01:53:22 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.26-a0-812-g31a745ddcd
02:18:47 <Cbrotell> Unstable branch on cbro.berotato.org updated to: 0.26-a0-812-g31a745ddcd (34)
02:53:02 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.26-a0-812-g31a745ddcd
03:30:51 <Kelbell> Fork (bcrawl) on crawl.kelbi.org updated to: 0.23-a0-3681-ga1082b90cb
09:12:16 <advil> lol I don't think I thought through the consequences of writing a script to automatically test for connectivity
09:18:20 <gammafunk> if you're not connected, you can't tell anyone you don't have a connection?
09:18:50 <advil> pretty much
09:19:00 <gammafunk> a classic
09:19:36 <advil> (the consequence being there's not a small number of vaults that turn out to have minor connectivity issues that now I guess I've set myself up to fix)
09:20:38 <gammafunk> oh lol
09:20:49 <gammafunk> I actually though you meant network connectivity
09:20:52 <advil> haha
09:21:47 <gammafunk> yeah, that issue about vaults being default opaque hiding issues like that was kind of in the back of my mind
09:21:55 <advil> but it's still basically right, e.g. dpeg_arrival_arbitrary_a would just quietly veto on 1/50 placements and no one would ever know
09:22:00 <gammafunk> because with the old system you didn't have to think about it as hard
09:22:31 <gammafunk> finding issues like that in vaults sounds very good though
09:22:33 <advil> !vault dpeg_arrival_arbitrary_a
09:22:34 <Sequell> 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/arrival/small.des#L1545
09:22:51 <gammafunk> oh right
09:23:30 <advil> there's actually two issues, both really hard to spot without a script
09:23:30 <gammafunk> the shotgun substitution that sometimes makes zones?
09:23:30 <gammafunk> I assume on ' glyph
09:23:30 <advil> yeah, there's one grid on the top that's a problem in the middle, rarely
09:23:35 <advil> also there's a blank space in the lower middle that is easy to miss
09:23:50 <gammafunk> oh huh, so there is
09:24:14 <advil> which will force a veto if the vault places over the wrong thing (though tbh I'm not 100% sure that's an issue with arrival vaults, but it's at least in principle an issue)
09:24:15 <gammafunk> fr: vim-style des highlighting for github
09:25:40 <gammafunk> hard for me to see the problematic other area
09:26:15 <gammafunk> I also really dislike multi-part substs like SUBST:   ' : . x ':40, '=.x
09:26:21 <gammafunk> where the randomization happens twice
09:27:09 <gammafunk> I guess it might be necessary in this case because of the : subst
09:28:10 <advil> in the second row, there's a square right above the middle of `xxx` which can become disconnected
09:28:27 <gammafunk> oh I see
09:29:07 <gammafunk> yeah that string of '
09:29:07 <advil> it really is like a 1 in 50 chance or something
09:29:07 <gammafunk> yep, that's a common problem with that style of subst
09:29:10 <gammafunk> "ruining" the walls but you put too many contiguous glyphs that can get ruined
09:31:09 <advil> yeah, I expect to see many of those in my future if I follow through with this
09:38:03 <advil> !vault hangedman_elven_crosses_minicolossus
09:38:04 <Sequell> 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/branches/elf.des#L1183
09:39:03 <advil> hm, on this one the exterior doors can go to walls and it becomes unplacable, but this looks fairly intentional to me
09:59:00 <advil> hahaha there's a vault that places a fire dragon at lair entrance but also apparently can't place because it almost always puts the lair entrance on an island (it's not completely impossible though)
10:06:08 <advil> on discord ebering advocates for this vault
10:06:55 <advil> but using vetoes to control probability is not ideal
10:07:26 <advil> !vault minmay_lair_entry_lava
10:07:27 <Sequell> 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/branches/lair.des#L291
10:10:16 <advil> one fix would be to make this vault better connected but not always place the dragon
10:13:16 <advil> %git 9a2f62deabe
10:13:16 <Cheibriados> 07gammafunk02 * 0.23-a0-92-g9a2f62d: Adjust a Lair entry vault (minmay) 10(2 years, 3 months ago, 1 file, 6+ 4-) 13https://github.com/crawl/crawl/commit/9a2f62deabe5
10:13:25 <advil> !blame gammafunk
10:13:26 <Sequell> I pronounce gammafunk... Guilty!
