00:17:07 <Cheibriados> The storm form gets caught in the web. 13https://crawl.develz.org/mantis/view.php?id=12627 by sdynet
00:17:07 -!- The topic of #crawl-dev is: Crawl Development | https://github.com/crawl/crawl | Logs: http://s-z.org/crawl-dev/ | People with +v have commit access | Please keep general Crawl-related chat to #crawl | Long stuff to a pastebin service, please
00:17:07 -!- The topic of #crawl is: Play Dungeon Crawl Stone Soup online now! Type ??online for instructions, ??lg / !lg for play stats | PM Sequell for long queries | http://crawl.develz.org | FooTV game replays: ??footv for instructions | #crawl-dev for dev discussion, #crawl-offtopic for offtopic
01:34:35 <Gretell> Unstable branch on crawl.develz.org updated to: 0.28-a0-124-g73ec313a74 (34)
01:56:38 <Gretell> Windows builds of master branch on crawl.develz.org updated to: 0.28-a0-124-g73ec313a74
02:22:56 -!- Guest52 is now known as twelwe
02:23:17 <Cbrotell> Unstable branch on cbro.berotato.org updated to: 0.28-a0-124-g73ec313a74 (34)
02:23:57 <twelwe> will high shields reduce the chance of riposting? i'm starting to rethink long blades on formicid because it seems im blocking attacks i'd rather outright avoid in melee
02:53:50 <Gretell> Monster database of master branch on crawl.develz.org updated to: 0.28-a0-124-g73ec313a74
03:01:40 <Yermak> Is this a feature that large abominations change their tile each time they're Irradiated? Text simply says "The large abomination is blasted with magical radiation".
03:47:00 <Kelbell> Fork (bcadrencrawl) on crawl.kelbi.org updated to: 0.03-622-g31332eff37
04:17:40 -!- allbery_b is now known as geekosaur
06:02:57 <Cheibriados> New branch created: pull/2062 (1 commit) 13https://github.com/crawl/crawl/pull/2062
06:02:57 <Cheibriados> 03Zhang Kai02 {GitHub} 07https://github.com/crawl/crawl/pull/2062 * 0.28-a0-125-ge06a0b3ce9: Update branches.txt 10(4 minutes ago, 1 file, 218+ 13-) 13https://github.com/crawl/crawl/commit/e06a0b3ce9c2
06:18:12 <Cheibriados> New branch created: pull/2063 (1 commit) 13https://github.com/crawl/crawl/pull/2063
06:18:12 <Cheibriados> 03Zhang Kai02 {GitHub} 07https://github.com/crawl/crawl/pull/2063 * 0.28-a0-125-gf353af8b66: Update branches.txt 10(5 minutes ago, 1 file, 223+ 13-) 13https://github.com/crawl/crawl/commit/f353af8b6665
06:22:29 <Lantell> Experimental (bcrawl) branch on underhound.eu updated to: 0.23-a0-4274-ga1336d5a09
08:05:04 <Cheibriados> Some keystrokes/combos do nothing 13https://crawl.develz.org/mantis/view.php?id=12628 by nubinia
08:44:07 <Cheibriados> 03ebering02 07[hot-hells] * 0.28-a0-121-gaaba44ed19: refactor: move padded_bounds to a layout library 10(20 hours ago, 4 files, 21+ 30-) 13https://github.com/crawl/crawl/commit/aaba44ed1969
08:44:07 <Cheibriados> 03ebering02 07[hot-hells] * 0.28-a0-122-gf8fb1a5dfd: fix: resolve a todo in zonify mapping 10(4 hours ago, 1 file, 10+ 6-) 13https://github.com/crawl/crawl/commit/f8fb1a5dfd55
08:44:07 <Cheibriados> 03ebering02 07[hot-hells] * 0.28-a0-123-g48fbf36410: feat: separate builder_bounds and max_bounds 10(3 hours ago, 4 files, 28+ 5-) 13https://github.com/crawl/crawl/commit/48fbf364106c
08:44:07 <Cheibriados> 03ebering02 07[hot-hells] * 0.28-a0-124-ge8565da66f: feat: shrink Hell subbranches on floors 1-6 10(3 hours ago, 13 files, 71+ 63-) 13https://github.com/crawl/crawl/commit/e8565da66f0e
09:03:09 <dcss-ci> 04Build failed for 08hot-hells @ e8565da6 06https://github.com/crawl/crawl/actions/runs/1130541108
09:26:27 <Discord|> <12e​bering> it begins
09:29:54 <Cheibriados> New branch created: pull/2064 (10 commits) 13https://github.com/crawl/crawl/pull/2064
09:29:54 <Cheibriados> Branch pull/2064 updated to be equal with hot-hells: 13https://github.com/crawl/crawl/pull/2064
09:35:30 <Discord|> <12e​bering> @nicolae you might be interested in the vaulting section of that PR
10:16:45 <Discord|> <12e​bering> @advil I seem to have hit some kind of seed stability or system portability bug with how I've modified levelgen for the hells. In CI for that PR test #29 fails with a builder crash but when I do a local fulldebug build the test does not fail...
