of stuff, so I'm not 100% sure
19:31:54 <cgettys> I think wec ould reduce to 16 and see but idk
19:32:02 <cgettys> 20 is awkward/results in worse codegen
19:32:11 <cgettys> I went back and forth on things and nerd-sniped myself a bit
19:33:12 <cgettys> Clang is smart enough to fully vectorize and unroll what I wrote on O2
19:35:12 <cgettys> GCC is grumpy but spits o ut ok non-vectorized code
19:36:59 <cgettys> but given GCC won't vectorize it, would be ideal to avoid going to 32 bytes
19:41:29 <NotCheibriados> 03regret-index02 07* 0.33-a0-774-ge11c45b518: Properly grant more body armour to deep elf / high priest monsters 10(70 minutes ago, 1 file, 13+ 6-) 13https://github.com/crawl/crawl/commit/e11c45b518cf
19:41:29 <NotCheibriados> 03regret-index02 07* 0.33-a0-775-gd154b68b10: Tactile mutation icons (CanofWorms) 10(7 minutes ago, 6 files, 6+ 2-) 13https://github.com/crawl/crawl/commit/d154b68b10df
19:44:15 <cgettys> Jeez, GCC is picky
19:44:34 <cgettys> I can get it to unroll at O3, but still not vectorize
20:13:16 <cgettys> D'oh, I'm a dummy
20:13:23 <cgettys> std::find would be perfeclty find here
20:25:54 <cgettys> Wow, the codegen is... not great for that
20:26:06 <cgettys> I mean, not dreadful, but I was expecting clang to still autovectorize nicely
20:30:57 <cgettys> Oh, right, yet again, I'm a dummy
20:31:55 <cgettys> if I do auto end = dngn_overlay.end(); bool found= std::find(dngn_overlay.begin(), end, tileidx) != end;
20:32:04 <cgettys> clang unrolls nicely just like the open coded version
20:32:07 <cgettys> auto end = dngn_overlay.begin() + num_dngn_overlay;
20:32:20 <cgettys> leads to just unrolling w/o vectorization
20:32:26 <cgettys> because loop bound not constant
20:32:27 <cgettys> duh
20:50:14 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4253 * 0.32.1-6-g8945a0d392: fix: overflow of overlays with Qazlal's Disaster area (prawnwizard) 10(5 hours ago, 6 files, 156+ 90-) 13https://github.com/crawl/crawl/commit/8945a0d39244
20:54:30 <NotCheibriados> New branch created: pull/4254 (1 commit) 13https://github.com/crawl/crawl/pull/4254
20:54:32 <NotCheibriados> 03Cgettys02 07https://github.com/crawl/crawl/pull/4254 * 0.33-a0-764-ge6a09ff5d5: fix: overflow of overlays with Qazlal's Disaster area (prawnwizard) 10(5 hours ago, 6 files, 156+ 90-) 13https://github.com/crawl/crawl/commit/e6a09ff5d573
21:25:25 <NotCheibriados> 03dolorous02 07* 0.33-a0-776-g802b72d441: Add a few Xom messages for fountains of eyes. 10(4 minutes ago, 2 files, 8+ 0-) 13https://github.com/crawl/crawl/commit/802b72d44139
22:11:41 <cgettys> I'm having a hard time repro'ing that crash even without my fix
22:12:04 <cgettys> looks like electrocution brand also involved
22:13:40 <cgettys> Oh, and FogMachine?
22:13:46 <cgettys> https://github.com/crawl/crawl/blob/802b72d44139372eae173a53d903e7b86e4b20dc/crawl-ref/source/dat/dlua/lm_fog.lua#L70
23:35:18 <Gretell> Unstable branch on crawl.develz.org updated to: 0.33-a0-776-g802b72d441 (34)
23:58:12 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.33-a0-776-g802b72d441