10:14:03 <advil> I guess I might just skip this one for now
11:20:31 <gammafunk> advil: yeah, we discussed that lair fire dragon entrance; obviously I didn't catch that error I made with the glyphing
11:21:30 <gammafunk> my vote is to just fix the glyphing, and looking at cumulative vault weights, it will be fine, but I can double check in mapstat that it's not placing too frequently
11:21:38 <gammafunk> because larger vaults tend to place more, so if it'd effectively place 1 in 10 times or something relatively high, that'd be bad
11:21:46 <gammafunk> however there's no way it's that frequent
11:23:42 <advil> that seems fine, I do think that of lair entrances it'll place relatively easily with that fixed
11:23:59 <advil> but it may well not be very frequent even so
11:39:24 <advil> gammafunk: do you know if there's a use case for putting @ internal to a map? The docs imply you shouldn't do this but I have found a few cases where this happens
11:41:59 <gammafunk> advil: not aware of one, no; could that be useful for, say, a donut-shaped map that has space (no non-vault) in the hole?
11:42:20 <gammafunk> s/no non-vault/so, non-vault/
11:42:33 <advil> hm, maybe, that doesn't match the cases I've seen though
11:42:42 <advil> !vault nicolae_shoals_arrival_archipelago
11:42:42 <Sequell> 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/branches/shoals.des#L568
11:42:52 <advil> I think this one might just be a mistake
11:43:59 <advil> maybe should just be `a:@` or something
11:43:59 <advil> and `de:.`
11:45:00 <gammafunk> ah, yeah
11:45:05 <gammafunk> maybe just a typo
11:45:17 <advil> there was one other I saw that also looked potentially like a typo, can't remember what
11:46:35 <gammafunk> I wonder if one of d or e should be w
11:47:27 <gammafunk> because nothing is really getting cut off if both d and e become floor, which seems to be the intent
11:47:44 <gammafunk> that something gets isolated from something else, but everything remains connected to something
11:48:12 <advil> I think that if there are explicit @s then the floor ones could get cut off depending on the surroundings
11:48:14 <gammafunk> but yeah, seems unintentional that those glyphs are all becoming @, and effectively floor
12:24:30 <Cheibriados> New branch created: xp_pool (1 commit) 13https://github.com/crawl/crawl/tree/xp_pool
12:24:30 <Cheibriados> 03PleasingFungus02 07[xp_pool] * 0.26-a0-813-g7f52359: Don't crash in the skills menu (advil) 10(65 seconds ago, 6 files, 32+ 10-) 13https://github.com/crawl/crawl/commit/7f52359de016
12:24:52 <Discord|> <10P​leasingFungus> very unconfident about whether this save_game is now safe, but maybe...?
12:37:31 <advil> looks plausible
12:38:13 <advil> is there a way to do something like `SHUFFLE: XYZ` but permute instead of potentially removing glyphs?
12:38:27 <advil> maybe something with this / syntax?
12:39:59 <dcss-ci> 04Build failed for 08xp_pool @ 7f52359d 06https://github.com/crawl/crawl/actions/runs/362035086
12:44:33 <Discord|> <10P​leasingFungus> ooh, good crash
12:47:21 <advil> the best I have come up with is something like `SHUFFLE: XYZ/XZY, XYZ/ZXY, XYZ/ZYX, XYZ/YXZ, XYZ/YZX`, but it seems unlikely that really makes the probability even
12:47:21 <advil> or maybe X/Y, X/Z, Y/X, Y/Z, Z/X, Z/Y
12:48:32 <advil> oh, that wouldn't work
12:51:22 <gammafunk> how do you mean by removing glyphs? can't you just do something with Y and Z that's not removal?