10:18:43 <advil> hm in CI it looks like the level is entirely filled with wall
10:19:08 <advil> in dis
10:20:58 <Discord|> <12e​bering> yes, that's a sign that the builder is crashing early on in processing the layout
10:21:05 <Discord|> <12e​bering> but the crash dump in CI doesn't say which layout
10:21:31 <Discord|> <12e​bering> there are a couple of dis layouts that use the box function that's segfaulting but repeatedly re-running them with &P hasn't done anything for me
10:23:48 <advil> it's too bad gdb in CI is lacking the symbol table, is that new?
10:27:46 <Discord|> <12e​bering> not sure...
10:28:31 <advil> yeah, also not replicating that locally, are you on your mac?
10:30:59 <advil> ah I do get it when I run the test directly
10:31:15 <Discord|> <12e​bering> I am on my mac
10:31:24 <Discord|> <12e​bering> how do you mean run the test directly? ./crawl -test 29?
10:34:09 <advil> ./crawl -test unique.lua
10:34:24 <advil> it may be sporadic, or something?
10:34:29 <advil> I'm not getting it in lldb now
10:35:01 <advil> it was specifically while running the randomized placement test for uniques in dis, so I would guess the crash is likely while resetting the level (which that test does a lot)
10:35:19 <advil> in dis 6
10:35:23 <advil> when I saw it at least
10:35:40 <advil> it's also reporting that the crash is happening in _fill_area
10:35:58 <advil> so that's at least slightly more specific
10:36:50 <advil> this is 85% an uninitialized variable
10:37:22 <advil> based on the signature "more likely in gcc, doesn't replicate consistently, somewhat different across machines"
10:39:58 <Discord|> <10P​leasingFungus> !time
10:39:58 <Discord|> <04C​erebot> Time: Aug 14, 2021, 02:39:58 PM, UTC. The 2021 0.27 tournament ends in 1 day, 5 hours, 20 minutes and 1 second.
10:40:03 <advil> aha, finally replicated in lldb
10:40:11 <Discord|> <10P​leasingFungus> well done!
10:41:19 <advil> ugh, my terminal is messed up so this is a bit hard to work with
10:41:58 <advil> it seems to have crashed on `lines(x, y) = fill` with x=20, y=42
10:42:26 <advil> maybe there's an off-by-one error on the size of lines w.r.t. map bounds for the reduced map size?
10:43:05 <advil> not really sure how that variable gets set up
10:45:12 <Discord|> <08n​icolae> hellvaults...
10:45:18 <Discord|> <12e​bering> hm yeah there's definitely an off-by-one error in there
10:45:33 <Discord|> <12e​bering> there have been several bounds checking issues like that that I've uncovered
10:45:58 <Discord|> <12e​bering> where things have just worked because they fit in GXM GYM bounds and these functions weren't being used on partial or smaller maps
10:47:02 <advil> it's too bad that _fill_area doesn't seem to do any explicit bounds checking for its args
10:47:21 <advil> or maybe dgn_fill_area
10:48:00 <advil> btw I replicated the crash maybe 1 out of 15 times, to get it I deleted everything but dis from uniques.lua (for speed) and reran many times
10:48:33 <advil> it's presumably contingent on exactly what's up with the adjacent memory to lines
10:49:18 <Discord|> <12e​bering> does it replicate on fill_area every time?