12:52:17 <advil> basically permute XYZ so that all three glyphs are guaranteed to be in the result
12:52:58 <advil> where just SHUFFLE: XYZ only guarantees that one of them will be there
12:52:58 <advil> I think I did figure out something better for the case I'm dealing with
12:52:58 <gammafunk> in this situation, are Y and Z not on the map?
12:54:01 <gammafunk> if so, then yeah you have to pace Y and Z before the shuffle somewhere, then do something with them after the shuffle
12:54:30 <advil> they are on the map...I'm trying to fix a map that used a shotgun shuffle + validation check to ensure at least one path to one of three exits; so my idea is to add explicit paths to each exit where at least one exit path needs to survive
12:55:04 <advil> where the paths are X Y and Z
12:55:25 <advil> I ended up doing `SHUFFLE: YZ, XY/YZ, XZ/ZX`, so X will always come out
12:56:42 <advil> but what I originally wanted was something like `SHUFFLE: XYZ` that gives you a result that is evenly weighted between XYZ, XZY, YXZ, YZX, ZXY, ZYX
12:58:35 <advil> permutation without repitition, basically
12:58:35 <advil> *repetition
13:01:10 <gammafunk> well, from my understanding, that's exactly what the non-slashed shuffle does.
13:01:19 <gammafunk> and from the docs:
13:01:20 <gammafunk> SHUFFLE: 123w     (i.e. list of glyphs, NOT slash-separated)
13:01:20 <gammafunk> could, for example, swap all occurrences of "1" with "2", as well as
13:01:23 <gammafunk> swapping all "3" with "w" (or any other of the 24 possibilities).
13:03:00 <gammafunk> it's my understanding that non-slashed shuffle is just a permutation of all glyphs weighted evenly, but I'm probably not understanding you well
13:03:04 <advil> shouldn't XYZ also include XXX, YYY, XYY, etc?
13:04:12 <advil> i.e. a permutation *with* repetition?
13:04:12 <gammafunk> no, it's a permutation in the sense of permutation of n objects, not (I guess you would say) sampling with replacement
13:05:01 <gammafunk> that's they the docs mention 24 possibilities and not 4^4 for four glyphs
13:05:08 <gammafunk> s/they/why/
13:05:08 <advil> oh hm
13:08:39 <advil> ok, yeah I was just misinterpreting the docs
13:54:52 <G-Flex> Can anyone tell me what slot the "big brained" Ds mutation takes up? Got it in trunk but the bot doesn't know about it
14:22:21 <Cheibriados> Minotaur generated on D:13, maybe a bug 13https://crawl.develz.org/mantis/view.php?id=12383 by Siegurt
14:22:58 <twelwe> FR: alternate swamp ending with 27 limb bunyip
15:00:38 <Discord|> <10P​leasingFungus> twelwe: into it
15:16:42 <advil> haha I stared at this for like 10 minutes before figuring out the mistake: SUBST:   rsn = . RSN = @, uUvVwWyYzZ = x
15:17:10 <advil> sometimes vault syntax is ...overly terse for my tastes
15:17:26 <advil> though I'm definitely getting a crash course in it
15:20:12 <geekosaur> missing comma?
15:20:18 <advil> yeah
15:20:36 <advil> which manifested only in weird positioning for @
15:21:14 <geekosaur> it's fortran all over again
16:03:54 <Cheibriados> New branch created: pull/1615 (1 commit) 13https://github.com/crawl/crawl/pull/1615
16:03:54 <Cheibriados> 03Nikolai Lavsky02 07https://github.com/crawl/crawl/pull/1615 * 0.26-a0-813-gad129e5: Colourize beam damage descriptions in WebTiles 10(9 minutes ago, 2 files, 6+ 3-) 13https://github.com/crawl/crawl/commit/ad129e5fd12a
17:55:50 <advil> !vault st_stairs_2
17:55:51 <Sequell> 1/1. https://github.com/crawl/crawl/blob/master/crawl-ref/source/dat/des/variable/mini_monsters.des#L3880
17:56:35 <advil> can anyone decode that NSUBST for me?
17:56:35 <advil> I'm pretty sure it's not doing anything reasonable, but I can't tell if it's doing that on purpose or not
17:56:56 -!- jfcaron_ is now known as jfcaron
17:56:58 <advil> (sequell is off by a few lines so I mean the one in the st_stairs_2 vault)
20:30:52 <Cbrotell> ziggurnaut (L5 FeNe) Crash caused by signal #6: Aborted (D:3)
20:53:43 <Cheibriados> 03PleasingFungus02 07* 0.26-a0-813-ge0c665c: Don't crash in the skills menu (advil) 10(9 hours ago, 6 files, 34+ 9-) 13https://github.com/crawl/crawl/commit/e0c665c49dfa
20:53:50 <Cheibriados> Branch xp_pool updated to be equal with master: 13https://github.com/crawl/crawl/tree/xp_pool
21:05:08 <dcss-ci> 04Build failed for 08xp_pool @ e0c665c4 06https://github.com/crawl/crawl/actions/runs/362643609
21:09:08 <dcss-ci> 04Build failed for 08master @ e0c665c4 06https://github.com/crawl/crawl/actions/runs/362643646
21:10:18 <advil> heh that ASSERT implies there is negative xp at some point
21:13:18 <Cheibriados> 03PleasingFungus02 07* 0.26-a0-814-ge366a86: Stop forcing Swamp 10(14 minutes ago, 3 files, 34+ 41-) 13https://github.com/crawl/crawl/commit/e366a860e629
21:17:22 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-813-ge0c665c49d (34)
21:21:31 <Kelbell> bgg1996 (L1 VSSu) ASSERT(you.exp_available >= 0) in 'skills.cc' at line 1354 failed. (D:1)
21:22:53 <dcss-ci> 04Build failed for 08master @ e366a860 06https://github.com/crawl/crawl/actions/runs/362661023
21:24:02 <advil> hm
21:25:01 <advil> !crashlog
21:25:01 <Sequell> 22722. bgg1996, XL1 VSSu, T:0 (milestone): https://crawl.kelbi.org/crawl/morgue/bgg1996/crash-bgg1996-20201114-022123.txt
21:26:48 <Kelbell> adviltest (L1 VSSu) ASSERT(you.exp_available >= 0) in 'skills.cc' at line 1354 failed. (D:1)
21:27:20 <Discord|> <06a​dvil> @PleasingFungus I was able to replicate that just by matching the combo it looks like
21:27:44 <advil> it's the same crash that's in the test in any case
21:28:03 <Discord|> <10P​leasingFungus> oh oops
21:28:05 <Discord|> <10P​leasingFungus> hm
21:28:32 <Discord|> <10P​leasingFungus> i was hoping i'd fixed that
21:28:44 <Discord|> <10P​leasingFungus> let me look
21:29:59 <Discord|> <10P​leasingFungus> i can also repro
21:30:01 <Discord|> <10P​leasingFungus> but don't really get it
21:30:38 <Discord|> <10P​leasingFungus> oh, startup issue maybe
21:31:02 <Discord|> <10P​leasingFungus> ...no, i don't get it.
21:31:46 <Cheibriados> 03PleasingFungus02 07* 0.26-a0-815-g3d2bcdb: Revert "Don't crash in the skills menu (advil)" 10(27 seconds ago, 6 files, 9+ 34-) 13https://github.com/crawl/crawl/commit/3d2bcdb4e9ce
21:32:14 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-814-ge366a860e6 (34)
21:33:05 <Kelbell> bgg1996 (L1 VSSu) ASSERT(you.exp_available >= 0) in 'skills.cc' at line 1354 failed. (D:1)
21:33:08 <Kelbell> Monjah (L1 OpIE) ASSERT(you.exp_available >= 0) in 'skills.cc' at line 1354 failed. (D:1)
21:36:54 <advil> yeah I got the crash on startup when starting with VSSu
21:38:29 <advil> is it related to xp aptitude somehow?