10:49:39 <Discord|> <12e​bering> well this will certainly be an adventure
10:49:39 <advil> well, I've only seen it twice, but yes so far
10:49:46 <Discord|> <12e​bering> 👍
10:55:02 <advil> it's always crashing in the same place but the coordinates have varied a bit: x=25 y=42; x=14 y=37; x=20 y=42 (again)
10:55:23 <advil> btw there may be something off with seeding here because a test like this in the lua test bed should be deterministic
10:55:36 <advil> it uses seed 1 iirc
11:02:26 <Discord|> <12e​bering> yeah thats why I pinged you when I couldn
11:02:30 <Discord|> <12e​bering> t repro locally
11:06:40 <advil> I think the variability is partly just that clang is more robust to crashes from bad memory access (somewhat annoyingly so)
11:13:13 <Discord|> <06a​dvil> there is actually some code here for checking coords with lines that you can use to get a more controlled version of this crash; When I insert a diagnostic ASSERT in the inner loop of _fill_area here's what I fairly reliably get:
11:13:16 <Discord|> <06a​dvil> ASSERT(_valid_coord(ls, lines, x, y, false)) in 'l-dgnbld.cc' at line 155 failed. (10,36 invalid relative to 10,10-15,40, lines size 40,35)
11:13:50 <Discord|> <06a​dvil> those numbers are x,y, x1,y1-x2,y2, and the sizes are from lines.width(), lines, height()
11:14:32 <Discord|> <06a​dvil> well here's the exact ASSERT in case you want to use it, I'm going to have to stop for the moment:
11:14:41 <Discord|> <06a​dvil>             ASSERTM(_valid_coord(ls, lines, x, y, false),                 "%d,%d invalid relative to %d,%d-%d,%d, lines size %d,%d", x, y, x1, y1, x2, y2, lines.width(), lines.height());
11:14:56 <Discord|> <12e​bering> Thanks!
11:15:11 <Discord|> <12e​bering> I should be able to hunt it down from there
11:15:13 <Discord|> <06a​dvil> it still doesn't happen every time but is a lot easier to get this way
11:16:00 <Discord|> <06a​dvil> for the seeding issue, not immediately sure, it could actually be caused by this (since uninit'd memory is a source of independent randomness)
11:18:53 -!- MakMorn_ is now known as MakMorn
12:02:40 <Cheibriados> New branch created: pull/2065 (1 commit) 13https://github.com/crawl/crawl/pull/2065
12:02:40 <Cheibriados> 03Matthew Daley02 07https://github.com/crawl/crawl/pull/2065 * 0.28-a0-125-g871d5a2e2c: feat: Show turn counts and real game times in lobby 10(23 minutes ago, 4 files, 59+ 17-) 13https://github.com/crawl/crawl/commit/871d5a2e2c3a
13:06:17 <Henzell> Unstable branch on crawl.akrasiac.org updated to: 0.28-a0-124-g73ec313 (34)
13:46:02 <Discord|> <03w​heals> @gammafunk seems like you can disable the github webhook now that chei is back
13:47:24 <Discord|> <09g​ammafunk> cool
13:48:15 <Discord|> <09g​ammafunk> disabled
15:23:02 <Discord|> <08n​icolae> i guess while the hell reform is on the table i should look into the hell wizmode crash bugs i talked about a while back
15:33:59 <Discord|> <08n​icolae> HOW TO DO THE WIZMODE HELL CRASH BUG: 0) Local tiles version 0.27 downloaded from crawl.develz.org. 1a) Go to the Vestibule of Hell in Wizmode via &~. OR 1b) Go to the Vestibule of Hell by going to Depths, finding a Hell entry, and entering Hell the standard way. 2) Regenerate the map using &^R. 3a) Take a few steps in some direction; I'm not sure what about walking triggers it OR 3b) Try to leave via
15:34:00 <Discord|> the stairs out of the Vestibule or by one of the stairs leading to a Hell. 4) Get the error c:\crawl\trunk\stone_soup-tiles-0.27>ASSERT(num_visits == 0 && levels_seen == 0 || num_visits > 0 && levels_seen > 0) in 'place-info.cc' at line 29 failed.
15:35:22 <Discord|> <08n​icolae> i don't know if i should post that on github or what
15:37:29 <Yermak> Orc knight dies, triggers its inner flame which damages newly generated spectral orc knight from Death Channel. Something seems off.
15:38:21 <Discord|> <10P​leasingFungus> does this get in the way of development?
15:38:42 <Discord|> <08n​icolae> well it makes it rather hard to test hell entrance subvaults
15:39:04 <Discord|> <10P​leasingFungus> why do you need to leave via the stairs?
15:39:37 <Discord|> <08n​icolae> it's not always leaving via the stairs, sometimes it triggers just by moving around the map, but it's easier to trigger the crash intentionally by taking the stairs
15:40:00 <Discord|> <10P​leasingFungus> i see
15:40:09 <Discord|> <10P​leasingFungus> a gh issue is reasonable
15:52:27 <Discord|> <08n​icolae> for some reason github doesn't accept attaching my morgue crash file
15:54:41 <Discord|> <08n​icolae> oh there we go
15:58:56 <Discord|> <05k​ate> i think that might also not be hell-specific, i think i've seen the same crash after using &^R elsewhere
16:02:08 <Discord|> <08n​icolae> insidious...