21:39:03 <advil> oh, nm, I didn't get it on HO and that is the same as VS
21:47:08 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-815-g3d2bcdb4e9 (34)
21:47:17 <Kelbell> Tezca (L1 FeTm) ASSERT(you.exp_available >= 0) in 'skills.cc' at line 1354 failed. (D:1)
21:55:22 <advil> til that there's a max xp that is not just some maxint value
21:55:59 <Discord|> <10P​leasingFungus> yeah i was surprised too
21:58:41 <advil> I don't *think* this is the crash (I'm in a state where recompiling to look at this is annoying) but probably you.experience_pool should be initialized in player::player()
22:01:51 <advil> although, that could do it
22:02:21 <advil> since the crash is in apply_xp which is apaprently called at startup assuming that experience_pool has a coherent value
22:03:49 <Discord|> <10P​leasingFungus> oops
22:03:50 <advil> and if it's an uninitialized variable that would explain why it has been coming and going a bit
22:03:55 <Discord|> <10P​leasingFungus> i only did it in tags
22:04:25 <advil> if that's it, I wonder if it was possible to get a positive number instead of a negative number :D
22:04:30 <Discord|> <10P​leasingFungus> are you fixing it?
22:04:39 <Discord|> <10P​leasingFungus> oh, great question
22:04:41 <advil> no, my repo is in a complcated state
22:04:53 <Discord|> <10P​leasingFungus> huh
22:37:13 <Bcadren> I'm usually for player/monster equality but a report on my fork is got me thinking that maybe AC_TRIPLE shouldn't work the same for players as enemies. It's just things like Iron Dragon's Scattershot breath and Deep Troll Earth Mage's LRD are in a state where they are very scary to felid and spriggan in robe; but do practically nothing through
22:37:13 <Bcadren> leather + auxiliaries due to triple reduction. (The impetus to me thinking about this was someone's spriggan getting oneshot by a monster with a Scattershot wand; which can't happen in mainline; but I think the AC_TRIPLE system is the root of the issue).
22:41:39 <Bcadren> Iron Dragon's Metal Splinters is 3d30 on hit which is average 46 before AC; 42 to a 3 AC felid, 31 at 10 AC, 16 for 20 AC. The heavy reduction means does nothing to medium and heavy armour characters. IDK. I'm stuck thinking about this.
22:43:37 <Cheibriados> 03PleasingFungus02 07* 0.26-a0-816-g42129fe: Revert "Revert "Don't crash in the skills menu (advil)"" 10(39 minutes ago, 6 files, 35+ 9-) 13https://github.com/crawl/crawl/commit/42129fe79a2c
22:47:29 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.26-a0-816-g42129fe79a (34)
23:45:24 <Cheibriados> New branch created: explode (2 commits) 13https://github.com/crawl/crawl/tree/explode
23:45:25 <Cheibriados> 03PleasingFungus02 07[explode] * 0.26-a0-817-gdbf270a: Remove dead code 10(29 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/dbf270a2a2cc
23:45:25 <Cheibriados> 03PleasingFungus02 07[explode] * 0.26-a0-818-g5a1802c: Split out mon-explode 10(20 seconds ago, 11 files, 326+ 279-) 13https://github.com/crawl/crawl/commit/5a1802c9b97b
23:46:59 <Cheibriados> 03PleasingFungus02 07* 0.26-a0-817-gdbf270a: Remove dead code 10(30 minutes ago, 1 file, 0+ 1-) 13https://github.com/crawl/crawl/commit/dbf270a2a2cc
23:55:04 <dcss-ci> 04Build failed for 08explode @ 5a1802c9 06https://github.com/crawl/crawl/actions/runs/362781187