16:19:09 <Yermak> FR: don't let inner flame autotarget already affected creature, and don't let its target to cycle through palyer.
18:24:06 <Lantell> Unstable branch on underhound.eu updated to: 0.28-a0-124-g73ec313a74 (34)
19:05:47 -!- allbery_b is now known as geekosaur
19:07:46 -!- Guest42 is now known as twelwe
19:42:28 <Cheibriados> 03kate-02 07* 0.28-a0-125-g8d3330d46d: Make potions of attraction wake up monsters 10(3 hours ago, 1 file, 1+ 0-) 13https://github.com/crawl/crawl/commit/8d3330d46d1c
19:42:28 <Cheibriados> 03kate-02 07* 0.28-a0-126-gf70498e230: Improve vampire transformation messages 10(2 hours ago, 1 file, 2+ 2-) 13https://github.com/crawl/crawl/commit/f70498e2302f
19:42:28 <Cheibriados> 03kate-02 07* 0.28-a0-127-ge31878e1fe: Don't consider rings of rPois useless for bloodless vampires 10(2 hours ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/e31878e1fe37
19:42:28 <Cheibriados> 03kate-02 07* 0.28-a0-128-gcbb86046cc: Rebrace 10(2 hours ago, 1 file, 2+ 1-) 13https://github.com/crawl/crawl/commit/cbb86046ccd9
19:47:17 <Kelbell> Unstable branch on crawl.kelbi.org updated to: 0.28-a0-128-gcbb86046cc (34)
20:05:50 <Discord|> <10P​leasingFungus> dang, kate, brutal nerf to attraction stab tech
20:07:28 <Discord|> <05k​ate> just too powerful...
20:07:29 <Discord|> <10P​leasingFungus> not sure how i feel about that…
20:07:33 <Discord|> <10P​leasingFungus> i mean
20:07:40 <Discord|> <10P​leasingFungus> people did seem to enjoy it
20:07:43 <Discord|> <10P​leasingFungus> idk
20:07:50 <Discord|> <10P​leasingFungus> will put in some deep thonks
20:16:35 <twelwe> does high shields reduce riposte chance?
20:29:11 <Discord|> <08n​icolae> oh nice, dungeon.lua is included by default by the .des handler, and dungeon.lua requires util.lua, so i think util.lua is also included by default
20:39:23 <Discord|> <10P​leasingFungus> twelwe: yeah, unfortunately. i need to redesign riposte at some point
20:39:29 <Discord|> <10P​leasingFungus> or hellmonk can meme it
20:41:22 <twelwe> ok thanks PF. I thought my formicids were taking less riposte when i went to shield of ignorance
20:43:35 <twelwe> honestly, i think thats alright - I actually would like to see riposte be a smaller chance, and a normally generating glove ego. long blades could get extra benefit from egos instead of riposte, or give them more unique interactions with monsters like hydras/fiery blades
20:44:28 <twelwe> long blades have more surface area covered in frost/fire/poison than, say, a dagger or mace
20:49:23 <twelwe> hey PF are you west coast canada?
20:49:40 <Discord|> <10P​leasingFungus> i am not canadian, sorry
20:50:24 <twelwe> wait cali?
20:50:32 <Discord|> <10P​leasingFungus> yeah, the other ca
20:50:35 <twelwe> ahh
20:51:13 <twelwe> i'm going to live on the beach tomorrow! seabrook washington
20:53:19 <Discord|> <10P​leasingFungus> congrat!
20:53:23 <Discord|> <10P​leasingFungus> sounds fun 🙂
20:54:34 <twelwe> i'll be sitting in the shade playing formicids, recovering from sunburn
21:27:00 -!- allbery_b is now known as geekosaur
21:36:36 <Discord|> <09g​ammafunk> California? I recognize the state. See you soon
21:42:45 <Discord|> <10P​leasingFungus> let me know when you’re here, i’ll put on a kettle for tea
22:23:40 <Discord|> <08n​icolae> what about califiveya
22:46:20 <twelwe> damn i was just down in thousand oaks 72 hours ago
22:47:05 <twelwe> ill be in that area three times a year for a weekend
23:26:54 <Yermak> Guys, I thought berserkering monsters can't perform tactical retreats! Duvessa just started running away from me swimming in deep water.
23:30:18 <Discord|> <10P​leasingFungus> lol
23:30:26 <Discord|> <10P​leasingFungus> pathing is hard
23:45:30 <twelwe> Bugfix: Duvessa no longer tries to drown herself during a fit of rage